feat: chat on mps

.gitattributes

@@ -0,0 +1 @@
+paper_plot/data/big_graph_degree_data.npz filter=lfs diff=lfs merge=lfs -text

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -1,50 +0,0 @@
-name: Bug Report
-description: Report a bug in LEANN
-labels: ["bug"]
-
-body:
-  - type: textarea
-    id: description
-    attributes:
-      label: What happened?
-      description: A clear description of the bug
-    validations:
-      required: true
-
-  - type: textarea
-    id: reproduce
-    attributes:
-      label: How to reproduce
-      placeholder: |
-        1. Install with...
-        2. Run command...
-        3. See error
-    validations:
-      required: true
-
-  - type: textarea
-    id: error
-    attributes:
-      label: Error message
-      description: Paste any error messages
-      render: shell
-
-  - type: input
-    id: version
-    attributes:
-      label: LEANN Version
-      placeholder: "0.1.0"
-    validations:
-      required: true
-
-  - type: dropdown
-    id: os
-    attributes:
-      label: Operating System
-      options:
-        - macOS
-        - Linux
-        - Windows
-        - Docker
-    validations:
-      required: true

.github/ISSUE_TEMPLATE/config.yml

@@ -1,8 +0,0 @@
-blank_issues_enabled: true
-contact_links:
-  - name: Documentation
-    url: https://github.com/LEANN-RAG/LEANN-RAG/tree/main/docs
-    about: Read the docs first
-  - name: Discussions
-    url: https://github.com/LEANN-RAG/LEANN-RAG/discussions
-    about: Ask questions and share ideas

.github/ISSUE_TEMPLATE/feature_request.yml

@@ -1,27 +0,0 @@
-name: Feature Request
-description: Suggest a new feature for LEANN
-labels: ["enhancement"]
-
-body:
-  - type: textarea
-    id: problem
-    attributes:
-      label: What problem does this solve?
-      description: Describe the problem or need
-    validations:
-      required: true
-
-  - type: textarea
-    id: solution
-    attributes:
-      label: Proposed solution
-      description: How would you like this to work?
-    validations:
-      required: true
-
-  - type: textarea
-    id: example
-    attributes:
-      label: Example usage
-      description: Show how the API might look
-      render: python

.github/pull_request_template.md

@@ -1,13 +0,0 @@
-## What does this PR do?
-
-<!-- Brief description of your changes -->
-
-## Related Issues
-
-Fixes #
-
-## Checklist
-
-- [ ] Tests pass (`uv run pytest`)
-- [ ] Code formatted (`ruff format` and `ruff check`)
-- [ ] Pre-commit hooks pass (`pre-commit run --all-files`)

.github/workflows/build-and-publish.yml

@@ -1,12 +0,0 @@
-name: CI
-
-on:
-  push:
-    branches: [ main ]
-  pull_request:
-    branches: [ main ]
-  workflow_dispatch:
-
-jobs:
-  build:
-    uses: ./.github/workflows/build-reusable.yml

.github/workflows/build-reusable.yml

@@ -1,402 +0,0 @@
-name: Reusable Build
-
-on:
-  workflow_call:
-    inputs:
-      ref:
-        description: 'Git ref to build'
-        required: false
-        type: string
-        default: ''
-
-jobs:
-  lint:
-    name: Lint and Format Check
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          ref: ${{ inputs.ref }}
-
-      - name: Setup Python
-        uses: actions/setup-python@v5
-        with:
-          python-version: '3.11'
-
-      - name: Install uv
-        uses: astral-sh/setup-uv@v4
-
-      - name: Install ruff
-        run: |
-          uv tool install ruff
-
-      - name: Run ruff check
-        run: |
-          ruff check .
-
-      - name: Run ruff format check
-        run: |
-          ruff format --check .
-
-  build:
-    needs: lint
-    name: Build ${{ matrix.os }} Python ${{ matrix.python }}
-    strategy:
-      matrix:
-        include:
-          - os: ubuntu-22.04
-            python: '3.9'
-          - os: ubuntu-22.04
-            python: '3.10'
-          - os: ubuntu-22.04
-            python: '3.11'
-          - os: ubuntu-22.04
-            python: '3.12'
-          - os: ubuntu-22.04
-            python: '3.13'
-          # ARM64 Linux builds
-          - os: ubuntu-24.04-arm
-            python: '3.9'
-          - os: ubuntu-24.04-arm
-            python: '3.10'
-          - os: ubuntu-24.04-arm
-            python: '3.11'
-          - os: ubuntu-24.04-arm
-            python: '3.12'
-          - os: ubuntu-24.04-arm
-            python: '3.13'
-          - os: macos-14
-            python: '3.9'
-          - os: macos-14
-            python: '3.10'
-          - os: macos-14
-            python: '3.11'
-          - os: macos-14
-            python: '3.12'
-          - os: macos-14
-            python: '3.13'
-          - os: macos-15
-            python: '3.9'
-          - os: macos-15
-            python: '3.10'
-          - os: macos-15
-            python: '3.11'
-          - os: macos-15
-            python: '3.12'
-          - os: macos-15
-            python: '3.13'
-          - os: macos-13
-            python: '3.9'
-          - os: macos-13
-            python: '3.10'
-          - os: macos-13
-            python: '3.11'
-          - os: macos-13
-            python: '3.12'
-          # Note: macos-13 + Python 3.13 excluded due to PyTorch compatibility
-          # (PyTorch 2.5+ supports Python 3.13 but not Intel Mac x86_64)
-    runs-on: ${{ matrix.os }}
-
-    steps:
-      - uses: actions/checkout@v5
-        with:
-          ref: ${{ inputs.ref }}
-          submodules: recursive
-
-      - name: Setup Python
-        uses: actions/setup-python@v5
-        with:
-          python-version: ${{ matrix.python }}
-
-      - name: Install uv
-        uses: astral-sh/setup-uv@v6
-
-      - name: Install system dependencies (Ubuntu)
-        if: runner.os == 'Linux'
-        run: |
-          sudo apt-get update
-          sudo apt-get install -y libomp-dev libboost-all-dev protobuf-compiler libzmq3-dev \
-            pkg-config libabsl-dev libaio-dev libprotobuf-dev \
-            patchelf
-
-          # Debug: Show system information
-          echo "🔍 System Information:"
-          echo "Architecture: $(uname -m)"
-          echo "OS: $(uname -a)"
-          echo "CPU info: $(lscpu | head -5)"
-
-          # Install math library based on architecture
-          ARCH=$(uname -m)
-          echo "🔍 Setting up math library for architecture: $ARCH"
-
-          if [[ "$ARCH" == "x86_64" ]]; then
-            # Install Intel MKL for DiskANN on x86_64
-            echo "📦 Installing Intel MKL for x86_64..."
-            wget -q https://registrationcenter-download.intel.com/akdlm/IRC_NAS/79153e0f-74d7-45af-b8c2-258941adf58a/intel-onemkl-2025.0.0.940.sh
-            sudo sh intel-onemkl-2025.0.0.940.sh -a --components intel.oneapi.lin.mkl.devel --action install --eula accept -s
-            source /opt/intel/oneapi/setvars.sh
-            echo "MKLROOT=/opt/intel/oneapi/mkl/latest" >> $GITHUB_ENV
-            echo "LD_LIBRARY_PATH=/opt/intel/oneapi/compiler/latest/linux/compiler/lib/intel64_lin" >> $GITHUB_ENV
-            echo "LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/intel/oneapi/mkl/latest/lib/intel64" >> $GITHUB_ENV
-            echo "✅ Intel MKL installed for x86_64"
-
-            # Debug: Check MKL installation
-            echo "🔍 MKL Installation Check:"
-            ls -la /opt/intel/oneapi/mkl/latest/ || echo "MKL directory not found"
-            ls -la /opt/intel/oneapi/mkl/latest/lib/ || echo "MKL lib directory not found"
-
-          elif [[ "$ARCH" == "aarch64" ]]; then
-            # Use OpenBLAS for ARM64 (MKL installer not compatible with ARM64)
-            echo "📦 Installing OpenBLAS for ARM64..."
-            sudo apt-get install -y libopenblas-dev liblapack-dev liblapacke-dev
-            echo "✅ OpenBLAS installed for ARM64"
-
-            # Debug: Check OpenBLAS installation
-            echo "🔍 OpenBLAS Installation Check:"
-            dpkg -l | grep openblas || echo "OpenBLAS package not found"
-            ls -la /usr/lib/aarch64-linux-gnu/openblas/ || echo "OpenBLAS directory not found"
-          fi
-
-          # Debug: Show final library paths
-          echo "🔍 Final LD_LIBRARY_PATH: $LD_LIBRARY_PATH"
-
-      - name: Install system dependencies (macOS)
-        if: runner.os == 'macOS'
-        run: |
-          # Don't install LLVM, use system clang for better compatibility
-          brew install libomp boost protobuf zeromq
-
-      - name: Install build dependencies
-        run: |
-          uv pip install --system scikit-build-core numpy swig Cython pybind11
-          if [[ "$RUNNER_OS" == "Linux" ]]; then
-            uv pip install --system auditwheel
-          else
-            uv pip install --system delocate
-          fi
-
-      - name: Set macOS environment variables
-        if: runner.os == 'macOS'
-        run: |
-          # Use brew --prefix to automatically detect Homebrew installation path
-          HOMEBREW_PREFIX=$(brew --prefix)
-          echo "HOMEBREW_PREFIX=${HOMEBREW_PREFIX}" >> $GITHUB_ENV
-          echo "OpenMP_ROOT=${HOMEBREW_PREFIX}/opt/libomp" >> $GITHUB_ENV
-
-          # Set CMAKE_PREFIX_PATH to let CMake find all packages automatically
-          echo "CMAKE_PREFIX_PATH=${HOMEBREW_PREFIX}" >> $GITHUB_ENV
-
-          # Set compiler flags for OpenMP (required for both backends)
-          echo "LDFLAGS=-L${HOMEBREW_PREFIX}/opt/libomp/lib" >> $GITHUB_ENV
-          echo "CPPFLAGS=-I${HOMEBREW_PREFIX}/opt/libomp/include" >> $GITHUB_ENV
-
-      - name: Build packages
-        run: |
-          # Build core (platform independent)
-          cd packages/leann-core
-          uv build
-          cd ../..
-
-          # Build HNSW backend
-          cd packages/leann-backend-hnsw
-          if [[ "${{ matrix.os }}" == macos-* ]]; then
-            # Use system clang for better compatibility
-            export CC=clang
-            export CXX=clang++
-            # Homebrew libraries on each macOS version require matching minimum version
-            if [[ "${{ matrix.os }}" == "macos-13" ]]; then
-              export MACOSX_DEPLOYMENT_TARGET=13.0
-            elif [[ "${{ matrix.os }}" == "macos-14" ]]; then
-              export MACOSX_DEPLOYMENT_TARGET=14.0
-            elif [[ "${{ matrix.os }}" == "macos-15" ]]; then
-              export MACOSX_DEPLOYMENT_TARGET=15.0
-            fi
-            uv build --wheel --python ${{ matrix.python }} --find-links ${GITHUB_WORKSPACE}/packages/leann-core/dist
-          else
-            uv build --wheel --python ${{ matrix.python }} --find-links ${GITHUB_WORKSPACE}/packages/leann-core/dist
-          fi
-          cd ../..
-
-          # Build DiskANN backend
-          cd packages/leann-backend-diskann
-          if [[ "${{ matrix.os }}" == macos-* ]]; then
-            # Use system clang for better compatibility
-            export CC=clang
-            export CXX=clang++
-            # DiskANN requires macOS 13.3+ for sgesdd_ LAPACK function
-            # But Homebrew libraries on each macOS version require matching minimum version
-            if [[ "${{ matrix.os }}" == "macos-13" ]]; then
-              export MACOSX_DEPLOYMENT_TARGET=13.3
-            elif [[ "${{ matrix.os }}" == "macos-14" ]]; then
-              export MACOSX_DEPLOYMENT_TARGET=14.0
-            elif [[ "${{ matrix.os }}" == "macos-15" ]]; then
-              export MACOSX_DEPLOYMENT_TARGET=15.0
-            fi
-            uv build --wheel --python ${{ matrix.python }} --find-links ${GITHUB_WORKSPACE}/packages/leann-core/dist
-          else
-            uv build --wheel --python ${{ matrix.python }} --find-links ${GITHUB_WORKSPACE}/packages/leann-core/dist
-          fi
-          cd ../..
-
-          # Build meta package (platform independent)
-          cd packages/leann
-          uv build
-          cd ../..
-
-      - name: Repair wheels (Linux)
-        if: runner.os == 'Linux'
-        run: |
-          # Repair HNSW wheel
-          cd packages/leann-backend-hnsw
-          if [ -d dist ]; then
-            auditwheel repair dist/*.whl -w dist_repaired
-            rm -rf dist
-            mv dist_repaired dist
-          fi
-          cd ../..
-
-          # Repair DiskANN wheel
-          cd packages/leann-backend-diskann
-          if [ -d dist ]; then
-            auditwheel repair dist/*.whl -w dist_repaired
-            rm -rf dist
-            mv dist_repaired dist
-          fi
-          cd ../..
-
-      - name: Repair wheels (macOS)
-        if: runner.os == 'macOS'
-        run: |
-          # Determine deployment target based on runner OS
-          # Must match the Homebrew libraries for each macOS version
-          if [[ "${{ matrix.os }}" == "macos-13" ]]; then
-            HNSW_TARGET="13.0"
-            DISKANN_TARGET="13.3"
-          elif [[ "${{ matrix.os }}" == "macos-14" ]]; then
-            HNSW_TARGET="14.0"
-            DISKANN_TARGET="14.0"
-          elif [[ "${{ matrix.os }}" == "macos-15" ]]; then
-            HNSW_TARGET="15.0"
-            DISKANN_TARGET="15.0"
-          fi
-
-          # Repair HNSW wheel
-          cd packages/leann-backend-hnsw
-          if [ -d dist ]; then
-            export MACOSX_DEPLOYMENT_TARGET=$HNSW_TARGET
-            delocate-wheel -w dist_repaired -v --require-target-macos-version $HNSW_TARGET dist/*.whl
-            rm -rf dist
-            mv dist_repaired dist
-          fi
-          cd ../..
-
-          # Repair DiskANN wheel
-          cd packages/leann-backend-diskann
-          if [ -d dist ]; then
-            export MACOSX_DEPLOYMENT_TARGET=$DISKANN_TARGET
-            delocate-wheel -w dist_repaired -v --require-target-macos-version $DISKANN_TARGET dist/*.whl
-            rm -rf dist
-            mv dist_repaired dist
-          fi
-          cd ../..
-
-      - name: List built packages
-        run: |
-          echo "📦 Built packages:"
-          find packages/*/dist -name "*.whl" -o -name "*.tar.gz" | sort
-
-
-      - name: Install built packages for testing
-        run: |
-          # Create a virtual environment with the correct Python version
-          uv venv --python ${{ matrix.python }}
-          source .venv/bin/activate || source .venv/Scripts/activate
-
-          # Install packages using --find-links to prioritize local builds
-          uv pip install --find-links packages/leann-core/dist --find-links packages/leann-backend-hnsw/dist --find-links packages/leann-backend-diskann/dist packages/leann-core/dist/*.whl || uv pip install --find-links packages/leann-core/dist packages/leann-core/dist/*.tar.gz
-          uv pip install --find-links packages/leann-core/dist packages/leann-backend-hnsw/dist/*.whl
-          uv pip install --find-links packages/leann-core/dist packages/leann-backend-diskann/dist/*.whl
-          uv pip install packages/leann/dist/*.whl || uv pip install packages/leann/dist/*.tar.gz
-
-          # Install test dependencies using extras
-          uv pip install -e ".[test]"
-
-      - name: Run tests with pytest
-        env:
-          CI: true
-          OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
-          HF_HUB_DISABLE_SYMLINKS: 1
-          TOKENIZERS_PARALLELISM: false
-          PYTORCH_ENABLE_MPS_FALLBACK: 0
-          OMP_NUM_THREADS: 1
-          MKL_NUM_THREADS: 1
-        run: |
-          source .venv/bin/activate || source .venv/Scripts/activate
-          pytest tests/ -v --tb=short
-
-      - name: Run sanity checks (optional)
-        run: |
-          # Activate virtual environment
-          source .venv/bin/activate || source .venv/Scripts/activate
-
-          # Run distance function tests if available
-          if [ -f test/sanity_checks/test_distance_functions.py ]; then
-            echo "Running distance function sanity checks..."
-            python test/sanity_checks/test_distance_functions.py || echo "⚠️ Distance function test failed, continuing..."
-          fi
-
-      - name: Upload artifacts
-        uses: actions/upload-artifact@v4
-        with:
-          name: packages-${{ matrix.os }}-py${{ matrix.python }}
-          path: packages/*/dist/
-
-
-  arch-smoke:
-    name: Arch Linux smoke test (install & import)
-    needs: build
-    runs-on: ubuntu-latest
-    container:
-      image: archlinux:latest
-
-    steps:
-      - name: Prepare system
-        run: |
-          pacman -Syu --noconfirm
-          pacman -S --noconfirm python python-pip gcc git zlib openssl
-
-      - name: Download ALL wheel artifacts from this run
-        uses: actions/download-artifact@v5
-        with:
-          # Don't specify name, download all artifacts
-          path: ./wheels
-
-      - name: Install uv
-        uses: astral-sh/setup-uv@v6
-
-      - name: Create virtual environment and install wheels
-        run: |
-          uv venv
-          source .venv/bin/activate || source .venv/Scripts/activate
-          uv pip install --find-links wheels leann-core
-          uv pip install --find-links wheels leann-backend-hnsw
-          uv pip install --find-links wheels leann-backend-diskann
-          uv pip install --find-links wheels leann
-
-      - name: Import & tiny runtime check
-        env:
-          OMP_NUM_THREADS: 1
-          MKL_NUM_THREADS: 1
-        run: |
-          source .venv/bin/activate || source .venv/Scripts/activate
-          python - <<'PY'
-          import leann
-          import leann_backend_hnsw as h
-          import leann_backend_diskann as d
-          from leann import LeannBuilder, LeannSearcher
-          b = LeannBuilder(backend_name="hnsw")
-          b.add_text("hello arch")
-          b.build_index("arch_demo.leann")
-          s = LeannSearcher("arch_demo.leann")
-          print("search:", s.search("hello", top_k=1))
-          PY

.github/workflows/link-check.yml

@@ -1,19 +0,0 @@
-name: Link Check
-
-on:
-  push:
-    branches: [ main, master ]
-  pull_request:
-  schedule:
-    - cron: "0 3 * * 1"
-
-jobs:
-  link-check:
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v4
-      - uses: lycheeverse/lychee-action@v2
-        with:
-          args: --no-progress --insecure --user-agent 'curl/7.68.0' README.md docs/ apps/ examples/ benchmarks/
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/release-manual.yml

@@ -1,129 +0,0 @@
-name: Release
-
-on:
-  workflow_dispatch:
-    inputs:
-      version:
-        description: 'Version to release (e.g., 0.1.2)'
-        required: true
-        type: string
-
-jobs:
-  update-version:
-    name: Update Version
-    runs-on: ubuntu-latest
-    permissions:
-      contents: write
-    outputs:
-      commit-sha: ${{ steps.push.outputs.commit-sha }}
-
-    steps:
-      - uses: actions/checkout@v4
-
-      - name: Validate version
-        run: |
-          # Remove 'v' prefix if present for validation
-          VERSION_CLEAN="${{ inputs.version }}"
-          VERSION_CLEAN="${VERSION_CLEAN#v}"
-          if ! [[ "$VERSION_CLEAN" =~ ^[0-9]+\.[0-9]+\.[0-9]+$ ]]; then
-            echo "❌ Invalid version format. Expected format: X.Y.Z or vX.Y.Z"
-            exit 1
-          fi
-          echo "✅ Version format valid: ${{ inputs.version }}"
-
-      - name: Update versions and push
-        id: push
-        run: |
-          # Check current version
-          CURRENT_VERSION=$(grep "^version" packages/leann-core/pyproject.toml | cut -d'"' -f2)
-          echo "Current version: $CURRENT_VERSION"
-          echo "Target version: ${{ inputs.version }}"
-
-          if [ "$CURRENT_VERSION" = "${{ inputs.version }}" ]; then
-            echo "⚠️  Version is already ${{ inputs.version }}, skipping update"
-            COMMIT_SHA=$(git rev-parse HEAD)
-          else
-            ./scripts/bump_version.sh ${{ inputs.version }}
-            git config user.name "GitHub Actions"
-            git config user.email "actions@github.com"
-            git add packages/*/pyproject.toml
-            git commit -m "chore: release v${{ inputs.version }}"
-            git push origin main
-            COMMIT_SHA=$(git rev-parse HEAD)
-            echo "✅ Pushed version update: $COMMIT_SHA"
-          fi
-
-          echo "commit-sha=$COMMIT_SHA" >> $GITHUB_OUTPUT
-
-  build-packages:
-    name: Build packages
-    needs: update-version
-    uses: ./.github/workflows/build-reusable.yml
-    with:
-      ref: 'main'
-
-  publish:
-    name: Publish and Release
-    needs: [update-version, build-packages]
-    if: always() && needs.update-version.result == 'success' && needs.build-packages.result == 'success'
-    runs-on: ubuntu-latest
-    permissions:
-      contents: write
-
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          ref: 'main'
-
-      - name: Download all artifacts
-        uses: actions/download-artifact@v4
-        with:
-          path: dist-artifacts
-
-      - name: Collect packages
-        run: |
-          mkdir -p dist
-          find dist-artifacts -name "*.whl" -exec cp {} dist/ \;
-          find dist-artifacts -name "*.tar.gz" -exec cp {} dist/ \;
-
-          echo "📦 Packages to publish:"
-          ls -la dist/
-
-      - name: Publish to PyPI
-        env:
-          TWINE_USERNAME: __token__
-          TWINE_PASSWORD: ${{ secrets.PYPI_API_TOKEN }}
-        run: |
-          if [ -z "$TWINE_PASSWORD" ]; then
-            echo "❌ PYPI_API_TOKEN not configured!"
-            exit 1
-          fi
-
-          pip install twine
-          twine upload dist/* --skip-existing --verbose
-
-          echo "✅ Published to PyPI!"
-
-      - name: Create release
-        run: |
-          # Check if tag already exists
-          if git rev-parse "v${{ inputs.version }}" >/dev/null 2>&1; then
-            echo "⚠️  Tag v${{ inputs.version }} already exists, skipping tag creation"
-          else
-            git tag "v${{ inputs.version }}"
-            git push origin "v${{ inputs.version }}"
-            echo "✅ Created and pushed tag v${{ inputs.version }}"
-          fi
-
-          # Check if release already exists
-          if gh release view "v${{ inputs.version }}" >/dev/null 2>&1; then
-            echo "⚠️  Release v${{ inputs.version }} already exists, skipping release creation"
-          else
-            gh release create "v${{ inputs.version }}" \
-              --title "Release v${{ inputs.version }}" \
-              --notes "🚀 Released to PyPI: https://pypi.org/project/leann/${{ inputs.version }}/" \
-              --latest
-            echo "✅ Created GitHub release v${{ inputs.version }}"
-          fi
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.gitignore

@@ -9,20 +9,17 @@ demo/indices/
 outputs/
 *.pkl
 *.pdf
-*.idx
-*.map
 .history/
+scripts/
 lm_eval.egg-info/
 demo/experiment_results/**/*.json
 *.jsonl
 *.eml
 *.emlx
 *.json
-!.vscode/*.json
 *.sh
 *.txt
 !CMakeLists.txt
-!llms.txt
 latency_breakdown*.json
 experiment_results/eval_results/diskann/*.json
 aws/
@@ -36,15 +33,7 @@ build/
 nprobe_logs/
 micro/results
 micro/contriever-INT8
-data/*
+examples/data/
-!data/2501.14312v1 (1).pdf
-!data/2506.08276v1.pdf
-!data/PrideandPrejudice.txt
-!data/huawei_pangu.md
-!data/ground_truth/
-!data/indices/
-!data/queries/
-!data/.gitattributes
 *.qdstrm
 benchmark_results/
 results/
@@ -89,15 +78,4 @@ test_*.py
 packages/leann-backend-diskann/third_party/DiskANN/_deps/
 
 *.meta.json
-*.passages.json
+*.passages.json
-
-batchtest.py
-tests/__pytest_cache__/
-tests/__pycache__/
-paru-bin/
-
-CLAUDE.md
-CLAUDE.local.md
-.claude/*.local.*
-.claude/local/*
-benchmarks/data/

.gitmodules

@@ -1,9 +1,9 @@
 [submodule "packages/leann-backend-diskann/third_party/DiskANN"]
 	path = packages/leann-backend-diskann/third_party/DiskANN
-	url = https://github.com/yichuan-w/DiskANN.git
+	url = https://github.com/yichuan520030910320/DiskANN.git
 [submodule "packages/leann-backend-hnsw/third_party/faiss"]
 	path = packages/leann-backend-hnsw/third_party/faiss
-	url = https://github.com/yichuan-w/faiss.git
+	url = https://github.com/yichuan520030910320/faiss.git
 [submodule "packages/leann-backend-hnsw/third_party/msgpack-c"]
 	path = packages/leann-backend-hnsw/third_party/msgpack-c
 	url = https://github.com/msgpack/msgpack-c.git
@@ -14,6 +14,3 @@
 [submodule "packages/leann-backend-hnsw/third_party/libzmq"]
 	path = packages/leann-backend-hnsw/third_party/libzmq
 	url = https://github.com/zeromq/libzmq.git
-[submodule "packages/astchunk-leann"]
-	path = packages/astchunk-leann
-	url = https://github.com/yichuan-w/astchunk-leann.git

.pre-commit-config.yaml

@@ -1,17 +0,0 @@
-repos:
-  - repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v5.0.0
-    hooks:
-      - id: trailing-whitespace
-      - id: end-of-file-fixer
-      - id: check-yaml
-      - id: check-added-large-files
-      - id: check-merge-conflict
-      - id: debug-statements
-
-  - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.12.7  # Fixed version to match pyproject.toml
-    hooks:
-      - id: ruff
-        args: [--fix, --exit-non-zero-on-fix]
-      - id: ruff-format

.vscode/extensions.json

@@ -1,5 +1,9 @@
 {
     "recommendations": [
-        "charliermarsh.ruff",
+        "llvm-vs-code-extensions.vscode-clangd",
+        "ms-python.python",
+        "ms-vscode.cmake-tools",
+        "vadimcn.vscode-lldb",
+        "eamodio.gitlens",
     ]
-}
+}

.vscode/launch.json

@@ -0,0 +1,283 @@
+{
+    // Use IntelliSense to learn about possible attributes.
+    // Hover to view descriptions of existing attributes.
+    // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
+    "version": "0.2.0",
+    "configurations": [
+        // new emdedder
+        {
+            "name": "New Embedder",
+            "type": "debugpy",
+            "request": "launch",
+            "program": "demo/main.py",
+            "console": "integratedTerminal",
+            "args": [
+                "--search",
+                "--use-original",
+                "--domain",
+                "dpr",
+                "--nprobe",
+                "5000",
+                "--load",
+                "flat",
+                "--embedder",
+                "intfloat/multilingual-e5-small"
+            ]
+        }
+        //python /home/ubuntu/Power-RAG/faiss/demo/simple_build.py
+        {
+            "name": "main.py",
+            "type": "debugpy",
+            "request": "launch",
+            "program": "demo/main.py",
+            "console": "integratedTerminal",
+            "cwd": "${workspaceFolder}",
+            "args": [
+                "--query",
+                "1000",
+                "--load",
+                "bm25"
+            ]
+        },
+        {
+            "name": "Simple Build",
+            "type": "lldb",
+            "request": "launch",
+            "program": "${workspaceFolder}/.venv/bin/python",
+            "console": "integratedTerminal",
+            "cwd": "${workspaceFolder}",
+            "args": [
+                "faiss/demo/simple_build.py"
+            ],
+            "env": {
+                "LD_PRELOAD": "/lib/x86_64-linux-gnu/libmkl_core.so:/lib/x86_64-linux-gnu/libmkl_intel_thread.so:/lib/x86_64-linux-gnu/libmkl_intel_lp64.so:/lib/x86_64-linux-gnu/libiomp5.so"
+            }
+        },
+        //# Fix for Intel MKL error
+        //export LD_PRELOAD=/lib/x86_64-linux-gnu/libmkl_core.so:/lib/x86_64-linux-gnu/libmkl_intel_thread.so:/lib/x86_64-linux-gnu/libmkl_intel_lp64.so:/lib/x86_64-linux-gnu/libiomp5.so
+        //python faiss/demo/build_demo.py
+        {
+            "name": "Build Demo",
+            "type": "lldb",
+            "request": "launch",
+            "program": "${workspaceFolder}/.venv/bin/python",
+            "console": "integratedTerminal",
+            "cwd": "${workspaceFolder}",
+            "args": [
+                "faiss/demo/build_demo.py"
+            ],
+            "env": {
+                "LD_PRELOAD": "/lib/x86_64-linux-gnu/libmkl_core.so:/lib/x86_64-linux-gnu/libmkl_intel_thread.so:/lib/x86_64-linux-gnu/libmkl_intel_lp64.so:/lib/x86_64-linux-gnu/libiomp5.so"
+            }
+        },
+        {
+            "name": "DiskANN Serve",
+            "type": "lldb",
+            "request": "launch",
+            "program": "${workspaceFolder}/.venv/bin/python",
+            "console": "integratedTerminal",
+            "cwd": "${workspaceFolder}",
+            "args": [
+                "demo/main.py",
+                "--mode",
+                "serve",
+                "--engine",
+                "sglang",
+                "--load-indices",
+                "diskann",
+                "--domain",
+                "rpj_wiki",
+                "--lazy-load",
+                "--recompute-beighbor-embeddings",
+                "--port",
+                "8082",
+                "--diskann-search-memory-maximum",
+                "2",
+                "--diskann-graph",
+                "240",
+                "--search-only"
+            ],
+            "env": {
+                "PYTHONPATH": "${workspaceFolder}/faiss_repo/build/faiss/python:$PYTHONPATH"
+            },
+            "preLaunchTask": "CMake: build",
+        },
+        {
+            "name": "DiskANN Serve MAC",
+            "type": "lldb",
+            "request": "launch",
+            "program": "${workspaceFolder}/.venv/bin/python",
+            "console": "integratedTerminal",
+            "cwd": "${workspaceFolder}",
+            "args": [
+                "demo/main.py",
+                "--mode",
+                "serve",
+                "--engine",
+                "ollama",
+                "--load-indices",
+                "diskann",
+                "--domain",
+                "rpj_wiki",
+                "--lazy-load",
+                "--recompute-beighbor-embeddings"
+            ],
+            "preLaunchTask": "CMake: build",
+            "env": {
+                "KMP_DUPLICATE_LIB_OK": "TRUE",
+                "OMP_NUM_THREADS": "1",
+                "MKL_NUM_THREADS": "1",
+                "DYLD_INSERT_LIBRARIES": "/Users/ec2-user/Power-RAG/.venv/lib/python3.10/site-packages/torch/lib/libomp.dylib",
+                "KMP_BLOCKTIME": "0"
+            }
+        },
+        {
+            "name": "Python Debugger: Current File with Arguments",
+            "type": "debugpy",
+            "request": "launch",
+            "program": "ric/main_ric.py",
+            "console": "integratedTerminal",
+            "cwd": "${workspaceFolder}",
+            "args": [
+                "--config-name",
+                "${input:configSelection}"
+            ],
+            "justMyCode": false
+        },
+        //python ./demo/validate_equivalence.py sglang
+        {
+            "name": "Validate Equivalence",
+            "type": "debugpy",
+            "request": "launch",
+            "program": "demo/validate_equivalence.py",
+            "console": "integratedTerminal",
+            "args": [
+                "sglang"
+            ],
+        },
+        //python demo/retrieval_demo.py --engine sglang  --skip-embeddings --domain dpr --load-indices flat ivf_flat
+        {
+            "name": "Retrieval Demo",
+            "type": "debugpy",
+            "request": "launch",
+            "program": "demo/retrieval_demo.py",
+            "console": "integratedTerminal",
+            "args": [
+                "--engine",
+                "vllm",
+                "--skip-embeddings",
+                "--domain",
+                "dpr",
+                "--load-indices",
+                // "flat",
+                "ivf_flat"
+            ],
+        },
+        //python demo/retrieval_demo.py --engine sglang  --skip-embeddings --domain dpr  --load-indices  diskann --hnsw-M 64 --hnsw-efConstruction 150 --hnsw-efSearch 128  --hnsw-sq-bits 8 
+        {
+            "name": "Retrieval Demo DiskANN",
+            "type": "debugpy",
+            "request": "launch",
+            "program": "demo/retrieval_demo.py",
+            "console": "integratedTerminal",
+            "args": [
+                "--engine",
+                "sglang",
+                "--skip-embeddings",
+                "--domain",
+                "dpr",
+                "--load-indices",
+                "diskann",
+                "--hnsw-M",
+                "64",
+                "--hnsw-efConstruction",
+                "150",
+                "--hnsw-efSearch",
+                "128",
+                "--hnsw-sq-bits",
+                "8"
+            ],
+        },
+        {
+            "name": "Find Probe",
+            "type": "debugpy",
+            "request": "launch",
+            "program": "find_probe.py",
+            "console": "integratedTerminal",
+            "cwd": "${workspaceFolder}",
+        },
+        {
+            "name": "Python: Attach",
+            "type": "debugpy",
+            "request": "attach",
+            "processId": "${command:pickProcess}",
+            "justMyCode": true
+        },
+        {
+            "name": "Edge RAG",
+            "type": "lldb",
+            "request": "launch",
+            "program": "${workspaceFolder}/.venv/bin/python",
+            "console": "integratedTerminal",
+            "cwd": "${workspaceFolder}",
+            "args": [
+                "edgerag_demo.py"
+            ],
+            "env": {
+                "LD_PRELOAD": "/lib/x86_64-linux-gnu/libiomp5.so /lib/x86_64-linux-gnu/libmkl_core.so /lib/x86_64-linux-gnu/libmkl_intel_lp64.so /lib/x86_64-linux-gnu/libmkl_intel_thread.so",
+                "MKL_NUM_THREADS": "1",
+                "OMP_NUM_THREADS": "1",
+            }
+        },
+        {
+            "name": "Launch Embedding Server",
+            "type": "debugpy",
+            "request": "launch",
+            "program": "demo/embedding_server.py",
+            "console": "integratedTerminal",
+            "cwd": "${workspaceFolder}",
+            "args": [
+                "--domain",
+                "rpj_wiki",
+                "--zmq-port",
+                "5556",
+            ]
+        },
+        {
+            "name": "HNSW Serve",
+            "type": "lldb",
+            "request": "launch",
+            "program": "${workspaceFolder}/.venv/bin/python",
+            "console": "integratedTerminal",
+            "cwd": "${workspaceFolder}",
+            "args": [
+                "demo/main.py",
+                "--domain",
+                "rpj_wiki",
+                "--load",
+                "hnsw",
+                "--mode",
+                "serve",
+                "--search",
+                "--skip-pa",
+                "--recompute",
+                "--hnsw-old"
+            ],
+            "env": {
+                "LD_PRELOAD": "/lib/x86_64-linux-gnu/libmkl_core.so:/lib/x86_64-linux-gnu/libmkl_intel_thread.so:/lib/x86_64-linux-gnu/libmkl_intel_lp64.so:/lib/x86_64-linux-gnu/libiomp5.so"
+            }
+        },
+    ],
+    "inputs": [
+        {
+            "id": "configSelection",
+            "type": "pickString",
+            "description": "Select a configuration",
+            "options": [
+                "example_config",
+                "vllm_gritlm"
+            ],
+            "default": "example_config"
+        }
+    ],
+}

.vscode/settings.json

@@ -1,22 +1,43 @@
 {
-  "python.defaultInterpreterPath": ".venv/bin/python",
+    "python.analysis.extraPaths": [
-  "python.terminal.activateEnvironment": true,
+        "./sglang_repo/python"
-  "[python]": {
+    ],
-    "editor.defaultFormatter": "charliermarsh.ruff",
+    "cmake.sourceDirectory": "${workspaceFolder}/DiskANN",
-    "editor.formatOnSave": true,
+    "cmake.configureArgs": [
-    "editor.codeActionsOnSave": {
+        "-DPYBIND=True",
-      "source.organizeImports": "explicit",
+        "-DUPDATE_EDITABLE_INSTALL=ON",
-      "source.fixAll": "explicit"
+    ],
+    "cmake.environment": {
+        "PATH": "/Users/ec2-user/Power-RAG/.venv/bin:${env:PATH}"
     },
-    "editor.insertSpaces": true,
+    "cmake.buildDirectory": "${workspaceFolder}/build",
-    "editor.tabSize": 4
+    "files.associations": {
-  },
+        "*.tcc": "cpp",
-  "ruff.enable": true,
+        "deque": "cpp",
-  "files.watcherExclude": {
+        "string": "cpp",
-    "**/.venv/**": true,
+        "unordered_map": "cpp",
-    "**/__pycache__/**": true,
+        "vector": "cpp",
-    "**/*.egg-info/**": true,
+        "map": "cpp",
-    "**/build/**": true,
+        "unordered_set": "cpp",
-    "**/dist/**": true
+        "atomic": "cpp",
-  }
+        "inplace_vector": "cpp",
-}
+        "*.ipp": "cpp",
+        "forward_list": "cpp",
+        "list": "cpp",
+        "any": "cpp",
+        "system_error": "cpp",
+        "__hash_table": "cpp",
+        "__split_buffer": "cpp",
+        "__tree": "cpp",
+        "ios": "cpp",
+        "set": "cpp",
+        "__string": "cpp",
+        "string_view": "cpp",
+        "ranges": "cpp",
+        "iosfwd": "cpp"
+    },
+    "lldb.displayFormat": "auto",
+    "lldb.showDisassembly": "auto",
+    "lldb.dereferencePointers": true,
+    "lldb.consoleMode": "commands",
+}

.vscode/tasks.json

@@ -0,0 +1,16 @@
+{
+	"version": "2.0.0",
+	"tasks": [
+		{
+			"type": "cmake",
+			"label": "CMake: build",
+			"command": "build",
+			"targets": [
+				"all"
+			],
+			"group": "build",
+			"problemMatcher": [],
+			"detail": "CMake template build task"
+		}
+	]
+}

LICENSE

@@ -1,6 +1,6 @@
 MIT License
 
-Copyright (c) 2025 LEANN Contributors
+Copyright (c) 2024 Rulin Shao
 
 Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
 

apps/base_rag_example.py

@@ -1,387 +0,0 @@
-"""
-Base class for unified RAG examples interface.
-Provides common parameters and functionality for all RAG examples.
-"""
-
-import argparse
-from abc import ABC, abstractmethod
-from pathlib import Path
-from typing import Any
-
-import dotenv
-from leann.api import LeannBuilder, LeannChat
-from leann.registry import register_project_directory
-from leann.settings import resolve_ollama_host, resolve_openai_api_key, resolve_openai_base_url
-
-dotenv.load_dotenv()
-
-
-class BaseRAGExample(ABC):
-    """Base class for all RAG examples with unified interface."""
-
-    def __init__(
-        self,
-        name: str,
-        description: str,
-        default_index_name: str,
-    ):
-        self.name = name
-        self.description = description
-        self.default_index_name = default_index_name
-        self.parser = self._create_parser()
-
-    def _create_parser(self) -> argparse.ArgumentParser:
-        """Create argument parser with common parameters."""
-        parser = argparse.ArgumentParser(
-            description=self.description, formatter_class=argparse.RawDescriptionHelpFormatter
-        )
-
-        # Core parameters (all examples share these)
-        core_group = parser.add_argument_group("Core Parameters")
-        core_group.add_argument(
-            "--index-dir",
-            type=str,
-            default=f"./{self.default_index_name}",
-            help=f"Directory to store the index (default: ./{self.default_index_name})",
-        )
-        core_group.add_argument(
-            "--query",
-            type=str,
-            default=None,
-            help="Query to run (if not provided, will run in interactive mode)",
-        )
-        # Allow subclasses to override default max_items
-        max_items_default = getattr(self, "max_items_default", -1)
-        core_group.add_argument(
-            "--max-items",
-            type=int,
-            default=max_items_default,
-            help="Maximum number of items to process  -1 for all, means index all documents, and you should set it to a reasonable number if you have a large dataset and try at the first time)",
-        )
-        core_group.add_argument(
-            "--force-rebuild", action="store_true", help="Force rebuild index even if it exists"
-        )
-
-        # Embedding parameters
-        embedding_group = parser.add_argument_group("Embedding Parameters")
-        # Allow subclasses to override default embedding_model
-        embedding_model_default = getattr(self, "embedding_model_default", "facebook/contriever")
-        embedding_group.add_argument(
-            "--embedding-model",
-            type=str,
-            default=embedding_model_default,
-            help=f"Embedding model to use (default: {embedding_model_default}), we provide facebook/contriever, text-embedding-3-small,mlx-community/Qwen3-Embedding-0.6B-8bit or nomic-embed-text",
-        )
-        embedding_group.add_argument(
-            "--embedding-mode",
-            type=str,
-            default="sentence-transformers",
-            choices=["sentence-transformers", "openai", "mlx", "ollama"],
-            help="Embedding backend mode (default: sentence-transformers), we provide sentence-transformers, openai, mlx, or ollama",
-        )
-        embedding_group.add_argument(
-            "--embedding-host",
-            type=str,
-            default=None,
-            help="Override Ollama-compatible embedding host",
-        )
-        embedding_group.add_argument(
-            "--embedding-api-base",
-            type=str,
-            default=None,
-            help="Base URL for OpenAI-compatible embedding services",
-        )
-        embedding_group.add_argument(
-            "--embedding-api-key",
-            type=str,
-            default=None,
-            help="API key for embedding service (defaults to OPENAI_API_KEY)",
-        )
-
-        # LLM parameters
-        llm_group = parser.add_argument_group("LLM Parameters")
-        llm_group.add_argument(
-            "--llm",
-            type=str,
-            default="openai",
-            choices=["openai", "ollama", "hf", "simulated"],
-            help="LLM backend: openai, ollama, or hf (default: openai)",
-        )
-        llm_group.add_argument(
-            "--llm-model",
-            type=str,
-            default=None,
-            help="Model name (default: gpt-4o) e.g., gpt-4o-mini, llama3.2:1b, Qwen/Qwen2.5-1.5B-Instruct",
-        )
-        llm_group.add_argument(
-            "--llm-host",
-            type=str,
-            default=None,
-            help="Host for Ollama-compatible APIs (defaults to LEANN_OLLAMA_HOST/OLLAMA_HOST)",
-        )
-        llm_group.add_argument(
-            "--thinking-budget",
-            type=str,
-            choices=["low", "medium", "high"],
-            default=None,
-            help="Thinking budget for reasoning models (low/medium/high). Supported by GPT-Oss:20b and other reasoning models.",
-        )
-        llm_group.add_argument(
-            "--llm-api-base",
-            type=str,
-            default=None,
-            help="Base URL for OpenAI-compatible APIs",
-        )
-        llm_group.add_argument(
-            "--llm-api-key",
-            type=str,
-            default=None,
-            help="API key for OpenAI-compatible APIs (defaults to OPENAI_API_KEY)",
-        )
-
-        # AST Chunking parameters
-        ast_group = parser.add_argument_group("AST Chunking Parameters")
-        ast_group.add_argument(
-            "--use-ast-chunking",
-            action="store_true",
-            help="Enable AST-aware chunking for code files (requires astchunk)",
-        )
-        ast_group.add_argument(
-            "--ast-chunk-size",
-            type=int,
-            default=512,
-            help="Maximum characters per AST chunk (default: 512)",
-        )
-        ast_group.add_argument(
-            "--ast-chunk-overlap",
-            type=int,
-            default=64,
-            help="Overlap between AST chunks (default: 64)",
-        )
-        ast_group.add_argument(
-            "--code-file-extensions",
-            nargs="+",
-            default=None,
-            help="Additional code file extensions to process with AST chunking (e.g., .py .java .cs .ts)",
-        )
-        ast_group.add_argument(
-            "--ast-fallback-traditional",
-            action="store_true",
-            default=True,
-            help="Fall back to traditional chunking if AST chunking fails (default: True)",
-        )
-
-        # Search parameters
-        search_group = parser.add_argument_group("Search Parameters")
-        search_group.add_argument(
-            "--top-k", type=int, default=20, help="Number of results to retrieve (default: 20)"
-        )
-        search_group.add_argument(
-            "--search-complexity",
-            type=int,
-            default=32,
-            help="Search complexity for graph traversal (default: 64)",
-        )
-
-        # Index building parameters
-        index_group = parser.add_argument_group("Index Building Parameters")
-        index_group.add_argument(
-            "--backend-name",
-            type=str,
-            default="hnsw",
-            choices=["hnsw", "diskann"],
-            help="Backend to use for index (default: hnsw)",
-        )
-        index_group.add_argument(
-            "--graph-degree",
-            type=int,
-            default=32,
-            help="Graph degree for index construction (default: 32)",
-        )
-        index_group.add_argument(
-            "--build-complexity",
-            type=int,
-            default=64,
-            help="Build complexity for index construction (default: 64)",
-        )
-        index_group.add_argument(
-            "--no-compact",
-            action="store_true",
-            help="Disable compact index storage",
-        )
-        index_group.add_argument(
-            "--no-recompute",
-            action="store_true",
-            help="Disable embedding recomputation",
-        )
-
-        # Add source-specific parameters
-        self._add_specific_arguments(parser)
-
-        return parser
-
-    @abstractmethod
-    def _add_specific_arguments(self, parser: argparse.ArgumentParser):
-        """Add source-specific arguments. Override in subclasses."""
-        pass
-
-    @abstractmethod
-    async def load_data(self, args) -> list[str]:
-        """Load data from the source. Returns list of text chunks."""
-        pass
-
-    def get_llm_config(self, args) -> dict[str, Any]:
-        """Get LLM configuration based on arguments."""
-        config = {"type": args.llm}
-
-        if args.llm == "openai":
-            config["model"] = args.llm_model or "gpt-4o"
-            config["base_url"] = resolve_openai_base_url(args.llm_api_base)
-            resolved_key = resolve_openai_api_key(args.llm_api_key)
-            if resolved_key:
-                config["api_key"] = resolved_key
-        elif args.llm == "ollama":
-            config["model"] = args.llm_model or "llama3.2:1b"
-            config["host"] = resolve_ollama_host(args.llm_host)
-        elif args.llm == "hf":
-            config["model"] = args.llm_model or "Qwen/Qwen2.5-1.5B-Instruct"
-        elif args.llm == "simulated":
-            # Simulated LLM doesn't need additional configuration
-            pass
-
-        return config
-
-    async def build_index(self, args, texts: list[str]) -> str:
-        """Build LEANN index from texts."""
-        index_path = str(Path(args.index_dir) / f"{self.default_index_name}.leann")
-
-        print(f"\n[Building Index] Creating {self.name} index...")
-        print(f"Total text chunks: {len(texts)}")
-
-        embedding_options: dict[str, Any] = {}
-        if args.embedding_mode == "ollama":
-            embedding_options["host"] = resolve_ollama_host(args.embedding_host)
-        elif args.embedding_mode == "openai":
-            embedding_options["base_url"] = resolve_openai_base_url(args.embedding_api_base)
-            resolved_embedding_key = resolve_openai_api_key(args.embedding_api_key)
-            if resolved_embedding_key:
-                embedding_options["api_key"] = resolved_embedding_key
-
-        builder = LeannBuilder(
-            backend_name=args.backend_name,
-            embedding_model=args.embedding_model,
-            embedding_mode=args.embedding_mode,
-            embedding_options=embedding_options or None,
-            graph_degree=args.graph_degree,
-            complexity=args.build_complexity,
-            is_compact=not args.no_compact,
-            is_recompute=not args.no_recompute,
-            num_threads=1,  # Force single-threaded mode
-        )
-
-        # Add texts in batches for better progress tracking
-        batch_size = 1000
-        for i in range(0, len(texts), batch_size):
-            batch = texts[i : i + batch_size]
-            for text in batch:
-                builder.add_text(text)
-            print(f"Added {min(i + batch_size, len(texts))}/{len(texts)} texts...")
-
-        print("Building index structure...")
-        builder.build_index(index_path)
-        print(f"Index saved to: {index_path}")
-
-        # Register project directory so leann list can discover this index
-        # The index is saved as args.index_dir/index_name.leann
-        # We want to register the current working directory where the app is run
-        register_project_directory(Path.cwd())
-
-        return index_path
-
-    async def run_interactive_chat(self, args, index_path: str):
-        """Run interactive chat with the index."""
-        chat = LeannChat(
-            index_path,
-            llm_config=self.get_llm_config(args),
-            system_prompt=f"You are a helpful assistant that answers questions about {self.name} data.",
-            complexity=args.search_complexity,
-        )
-
-        print(f"\n[Interactive Mode] Chat with your {self.name} data!")
-        print("Type 'quit' or 'exit' to stop.\n")
-
-        while True:
-            try:
-                query = input("You: ").strip()
-                if query.lower() in ["quit", "exit", "q"]:
-                    print("Goodbye!")
-                    break
-
-                if not query:
-                    continue
-
-                # Prepare LLM kwargs with thinking budget if specified
-                llm_kwargs = {}
-                if hasattr(args, "thinking_budget") and args.thinking_budget:
-                    llm_kwargs["thinking_budget"] = args.thinking_budget
-
-                response = chat.ask(
-                    query,
-                    top_k=args.top_k,
-                    complexity=args.search_complexity,
-                    llm_kwargs=llm_kwargs,
-                )
-                print(f"\nAssistant: {response}\n")
-
-            except KeyboardInterrupt:
-                print("\nGoodbye!")
-                break
-            except Exception as e:
-                print(f"Error: {e}")
-
-    async def run_single_query(self, args, index_path: str, query: str):
-        """Run a single query against the index."""
-        chat = LeannChat(
-            index_path,
-            llm_config=self.get_llm_config(args),
-            complexity=args.search_complexity,
-        )
-
-        print(f"\n[Query]: \033[36m{query}\033[0m")
-
-        # Prepare LLM kwargs with thinking budget if specified
-        llm_kwargs = {}
-        if hasattr(args, "thinking_budget") and args.thinking_budget:
-            llm_kwargs["thinking_budget"] = args.thinking_budget
-
-        response = chat.ask(
-            query, top_k=args.top_k, complexity=args.search_complexity, llm_kwargs=llm_kwargs
-        )
-        print(f"\n[Response]: \033[36m{response}\033[0m")
-
-    async def run(self):
-        """Main entry point for the example."""
-        args = self.parser.parse_args()
-
-        # Check if index exists
-        index_path = str(Path(args.index_dir) / f"{self.default_index_name}.leann")
-        index_exists = Path(args.index_dir).exists()
-
-        if not index_exists or args.force_rebuild:
-            # Load data and build index
-            print(f"\n{'Rebuilding' if index_exists else 'Building'} index...")
-            texts = await self.load_data(args)
-
-            if not texts:
-                print("No data found to index!")
-                return
-
-            index_path = await self.build_index(args, texts)
-        else:
-            print(f"\nUsing existing index in {args.index_dir}")
-
-        # Run query or interactive mode
-        if args.query:
-            await self.run_single_query(args, index_path, args.query)
-        else:
-            await self.run_interactive_chat(args, index_path)

apps/browser_rag.py

@@ -1,171 +0,0 @@
-"""
-Browser History RAG example using the unified interface.
-Supports Chrome browser history.
-"""
-
-import os
-import sys
-from pathlib import Path
-
-# Add parent directory to path for imports
-sys.path.insert(0, str(Path(__file__).parent))
-
-from base_rag_example import BaseRAGExample
-from chunking import create_text_chunks
-
-from .history_data.history import ChromeHistoryReader
-
-
-class BrowserRAG(BaseRAGExample):
-    """RAG example for Chrome browser history."""
-
-    def __init__(self):
-        # Set default values BEFORE calling super().__init__
-        self.embedding_model_default = (
-            "sentence-transformers/all-MiniLM-L6-v2"  # Fast 384-dim model
-        )
-
-        super().__init__(
-            name="Browser History",
-            description="Process and query Chrome browser history with LEANN",
-            default_index_name="google_history_index",
-        )
-
-    def _add_specific_arguments(self, parser):
-        """Add browser-specific arguments."""
-        browser_group = parser.add_argument_group("Browser Parameters")
-        browser_group.add_argument(
-            "--chrome-profile",
-            type=str,
-            default=None,
-            help="Path to Chrome profile directory (auto-detected if not specified)",
-        )
-        browser_group.add_argument(
-            "--auto-find-profiles",
-            action="store_true",
-            default=True,
-            help="Automatically find all Chrome profiles (default: True)",
-        )
-        browser_group.add_argument(
-            "--chunk-size", type=int, default=256, help="Text chunk size (default: 256)"
-        )
-        browser_group.add_argument(
-            "--chunk-overlap", type=int, default=128, help="Text chunk overlap (default: 128)"
-        )
-
-    def _get_chrome_base_path(self) -> Path:
-        """Get the base Chrome profile path based on OS."""
-        if sys.platform == "darwin":
-            return Path.home() / "Library" / "Application Support" / "Google" / "Chrome"
-        elif sys.platform.startswith("linux"):
-            return Path.home() / ".config" / "google-chrome"
-        elif sys.platform == "win32":
-            return Path(os.environ["LOCALAPPDATA"]) / "Google" / "Chrome" / "User Data"
-        else:
-            raise ValueError(f"Unsupported platform: {sys.platform}")
-
-    def _find_chrome_profiles(self) -> list[Path]:
-        """Auto-detect all Chrome profiles."""
-        base_path = self._get_chrome_base_path()
-        if not base_path.exists():
-            return []
-
-        profiles = []
-
-        # Check Default profile
-        default_profile = base_path / "Default"
-        if default_profile.exists() and (default_profile / "History").exists():
-            profiles.append(default_profile)
-
-        # Check numbered profiles
-        for item in base_path.iterdir():
-            if item.is_dir() and item.name.startswith("Profile "):
-                if (item / "History").exists():
-                    profiles.append(item)
-
-        return profiles
-
-    async def load_data(self, args) -> list[str]:
-        """Load browser history and convert to text chunks."""
-        # Determine Chrome profiles
-        if args.chrome_profile and not args.auto_find_profiles:
-            profile_dirs = [Path(args.chrome_profile)]
-        else:
-            print("Auto-detecting Chrome profiles...")
-            profile_dirs = self._find_chrome_profiles()
-
-            # If specific profile given, filter to just that one
-            if args.chrome_profile:
-                profile_path = Path(args.chrome_profile)
-                profile_dirs = [p for p in profile_dirs if p == profile_path]
-
-        if not profile_dirs:
-            print("No Chrome profiles found!")
-            print("Please specify --chrome-profile manually")
-            return []
-
-        print(f"Found {len(profile_dirs)} Chrome profiles")
-
-        # Create reader
-        reader = ChromeHistoryReader()
-
-        # Process each profile
-        all_documents = []
-        total_processed = 0
-
-        for i, profile_dir in enumerate(profile_dirs):
-            print(f"\nProcessing profile {i + 1}/{len(profile_dirs)}: {profile_dir.name}")
-
-            try:
-                # Apply max_items limit per profile
-                max_per_profile = -1
-                if args.max_items > 0:
-                    remaining = args.max_items - total_processed
-                    if remaining <= 0:
-                        break
-                    max_per_profile = remaining
-
-                # Load history
-                documents = reader.load_data(
-                    chrome_profile_path=str(profile_dir),
-                    max_count=max_per_profile,
-                )
-
-                if documents:
-                    all_documents.extend(documents)
-                    total_processed += len(documents)
-                    print(f"Processed {len(documents)} history entries from this profile")
-
-            except Exception as e:
-                print(f"Error processing {profile_dir}: {e}")
-                continue
-
-        if not all_documents:
-            print("No browser history found to process!")
-            return []
-
-        print(f"\nTotal history entries processed: {len(all_documents)}")
-
-        # Convert to text chunks
-        all_texts = create_text_chunks(
-            all_documents, chunk_size=args.chunk_size, chunk_overlap=args.chunk_overlap
-        )
-
-        return all_texts
-
-
-if __name__ == "__main__":
-    import asyncio
-
-    # Example queries for browser history RAG
-    print("\n🌐 Browser History RAG Example")
-    print("=" * 50)
-    print("\nExample queries you can try:")
-    print("- 'What websites did I visit about machine learning?'")
-    print("- 'Find my search history about programming'")
-    print("- 'What YouTube videos did I watch recently?'")
-    print("- 'Show me websites about travel planning'")
-    print("\nNote: Make sure Chrome is closed before running\n")
-
-    rag = BrowserRAG()
-    asyncio.run(rag.run())

apps/chunking/__init__.py

@@ -1,44 +0,0 @@
-"""Unified chunking utilities facade.
-
-This module re-exports the packaged utilities from `leann.chunking_utils` so
-that both repo apps (importing `chunking`) and installed wheels share one
-single implementation. When running from the repo without installation, it
-adds the `packages/leann-core/src` directory to `sys.path` as a fallback.
-"""
-
-import sys
-from pathlib import Path
-
-try:
-    from leann.chunking_utils import (
-        CODE_EXTENSIONS,
-        create_ast_chunks,
-        create_text_chunks,
-        create_traditional_chunks,
-        detect_code_files,
-        get_language_from_extension,
-    )
-except Exception:  # pragma: no cover - best-effort fallback for dev environment
-    repo_root = Path(__file__).resolve().parents[2]
-    leann_src = repo_root / "packages" / "leann-core" / "src"
-    if leann_src.exists():
-        sys.path.insert(0, str(leann_src))
-        from leann.chunking_utils import (
-            CODE_EXTENSIONS,
-            create_ast_chunks,
-            create_text_chunks,
-            create_traditional_chunks,
-            detect_code_files,
-            get_language_from_extension,
-        )
-    else:
-        raise
-
-__all__ = [
-    "CODE_EXTENSIONS",
-    "create_ast_chunks",
-    "create_text_chunks",
-    "create_traditional_chunks",
-    "detect_code_files",
-    "get_language_from_extension",
-]

apps/code_rag.py

@@ -1,211 +0,0 @@
-"""
-Code RAG example using AST-aware chunking for optimal code understanding.
-Specialized for code repositories with automatic language detection and
-optimized chunking parameters.
-"""
-
-import sys
-from pathlib import Path
-
-# Add parent directory to path for imports
-sys.path.insert(0, str(Path(__file__).parent))
-
-from base_rag_example import BaseRAGExample
-from chunking import CODE_EXTENSIONS, create_text_chunks
-from llama_index.core import SimpleDirectoryReader
-
-
-class CodeRAG(BaseRAGExample):
-    """Specialized RAG example for code repositories with AST-aware chunking."""
-
-    def __init__(self):
-        super().__init__(
-            name="Code",
-            description="Process and query code repositories with AST-aware chunking",
-            default_index_name="code_index",
-        )
-        # Override defaults for code-specific usage
-        self.embedding_model_default = "facebook/contriever"  # Good for code
-        self.max_items_default = -1  # Process all code files by default
-
-    def _add_specific_arguments(self, parser):
-        """Add code-specific arguments."""
-        code_group = parser.add_argument_group("Code Repository Parameters")
-
-        code_group.add_argument(
-            "--repo-dir",
-            type=str,
-            default=".",
-            help="Code repository directory to index (default: current directory)",
-        )
-        code_group.add_argument(
-            "--include-extensions",
-            nargs="+",
-            default=list(CODE_EXTENSIONS.keys()),
-            help="File extensions to include (default: supported code extensions)",
-        )
-        code_group.add_argument(
-            "--exclude-dirs",
-            nargs="+",
-            default=[
-                ".git",
-                "__pycache__",
-                "node_modules",
-                "venv",
-                ".venv",
-                "build",
-                "dist",
-                "target",
-            ],
-            help="Directories to exclude from indexing",
-        )
-        code_group.add_argument(
-            "--max-file-size",
-            type=int,
-            default=1000000,  # 1MB
-            help="Maximum file size in bytes to process (default: 1MB)",
-        )
-        code_group.add_argument(
-            "--include-comments",
-            action="store_true",
-            help="Include comments in chunking (useful for documentation)",
-        )
-        code_group.add_argument(
-            "--preserve-imports",
-            action="store_true",
-            default=True,
-            help="Try to preserve import statements in chunks (default: True)",
-        )
-
-    async def load_data(self, args) -> list[str]:
-        """Load code files and convert to AST-aware chunks."""
-        print(f"🔍 Scanning code repository: {args.repo_dir}")
-        print(f"📁 Including extensions: {args.include_extensions}")
-        print(f"🚫 Excluding directories: {args.exclude_dirs}")
-
-        # Check if repository directory exists
-        repo_path = Path(args.repo_dir)
-        if not repo_path.exists():
-            raise ValueError(f"Repository directory not found: {args.repo_dir}")
-
-        # Load code files with filtering
-        reader_kwargs = {
-            "recursive": True,
-            "encoding": "utf-8",
-            "required_exts": args.include_extensions,
-            "exclude_hidden": True,
-        }
-
-        # Create exclusion filter
-        def file_filter(file_path: str) -> bool:
-            """Filter out unwanted files and directories."""
-            path = Path(file_path)
-
-            # Check file size
-            try:
-                if path.stat().st_size > args.max_file_size:
-                    print(f"⚠️ Skipping large file: {path.name} ({path.stat().st_size} bytes)")
-                    return False
-            except Exception:
-                return False
-
-            # Check if in excluded directory
-            for exclude_dir in args.exclude_dirs:
-                if exclude_dir in path.parts:
-                    return False
-
-            return True
-
-        try:
-            # Load documents with file filtering
-            documents = SimpleDirectoryReader(
-                args.repo_dir,
-                file_extractor=None,  # Use default extractors
-                **reader_kwargs,
-            ).load_data(show_progress=True)
-
-            # Apply custom filtering
-            filtered_docs = []
-            for doc in documents:
-                file_path = doc.metadata.get("file_path", "")
-                if file_filter(file_path):
-                    filtered_docs.append(doc)
-
-            documents = filtered_docs
-
-        except Exception as e:
-            print(f"❌ Error loading code files: {e}")
-            return []
-
-        if not documents:
-            print(
-                f"❌ No code files found in {args.repo_dir} with extensions {args.include_extensions}"
-            )
-            return []
-
-        print(f"✅ Loaded {len(documents)} code files")
-
-        # Show breakdown by language/extension
-        ext_counts = {}
-        for doc in documents:
-            file_path = doc.metadata.get("file_path", "")
-            if file_path:
-                ext = Path(file_path).suffix.lower()
-                ext_counts[ext] = ext_counts.get(ext, 0) + 1
-
-        print("📊 Files by extension:")
-        for ext, count in sorted(ext_counts.items()):
-            print(f"   {ext}: {count} files")
-
-        # Use AST-aware chunking by default for code
-        print(
-            f"🧠 Using AST-aware chunking (chunk_size: {args.ast_chunk_size}, overlap: {args.ast_chunk_overlap})"
-        )
-
-        all_texts = create_text_chunks(
-            documents,
-            chunk_size=256,  # Fallback for non-code files
-            chunk_overlap=64,
-            use_ast_chunking=True,  # Always use AST for code RAG
-            ast_chunk_size=args.ast_chunk_size,
-            ast_chunk_overlap=args.ast_chunk_overlap,
-            code_file_extensions=args.include_extensions,
-            ast_fallback_traditional=True,
-        )
-
-        # Apply max_items limit if specified
-        if args.max_items > 0 and len(all_texts) > args.max_items:
-            print(f"⏳ Limiting to {args.max_items} chunks (from {len(all_texts)})")
-            all_texts = all_texts[: args.max_items]
-
-        print(f"✅ Generated {len(all_texts)} code chunks")
-        return all_texts
-
-
-if __name__ == "__main__":
-    import asyncio
-
-    # Example queries for code RAG
-    print("\n💻 Code RAG Example")
-    print("=" * 50)
-    print("\nExample queries you can try:")
-    print("- 'How does the embedding computation work?'")
-    print("- 'What are the main classes in this codebase?'")
-    print("- 'Show me the search implementation'")
-    print("- 'How is error handling implemented?'")
-    print("- 'What design patterns are used?'")
-    print("- 'Explain the chunking logic'")
-    print("\n🚀 Features:")
-    print("- ✅ AST-aware chunking preserves code structure")
-    print("- ✅ Automatic language detection")
-    print("- ✅ Smart filtering of large files and common excludes")
-    print("- ✅ Optimized for code understanding")
-    print("\nUsage examples:")
-    print("  python -m apps.code_rag --repo-dir ./my_project")
-    print(
-        "  python -m apps.code_rag --include-extensions .py .js --query 'How does authentication work?'"
-    )
-    print("\nOr run without --query for interactive mode\n")
-
-    rag = CodeRAG()
-    asyncio.run(rag.run())

apps/document_rag.py

@@ -1,131 +0,0 @@
-"""
-Document RAG example using the unified interface.
-Supports PDF, TXT, MD, and other document formats.
-"""
-
-import sys
-from pathlib import Path
-
-# Add parent directory to path for imports
-sys.path.insert(0, str(Path(__file__).parent))
-
-from base_rag_example import BaseRAGExample
-from chunking import create_text_chunks
-from llama_index.core import SimpleDirectoryReader
-
-
-class DocumentRAG(BaseRAGExample):
-    """RAG example for document processing (PDF, TXT, MD, etc.)."""
-
-    def __init__(self):
-        super().__init__(
-            name="Document",
-            description="Process and query documents (PDF, TXT, MD, etc.) with LEANN",
-            default_index_name="test_doc_files",
-        )
-
-    def _add_specific_arguments(self, parser):
-        """Add document-specific arguments."""
-        doc_group = parser.add_argument_group("Document Parameters")
-        doc_group.add_argument(
-            "--data-dir",
-            type=str,
-            default="data",
-            help="Directory containing documents to index (default: data)",
-        )
-        doc_group.add_argument(
-            "--file-types",
-            nargs="+",
-            default=None,
-            help="Filter by file types (e.g., .pdf .txt .md). If not specified, all supported types are processed",
-        )
-        doc_group.add_argument(
-            "--chunk-size", type=int, default=256, help="Text chunk size (default: 256)"
-        )
-        doc_group.add_argument(
-            "--chunk-overlap", type=int, default=128, help="Text chunk overlap (default: 128)"
-        )
-        doc_group.add_argument(
-            "--enable-code-chunking",
-            action="store_true",
-            help="Enable AST-aware chunking for code files in the data directory",
-        )
-
-    async def load_data(self, args) -> list[str]:
-        """Load documents and convert to text chunks."""
-        print(f"Loading documents from: {args.data_dir}")
-        if args.file_types:
-            print(f"Filtering by file types: {args.file_types}")
-        else:
-            print("Processing all supported file types")
-
-        # Check if data directory exists
-        data_path = Path(args.data_dir)
-        if not data_path.exists():
-            raise ValueError(f"Data directory not found: {args.data_dir}")
-
-        # Load documents
-        reader_kwargs = {
-            "recursive": True,
-            "encoding": "utf-8",
-        }
-        if args.file_types:
-            reader_kwargs["required_exts"] = args.file_types
-
-        documents = SimpleDirectoryReader(args.data_dir, **reader_kwargs).load_data(
-            show_progress=True
-        )
-
-        if not documents:
-            print(f"No documents found in {args.data_dir} with extensions {args.file_types}")
-            return []
-
-        print(f"Loaded {len(documents)} documents")
-
-        # Determine chunking strategy
-        use_ast = args.enable_code_chunking or getattr(args, "use_ast_chunking", False)
-
-        if use_ast:
-            print("Using AST-aware chunking for code files")
-
-        # Convert to text chunks with optional AST support
-        all_texts = create_text_chunks(
-            documents,
-            chunk_size=args.chunk_size,
-            chunk_overlap=args.chunk_overlap,
-            use_ast_chunking=use_ast,
-            ast_chunk_size=getattr(args, "ast_chunk_size", 512),
-            ast_chunk_overlap=getattr(args, "ast_chunk_overlap", 64),
-            code_file_extensions=getattr(args, "code_file_extensions", None),
-            ast_fallback_traditional=getattr(args, "ast_fallback_traditional", True),
-        )
-
-        # Apply max_items limit if specified
-        if args.max_items > 0 and len(all_texts) > args.max_items:
-            print(f"Limiting to {args.max_items} chunks (from {len(all_texts)})")
-            all_texts = all_texts[: args.max_items]
-
-        return all_texts
-
-
-if __name__ == "__main__":
-    import asyncio
-
-    # Example queries for document RAG
-    print("\n📄 Document RAG Example")
-    print("=" * 50)
-    print("\nExample queries you can try:")
-    print("- 'What are the main techniques LEANN uses?'")
-    print("- 'What is the technique DLPM?'")
-    print("- 'Who does Elizabeth Bennet marry?'")
-    print(
-        "- 'What is the problem of developing pan gu model Huawei meets? (盘古大模型开发中遇到什么问题?)'"
-    )
-    print("\n🚀 NEW: Code-aware chunking available!")
-    print("- Use --enable-code-chunking to enable AST-aware chunking for code files")
-    print("- Supports Python, Java, C#, TypeScript files")
-    print("- Better semantic understanding of code structure")
-    print("\nOr run without --query for interactive mode\n")
-
-    rag = DocumentRAG()
-    asyncio.run(rag.run())

apps/email_data/LEANN_email_reader.py

@@ -1,167 +0,0 @@
-import email
-import os
-from pathlib import Path
-from typing import Any
-
-from llama_index.core import Document
-from llama_index.core.readers.base import BaseReader
-
-
-def find_all_messages_directories(root: str | None = None) -> list[Path]:
-    """
-    Recursively find all 'Messages' directories under the given root.
-    Returns a list of Path objects.
-    """
-    if root is None:
-        # Auto-detect user's mail path
-        home_dir = os.path.expanduser("~")
-        root = os.path.join(home_dir, "Library", "Mail")
-
-    messages_dirs = []
-    for dirpath, _dirnames, _filenames in os.walk(root):
-        if os.path.basename(dirpath) == "Messages":
-            messages_dirs.append(Path(dirpath))
-    return messages_dirs
-
-
-class EmlxReader(BaseReader):
-    """
-    Apple Mail .emlx file reader with embedded metadata.
-
-    Reads individual .emlx files from Apple Mail's storage format.
-    """
-
-    def __init__(self, include_html: bool = False) -> None:
-        """
-        Initialize.
-
-        Args:
-            include_html: Whether to include HTML content in the email body (default: False)
-        """
-        self.include_html = include_html
-
-    def load_data(self, input_dir: str, **load_kwargs: Any) -> list[Document]:
-        """
-        Load data from the input directory containing .emlx files.
-
-        Args:
-            input_dir: Directory containing .emlx files
-            **load_kwargs:
-                max_count (int): Maximum amount of messages to read.
-        """
-        docs: list[Document] = []
-        max_count = load_kwargs.get("max_count", 1000)
-        count = 0
-        total_files = 0
-        successful_files = 0
-        failed_files = 0
-
-        print(f"Starting to process directory: {input_dir}")
-
-        # Walk through the directory recursively
-        for dirpath, dirnames, filenames in os.walk(input_dir):
-            # Skip hidden directories
-            dirnames[:] = [d for d in dirnames if not d.startswith(".")]
-
-            for filename in filenames:
-                # Check if we've reached the max count (skip if max_count == -1)
-                if max_count > 0 and count >= max_count:
-                    break
-
-                if filename.endswith(".emlx"):
-                    total_files += 1
-                    filepath = os.path.join(dirpath, filename)
-                    try:
-                        # Read the .emlx file
-                        with open(filepath, encoding="utf-8", errors="ignore") as f:
-                            content = f.read()
-
-                        # .emlx files have a length prefix followed by the email content
-                        # The first line contains the length, followed by the email
-                        lines = content.split("\n", 1)
-                        if len(lines) >= 2:
-                            email_content = lines[1]
-
-                            # Parse the email using Python's email module
-                            try:
-                                msg = email.message_from_string(email_content)
-
-                                # Extract email metadata
-                                subject = msg.get("Subject", "No Subject")
-                                from_addr = msg.get("From", "Unknown")
-                                to_addr = msg.get("To", "Unknown")
-                                date = msg.get("Date", "Unknown")
-
-                                # Extract email body
-                                body = ""
-                                if msg.is_multipart():
-                                    for part in msg.walk():
-                                        if (
-                                            part.get_content_type() == "text/plain"
-                                            or part.get_content_type() == "text/html"
-                                        ):
-                                            if (
-                                                part.get_content_type() == "text/html"
-                                                and not self.include_html
-                                            ):
-                                                continue
-                                            try:
-                                                payload = part.get_payload(decode=True)
-                                                if payload:
-                                                    body += payload.decode("utf-8", errors="ignore")
-                                            except Exception as e:
-                                                print(f"Error decoding payload: {e}")
-                                                continue
-                                else:
-                                    try:
-                                        payload = msg.get_payload(decode=True)
-                                        if payload:
-                                            body = payload.decode("utf-8", errors="ignore")
-                                    except Exception as e:
-                                        print(f"Error decoding single part payload: {e}")
-                                        body = ""
-
-                                # Only create document if we have some content
-                                if body.strip() or subject != "No Subject":
-                                    # Create document content with metadata embedded in text
-                                    doc_content = f"""
-[File]: {filename}
-[From]: {from_addr}
-[To]: {to_addr}
-[Subject]: {subject}
-[Date]: {date}
-[EMAIL BODY Start]:
-{body}
-"""
-
-                                    # No separate metadata - everything is in the text
-                                    doc = Document(text=doc_content, metadata={})
-                                    docs.append(doc)
-                                    count += 1
-                                    successful_files += 1
-
-                                    # Print first few successful files for debugging
-                                    if successful_files <= 3:
-                                        print(
-                                            f"Successfully loaded: {filename} - Subject: {subject[:50]}..."
-                                        )
-
-                            except Exception as e:
-                                failed_files += 1
-                                if failed_files <= 5:  # Only print first few errors
-                                    print(f"Error parsing email from {filepath}: {e}")
-                                continue
-
-                    except Exception as e:
-                        failed_files += 1
-                        if failed_files <= 5:  # Only print first few errors
-                            print(f"Error reading file {filepath}: {e}")
-                        continue
-
-        print("Processing summary:")
-        print(f"  Total .emlx files found: {total_files}")
-        print(f"  Successfully loaded: {successful_files}")
-        print(f"  Failed to load: {failed_files}")
-        print(f"  Final documents: {len(docs)}")
-
-        return docs

apps/email_rag.py

@@ -1,157 +0,0 @@
-"""
-Email RAG example using the unified interface.
-Supports Apple Mail on macOS.
-"""
-
-import sys
-from pathlib import Path
-
-# Add parent directory to path for imports
-sys.path.insert(0, str(Path(__file__).parent))
-
-from base_rag_example import BaseRAGExample
-from chunking import create_text_chunks
-
-from .email_data.LEANN_email_reader import EmlxReader
-
-
-class EmailRAG(BaseRAGExample):
-    """RAG example for Apple Mail processing."""
-
-    def __init__(self):
-        # Set default values BEFORE calling super().__init__
-        self.max_items_default = -1  # Process all emails by default
-        self.embedding_model_default = (
-            "sentence-transformers/all-MiniLM-L6-v2"  # Fast 384-dim model
-        )
-
-        super().__init__(
-            name="Email",
-            description="Process and query Apple Mail emails with LEANN",
-            default_index_name="mail_index",
-        )
-
-    def _add_specific_arguments(self, parser):
-        """Add email-specific arguments."""
-        email_group = parser.add_argument_group("Email Parameters")
-        email_group.add_argument(
-            "--mail-path",
-            type=str,
-            default=None,
-            help="Path to Apple Mail directory (auto-detected if not specified)",
-        )
-        email_group.add_argument(
-            "--include-html", action="store_true", help="Include HTML content in email processing"
-        )
-        email_group.add_argument(
-            "--chunk-size", type=int, default=256, help="Text chunk size (default: 256)"
-        )
-        email_group.add_argument(
-            "--chunk-overlap", type=int, default=25, help="Text chunk overlap (default: 25)"
-        )
-
-    def _find_mail_directories(self) -> list[Path]:
-        """Auto-detect all Apple Mail directories."""
-        mail_base = Path.home() / "Library" / "Mail"
-        if not mail_base.exists():
-            return []
-
-        # Find all Messages directories
-        messages_dirs = []
-        for item in mail_base.rglob("Messages"):
-            if item.is_dir():
-                messages_dirs.append(item)
-
-        return messages_dirs
-
-    async def load_data(self, args) -> list[str]:
-        """Load emails and convert to text chunks."""
-        # Determine mail directories
-        if args.mail_path:
-            messages_dirs = [Path(args.mail_path)]
-        else:
-            print("Auto-detecting Apple Mail directories...")
-            messages_dirs = self._find_mail_directories()
-
-        if not messages_dirs:
-            print("No Apple Mail directories found!")
-            print("Please specify --mail-path manually")
-            return []
-
-        print(f"Found {len(messages_dirs)} mail directories")
-
-        # Create reader
-        reader = EmlxReader(include_html=args.include_html)
-
-        # Process each directory
-        all_documents = []
-        total_processed = 0
-
-        for i, messages_dir in enumerate(messages_dirs):
-            print(f"\nProcessing directory {i + 1}/{len(messages_dirs)}: {messages_dir}")
-
-            try:
-                # Count emlx files
-                emlx_files = list(messages_dir.glob("*.emlx"))
-                print(f"Found {len(emlx_files)} email files")
-
-                # Apply max_items limit per directory
-                max_per_dir = -1  # Default to process all
-                if args.max_items > 0:
-                    remaining = args.max_items - total_processed
-                    if remaining <= 0:
-                        break
-                    max_per_dir = remaining
-                # If args.max_items == -1, max_per_dir stays -1 (process all)
-
-                # Load emails - fix the parameter passing
-                documents = reader.load_data(
-                    input_dir=str(messages_dir),
-                    max_count=max_per_dir,
-                )
-
-                if documents:
-                    all_documents.extend(documents)
-                    total_processed += len(documents)
-                    print(f"Processed {len(documents)} emails from this directory")
-
-            except Exception as e:
-                print(f"Error processing {messages_dir}: {e}")
-                continue
-
-        if not all_documents:
-            print("No emails found to process!")
-            return []
-
-        print(f"\nTotal emails processed: {len(all_documents)}")
-        print("now starting to split into text chunks ... take some time")
-
-        # Convert to text chunks
-        # Email reader uses chunk_overlap=25 as in original
-        all_texts = create_text_chunks(
-            all_documents, chunk_size=args.chunk_size, chunk_overlap=args.chunk_overlap
-        )
-
-        return all_texts
-
-
-if __name__ == "__main__":
-    import asyncio
-
-    # Check platform
-    if sys.platform != "darwin":
-        print("\n⚠️  Warning: This example is designed for macOS (Apple Mail)")
-        print("   Windows/Linux support coming soon!\n")
-
-    # Example queries for email RAG
-    print("\n📧 Email RAG Example")
-    print("=" * 50)
-    print("\nExample queries you can try:")
-    print("- 'What did my boss say about deadlines?'")
-    print("- 'Find emails about travel expenses'")
-    print("- 'Show me emails from last month about the project'")
-    print("- 'What food did I order from DoorDash?'")
-    print("\nNote: You may need to grant Full Disk Access to your terminal\n")
-
-    rag = EmailRAG()
-    asyncio.run(rag.run())

apps/history_data/__init__.py

@@ -1,3 +0,0 @@
-from .history import ChromeHistoryReader
-
-__all__ = ["ChromeHistoryReader"]

apps/history_data/history.py

@@ -1,186 +0,0 @@
-import os
-import sqlite3
-from pathlib import Path
-from typing import Any
-
-from llama_index.core import Document
-from llama_index.core.readers.base import BaseReader
-
-
-class ChromeHistoryReader(BaseReader):
-    """
-    Chrome browser history reader that extracts browsing data from SQLite database.
-
-    Reads Chrome history from the default Chrome profile location and creates documents
-    with embedded metadata similar to the email reader structure.
-    """
-
-    def __init__(self) -> None:
-        """Initialize."""
-        pass
-
-    def load_data(self, input_dir: str | None = None, **load_kwargs: Any) -> list[Document]:
-        """
-        Load Chrome history data from the default Chrome profile location.
-
-        Args:
-            input_dir: Not used for Chrome history (kept for compatibility)
-            **load_kwargs:
-                max_count (int): Maximum amount of history entries to read.
-                chrome_profile_path (str): Custom path to Chrome profile directory.
-        """
-        docs: list[Document] = []
-        max_count = load_kwargs.get("max_count", 1000)
-        chrome_profile_path = load_kwargs.get("chrome_profile_path", None)
-
-        # Default Chrome profile path on macOS
-        if chrome_profile_path is None:
-            chrome_profile_path = os.path.expanduser(
-                "~/Library/Application Support/Google/Chrome/Default"
-            )
-
-        history_db_path = os.path.join(chrome_profile_path, "History")
-
-        if not os.path.exists(history_db_path):
-            print(f"Chrome history database not found at: {history_db_path}")
-            return docs
-
-        try:
-            # Connect to the Chrome history database
-            print(f"Connecting to database: {history_db_path}")
-            conn = sqlite3.connect(history_db_path)
-            cursor = conn.cursor()
-
-            # Query to get browsing history with metadata (removed created_time column)
-            query = """
-            SELECT
-                datetime(last_visit_time/1000000-11644473600,'unixepoch','localtime') as last_visit,
-                url,
-                title,
-                visit_count,
-                typed_count,
-                hidden
-            FROM urls
-            ORDER BY last_visit_time DESC
-            """
-
-            print(f"Executing query on database: {history_db_path}")
-            cursor.execute(query)
-            rows = cursor.fetchall()
-            print(f"Query returned {len(rows)} rows")
-
-            count = 0
-            for row in rows:
-                if count >= max_count and max_count > 0:
-                    break
-
-                last_visit, url, title, visit_count, typed_count, _hidden = row
-
-                # Create document content with metadata embedded in text
-                doc_content = f"""
-[Title]: {title}
-[URL of the page]: {url}
-[Last visited time]: {last_visit}
-[Visit times]: {visit_count}
-[Typed times]: {typed_count}
-"""
-
-                # Create document with embedded metadata
-                doc = Document(text=doc_content, metadata={"title": title[0:150]})
-                # if len(title) > 150:
-                #     print(f"Title is too long: {title}")
-                docs.append(doc)
-                count += 1
-
-            conn.close()
-            print(f"Loaded {len(docs)} Chrome history documents")
-
-        except Exception as e:
-            print(f"Error reading Chrome history: {e}")
-            # add you may need to close your browser to make the database file available
-            # also highlight in red
-            print(
-                "\033[91mYou may need to close your browser to make the database file available\033[0m"
-            )
-            return docs
-
-        return docs
-
-    @staticmethod
-    def find_chrome_profiles() -> list[Path]:
-        """
-        Find all Chrome profile directories.
-
-        Returns:
-            List of Path objects pointing to Chrome profile directories
-        """
-        chrome_base_path = Path(os.path.expanduser("~/Library/Application Support/Google/Chrome"))
-        profile_dirs = []
-
-        if not chrome_base_path.exists():
-            print(f"Chrome directory not found at: {chrome_base_path}")
-            return profile_dirs
-
-        # Find all profile directories
-        for profile_dir in chrome_base_path.iterdir():
-            if profile_dir.is_dir() and profile_dir.name != "System Profile":
-                history_path = profile_dir / "History"
-                if history_path.exists():
-                    profile_dirs.append(profile_dir)
-                    print(f"Found Chrome profile: {profile_dir}")
-
-        print(f"Found {len(profile_dirs)} Chrome profiles")
-        return profile_dirs
-
-    @staticmethod
-    def export_history_to_file(
-        output_file: str = "chrome_history_export.txt", max_count: int = 1000
-    ):
-        """
-        Export Chrome history to a text file using the same SQL query format.
-
-        Args:
-            output_file: Path to the output file
-            max_count: Maximum number of entries to export
-        """
-        chrome_profile_path = os.path.expanduser(
-            "~/Library/Application Support/Google/Chrome/Default"
-        )
-        history_db_path = os.path.join(chrome_profile_path, "History")
-
-        if not os.path.exists(history_db_path):
-            print(f"Chrome history database not found at: {history_db_path}")
-            return
-
-        try:
-            conn = sqlite3.connect(history_db_path)
-            cursor = conn.cursor()
-
-            query = """
-            SELECT
-                datetime(last_visit_time/1000000-11644473600,'unixepoch','localtime') as last_visit,
-                url,
-                title,
-                visit_count,
-                typed_count,
-                hidden
-            FROM urls
-            ORDER BY last_visit_time DESC
-            LIMIT ?
-            """
-
-            cursor.execute(query, (max_count,))
-            rows = cursor.fetchall()
-
-            with open(output_file, "w", encoding="utf-8") as f:
-                for row in rows:
-                    last_visit, url, title, visit_count, typed_count, hidden = row
-                    f.write(
-                        f"{last_visit}\t{url}\t{title}\t{visit_count}\t{typed_count}\t{hidden}\n"
-                    )
-
-            conn.close()
-            print(f"Exported {len(rows)} history entries to {output_file}")
-
-        except Exception as e:
-            print(f"Error exporting Chrome history: {e}")

apps/history_data/wechat_history.py

@@ -1,774 +0,0 @@
-import json
-import os
-import re
-import subprocess
-import time
-from datetime import datetime
-from pathlib import Path
-from typing import Any
-
-from llama_index.core import Document
-from llama_index.core.readers.base import BaseReader
-
-
-class WeChatHistoryReader(BaseReader):
-    """
-    WeChat chat history reader that extracts chat data from exported JSON files.
-
-    Reads WeChat chat history from exported JSON files (from wechat-exporter tool)
-    and creates documents with embedded metadata similar to the Chrome history reader structure.
-
-    Also includes utilities for automatic WeChat chat history export.
-    """
-
-    def __init__(self) -> None:
-        """Initialize."""
-        self.packages_dir = Path(__file__).parent.parent.parent / "packages"
-        self.wechat_exporter_dir = self.packages_dir / "wechat-exporter"
-        self.wechat_decipher_dir = self.packages_dir / "wechat-decipher-macos"
-
-    def check_wechat_running(self) -> bool:
-        """Check if WeChat is currently running."""
-        try:
-            result = subprocess.run(["pgrep", "-f", "WeChat"], capture_output=True, text=True)
-            return result.returncode == 0
-        except Exception:
-            return False
-
-    def install_wechattweak(self) -> bool:
-        """Install WeChatTweak CLI tool."""
-        try:
-            # Create wechat-exporter directory if it doesn't exist
-            self.wechat_exporter_dir.mkdir(parents=True, exist_ok=True)
-
-            wechattweak_path = self.wechat_exporter_dir / "wechattweak-cli"
-            if not wechattweak_path.exists():
-                print("Downloading WeChatTweak CLI...")
-                subprocess.run(
-                    [
-                        "curl",
-                        "-L",
-                        "-o",
-                        str(wechattweak_path),
-                        "https://github.com/JettChenT/WeChatTweak-CLI/releases/latest/download/wechattweak-cli",
-                    ],
-                    check=True,
-                )
-
-            # Make executable
-            wechattweak_path.chmod(0o755)
-
-            # Install WeChatTweak
-            print("Installing WeChatTweak...")
-            subprocess.run(["sudo", str(wechattweak_path), "install"], check=True)
-            return True
-        except Exception as e:
-            print(f"Error installing WeChatTweak: {e}")
-            return False
-
-    def restart_wechat(self):
-        """Restart WeChat to apply WeChatTweak."""
-        try:
-            print("Restarting WeChat...")
-            subprocess.run(["pkill", "-f", "WeChat"], check=False)
-            time.sleep(2)
-            subprocess.run(["open", "-a", "WeChat"], check=True)
-            time.sleep(5)  # Wait for WeChat to start
-        except Exception as e:
-            print(f"Error restarting WeChat: {e}")
-
-    def check_api_available(self) -> bool:
-        """Check if WeChatTweak API is available."""
-        try:
-            result = subprocess.run(
-                ["curl", "-s", "http://localhost:48065/wechat/allcontacts"],
-                capture_output=True,
-                text=True,
-                timeout=5,
-            )
-            return result.returncode == 0 and result.stdout.strip()
-        except Exception:
-            return False
-
-    def _extract_readable_text(self, content: str) -> str:
-        """
-        Extract readable text from message content, removing XML and system messages.
-
-        Args:
-            content: The raw message content (can be string or dict)
-
-        Returns:
-            Cleaned, readable text
-        """
-        if not content:
-            return ""
-
-        # Handle dictionary content (like quoted messages)
-        if isinstance(content, dict):
-            # Extract text from dictionary structure
-            text_parts = []
-            if "title" in content:
-                text_parts.append(str(content["title"]))
-            if "quoted" in content:
-                text_parts.append(str(content["quoted"]))
-            if "content" in content:
-                text_parts.append(str(content["content"]))
-            if "text" in content:
-                text_parts.append(str(content["text"]))
-
-            if text_parts:
-                return " | ".join(text_parts)
-            else:
-                # If we can't extract meaningful text from dict, return empty
-                return ""
-
-        # Handle string content
-        if not isinstance(content, str):
-            return ""
-
-        # Remove common prefixes like "wxid_xxx:\n"
-        clean_content = re.sub(r"^wxid_[^:]+:\s*", "", content)
-        clean_content = re.sub(r"^[^:]+:\s*", "", clean_content)
-
-        # If it's just XML or system message, return empty
-        if clean_content.strip().startswith("<") or "recalled a message" in clean_content:
-            return ""
-
-        return clean_content.strip()
-
-    def _is_text_message(self, content: str) -> bool:
-        """
-        Check if a message contains readable text content.
-
-        Args:
-            content: The message content (can be string or dict)
-
-        Returns:
-            True if the message contains readable text, False otherwise
-        """
-        if not content:
-            return False
-
-        # Handle dictionary content
-        if isinstance(content, dict):
-            # Check if dict has any readable text fields
-            text_fields = ["title", "quoted", "content", "text"]
-            for field in text_fields:
-                if content.get(field):
-                    return True
-            return False
-
-        # Handle string content
-        if not isinstance(content, str):
-            return False
-
-        # Skip image messages (contain XML with img tags)
-        if "<img" in content and "cdnurl" in content:
-            return False
-
-        # Skip emoji messages (contain emoji XML tags)
-        if "<emoji" in content and "productid" in content:
-            return False
-
-        # Skip voice messages
-        if "<voice" in content:
-            return False
-
-        # Skip video messages
-        if "<video" in content:
-            return False
-
-        # Skip file messages
-        if "<appmsg" in content and "appid" in content:
-            return False
-
-        # Skip system messages (like "recalled a message")
-        if "recalled a message" in content:
-            return False
-
-        # Check if there's actual readable text (not just XML or system messages)
-        # Remove common prefixes like "wxid_xxx:\n" and check for actual content
-        clean_content = re.sub(r"^wxid_[^:]+:\s*", "", content)
-        clean_content = re.sub(r"^[^:]+:\s*", "", clean_content)
-
-        # If after cleaning we have meaningful text, consider it readable
-        if len(clean_content.strip()) > 0 and not clean_content.strip().startswith("<"):
-            return True
-
-        return False
-
-    def _concatenate_messages(
-        self,
-        messages: list[dict],
-        max_length: int = 128,
-        time_window_minutes: int = 30,
-        overlap_messages: int = 0,
-    ) -> list[dict]:
-        """
-        Concatenate messages based on length and time rules.
-
-        Args:
-            messages: List of message dictionaries
-            max_length: Maximum length for concatenated message groups. Use -1 to disable length constraint.
-            time_window_minutes: Time window in minutes to group messages together. Use -1 to disable time constraint.
-            overlap_messages: Number of messages to overlap between consecutive groups
-
-        Returns:
-            List of concatenated message groups
-        """
-        if not messages:
-            return []
-
-        concatenated_groups = []
-        current_group = []
-        current_length = 0
-        last_timestamp = None
-
-        for message in messages:
-            # Extract message info
-            content = message.get("content", "")
-            message_text = message.get("message", "")
-            create_time = message.get("createTime", 0)
-            message.get("fromUser", "")
-            message.get("toUser", "")
-            message.get("isSentFromSelf", False)
-
-            # Extract readable text
-            readable_text = self._extract_readable_text(content)
-            if not readable_text:
-                readable_text = message_text
-
-            # Skip empty messages
-            if not readable_text.strip():
-                continue
-
-            # Check time window constraint (only if time_window_minutes != -1)
-            if time_window_minutes != -1 and last_timestamp is not None and create_time > 0:
-                time_diff_minutes = (create_time - last_timestamp) / 60
-                if time_diff_minutes > time_window_minutes:
-                    # Time gap too large, start new group
-                    if current_group:
-                        concatenated_groups.append(
-                            {
-                                "messages": current_group,
-                                "total_length": current_length,
-                                "start_time": current_group[0].get("createTime", 0),
-                                "end_time": current_group[-1].get("createTime", 0),
-                            }
-                        )
-                        # Keep last few messages for overlap
-                        if overlap_messages > 0 and len(current_group) > overlap_messages:
-                            current_group = current_group[-overlap_messages:]
-                            current_length = sum(
-                                len(
-                                    self._extract_readable_text(msg.get("content", ""))
-                                    or msg.get("message", "")
-                                )
-                                for msg in current_group
-                            )
-                        else:
-                            current_group = []
-                            current_length = 0
-
-            # Check length constraint (only if max_length != -1)
-            message_length = len(readable_text)
-            if max_length != -1 and current_length + message_length > max_length and current_group:
-                # Current group would exceed max length, save it and start new
-                concatenated_groups.append(
-                    {
-                        "messages": current_group,
-                        "total_length": current_length,
-                        "start_time": current_group[0].get("createTime", 0),
-                        "end_time": current_group[-1].get("createTime", 0),
-                    }
-                )
-                # Keep last few messages for overlap
-                if overlap_messages > 0 and len(current_group) > overlap_messages:
-                    current_group = current_group[-overlap_messages:]
-                    current_length = sum(
-                        len(
-                            self._extract_readable_text(msg.get("content", ""))
-                            or msg.get("message", "")
-                        )
-                        for msg in current_group
-                    )
-                else:
-                    current_group = []
-                    current_length = 0
-
-            # Add message to current group
-            current_group.append(message)
-            current_length += message_length
-            last_timestamp = create_time
-
-        # Add the last group if it exists
-        if current_group:
-            concatenated_groups.append(
-                {
-                    "messages": current_group,
-                    "total_length": current_length,
-                    "start_time": current_group[0].get("createTime", 0),
-                    "end_time": current_group[-1].get("createTime", 0),
-                }
-            )
-
-        return concatenated_groups
-
-    def _create_concatenated_content(self, message_group: dict, contact_name: str) -> str:
-        """
-        Create concatenated content from a group of messages.
-
-        Args:
-            message_group: Dictionary containing messages and metadata
-            contact_name: Name of the contact
-
-        Returns:
-            Formatted concatenated content
-        """
-        messages = message_group["messages"]
-        start_time = message_group["start_time"]
-        end_time = message_group["end_time"]
-
-        # Format timestamps
-        if start_time:
-            try:
-                start_timestamp = datetime.fromtimestamp(start_time)
-                start_time_str = start_timestamp.strftime("%Y-%m-%d %H:%M:%S")
-            except (ValueError, OSError):
-                start_time_str = str(start_time)
-        else:
-            start_time_str = "Unknown"
-
-        if end_time:
-            try:
-                end_timestamp = datetime.fromtimestamp(end_time)
-                end_time_str = end_timestamp.strftime("%Y-%m-%d %H:%M:%S")
-            except (ValueError, OSError):
-                end_time_str = str(end_time)
-        else:
-            end_time_str = "Unknown"
-
-        # Build concatenated message content
-        message_parts = []
-        for message in messages:
-            content = message.get("content", "")
-            message_text = message.get("message", "")
-            create_time = message.get("createTime", 0)
-            is_sent_from_self = message.get("isSentFromSelf", False)
-
-            # Extract readable text
-            readable_text = self._extract_readable_text(content)
-            if not readable_text:
-                readable_text = message_text
-
-            # Format individual message
-            if create_time:
-                try:
-                    timestamp = datetime.fromtimestamp(create_time)
-                    # change to YYYY-MM-DD HH:MM:SS
-                    time_str = timestamp.strftime("%Y-%m-%d %H:%M:%S")
-                except (ValueError, OSError):
-                    time_str = str(create_time)
-            else:
-                time_str = "Unknown"
-
-            sender = "[Me]" if is_sent_from_self else "[Contact]"
-            message_parts.append(f"({time_str}) {sender}: {readable_text}")
-
-        concatenated_text = "\n".join(message_parts)
-
-        # Create final document content
-        doc_content = f"""
-Contact: {contact_name}
-Time Range: {start_time_str} - {end_time_str}
-Messages ({len(messages)} messages, {message_group["total_length"]} chars):
-
-{concatenated_text}
-"""
-        # TODO @yichuan give better format and rich info here!
-        doc_content = f"""
-{concatenated_text}
-"""
-        return doc_content, contact_name
-
-    def load_data(self, input_dir: str | None = None, **load_kwargs: Any) -> list[Document]:
-        """
-        Load WeChat chat history data from exported JSON files.
-
-        Args:
-            input_dir: Directory containing exported WeChat JSON files
-            **load_kwargs:
-                max_count (int): Maximum amount of chat entries to read.
-                wechat_export_dir (str): Custom path to WeChat export directory.
-                include_non_text (bool): Whether to include non-text messages (images, emojis, etc.)
-                concatenate_messages (bool): Whether to concatenate messages based on length rules.
-                max_length (int): Maximum length for concatenated message groups (default: 1000).
-                time_window_minutes (int): Time window in minutes to group messages together (default: 30).
-                overlap_messages (int): Number of messages to overlap between consecutive groups (default: 2).
-        """
-        docs: list[Document] = []
-        max_count = load_kwargs.get("max_count", 1000)
-        wechat_export_dir = load_kwargs.get("wechat_export_dir", None)
-        include_non_text = load_kwargs.get("include_non_text", False)
-        concatenate_messages = load_kwargs.get("concatenate_messages", False)
-        max_length = load_kwargs.get("max_length", 1000)
-        time_window_minutes = load_kwargs.get("time_window_minutes", 30)
-
-        # Default WeChat export path
-        if wechat_export_dir is None:
-            wechat_export_dir = "./wechat_export_test"
-
-        if not os.path.exists(wechat_export_dir):
-            print(f"WeChat export directory not found at: {wechat_export_dir}")
-            return docs
-
-        try:
-            # Find all JSON files in the export directory
-            json_files = list(Path(wechat_export_dir).glob("*.json"))
-            print(f"Found {len(json_files)} WeChat chat history files")
-
-            count = 0
-            for json_file in json_files:
-                if count >= max_count and max_count > 0:
-                    break
-
-                try:
-                    with open(json_file, encoding="utf-8") as f:
-                        chat_data = json.load(f)
-
-                    # Extract contact name from filename
-                    contact_name = json_file.stem
-
-                    if concatenate_messages:
-                        # Filter messages to only include readable text messages
-                        readable_messages = []
-                        for message in chat_data:
-                            try:
-                                content = message.get("content", "")
-                                if not include_non_text and not self._is_text_message(content):
-                                    continue
-
-                                readable_text = self._extract_readable_text(content)
-                                if not readable_text and not include_non_text:
-                                    continue
-
-                                readable_messages.append(message)
-                            except Exception as e:
-                                print(f"Error processing message in {json_file}: {e}")
-                                continue
-
-                        # Concatenate messages based on rules
-                        message_groups = self._concatenate_messages(
-                            readable_messages,
-                            max_length=max_length,
-                            time_window_minutes=time_window_minutes,
-                            overlap_messages=0,  # No overlap between groups
-                        )
-
-                        # Create documents from concatenated groups
-                        for message_group in message_groups:
-                            if count >= max_count and max_count > 0:
-                                break
-
-                            doc_content, contact_name = self._create_concatenated_content(
-                                message_group, contact_name
-                            )
-                            doc = Document(
-                                text=doc_content,
-                                metadata={"contact_name": contact_name},
-                            )
-                            docs.append(doc)
-                            count += 1
-
-                        print(
-                            f"Created {len(message_groups)} concatenated message groups for {contact_name}"
-                        )
-
-                    else:
-                        # Original single-message processing
-                        for message in chat_data:
-                            if count >= max_count and max_count > 0:
-                                break
-
-                            # Extract message information
-                            message.get("fromUser", "")
-                            message.get("toUser", "")
-                            content = message.get("content", "")
-                            message_text = message.get("message", "")
-                            create_time = message.get("createTime", 0)
-                            is_sent_from_self = message.get("isSentFromSelf", False)
-
-                            # Handle content that might be dict or string
-                            try:
-                                # Check if this is a readable text message
-                                if not include_non_text and not self._is_text_message(content):
-                                    continue
-
-                                # Extract readable text
-                                readable_text = self._extract_readable_text(content)
-                                if not readable_text and not include_non_text:
-                                    continue
-                            except Exception as e:
-                                # Skip messages that cause processing errors
-                                print(f"Error processing message in {json_file}: {e}")
-                                continue
-
-                            # Convert timestamp to readable format
-                            if create_time:
-                                try:
-                                    timestamp = datetime.fromtimestamp(create_time)
-                                    time_str = timestamp.strftime("%Y-%m-%d %H:%M:%S")
-                                except (ValueError, OSError):
-                                    time_str = str(create_time)
-                            else:
-                                time_str = "Unknown"
-
-                            # Create document content with metadata header and contact info
-                            doc_content = f"""
-Contact: {contact_name}
-Is sent from self: {is_sent_from_self}
-Time: {time_str}
-Message: {readable_text if readable_text else message_text}
-"""
-
-                            # Create document with embedded metadata
-                            doc = Document(
-                                text=doc_content, metadata={"contact_name": contact_name}
-                            )
-                            docs.append(doc)
-                            count += 1
-
-                except Exception as e:
-                    print(f"Error reading {json_file}: {e}")
-                    continue
-
-            print(f"Loaded {len(docs)} WeChat chat documents")
-
-        except Exception as e:
-            print(f"Error reading WeChat history: {e}")
-            return docs
-
-        return docs
-
-    @staticmethod
-    def find_wechat_export_dirs() -> list[Path]:
-        """
-        Find all WeChat export directories.
-
-        Returns:
-            List of Path objects pointing to WeChat export directories
-        """
-        export_dirs = []
-
-        # Look for common export directory names
-        possible_dirs = [
-            Path("./wechat_export"),
-            Path("./wechat_export_direct"),
-            Path("./wechat_chat_history"),
-            Path("./chat_export"),
-        ]
-
-        for export_dir in possible_dirs:
-            if export_dir.exists() and export_dir.is_dir():
-                json_files = list(export_dir.glob("*.json"))
-                if json_files:
-                    export_dirs.append(export_dir)
-                    print(
-                        f"Found WeChat export directory: {export_dir} with {len(json_files)} files"
-                    )
-
-        print(f"Found {len(export_dirs)} WeChat export directories")
-        return export_dirs
-
-    @staticmethod
-    def export_chat_to_file(
-        output_file: str = "wechat_chat_export.txt",
-        max_count: int = 1000,
-        export_dir: str | None = None,
-        include_non_text: bool = False,
-    ):
-        """
-        Export WeChat chat history to a text file.
-
-        Args:
-            output_file: Path to the output file
-            max_count: Maximum number of entries to export
-            export_dir: Directory containing WeChat JSON files
-            include_non_text: Whether to include non-text messages
-        """
-        if export_dir is None:
-            export_dir = "./wechat_export_test"
-
-        if not os.path.exists(export_dir):
-            print(f"WeChat export directory not found at: {export_dir}")
-            return
-
-        try:
-            json_files = list(Path(export_dir).glob("*.json"))
-
-            with open(output_file, "w", encoding="utf-8") as f:
-                count = 0
-                for json_file in json_files:
-                    if count >= max_count and max_count > 0:
-                        break
-
-                    try:
-                        with open(json_file, encoding="utf-8") as json_f:
-                            chat_data = json.load(json_f)
-
-                        contact_name = json_file.stem
-                        f.write(f"\n=== Chat with {contact_name} ===\n")
-
-                        for message in chat_data:
-                            if count >= max_count and max_count > 0:
-                                break
-
-                            from_user = message.get("fromUser", "")
-                            content = message.get("content", "")
-                            message_text = message.get("message", "")
-                            create_time = message.get("createTime", 0)
-
-                            # Skip non-text messages unless requested
-                            if not include_non_text:
-                                reader = WeChatHistoryReader()
-                                if not reader._is_text_message(content):
-                                    continue
-                                readable_text = reader._extract_readable_text(content)
-                                if not readable_text:
-                                    continue
-                                message_text = readable_text
-
-                            if create_time:
-                                try:
-                                    timestamp = datetime.fromtimestamp(create_time)
-                                    time_str = timestamp.strftime("%Y-%m-%d %H:%M:%S")
-                                except (ValueError, OSError):
-                                    time_str = str(create_time)
-                            else:
-                                time_str = "Unknown"
-
-                            f.write(f"[{time_str}] {from_user}: {message_text}\n")
-                            count += 1
-
-                    except Exception as e:
-                        print(f"Error processing {json_file}: {e}")
-                        continue
-
-            print(f"Exported {count} chat entries to {output_file}")
-
-        except Exception as e:
-            print(f"Error exporting WeChat chat history: {e}")
-
-    def export_wechat_chat_history(self, export_dir: str = "./wechat_export_direct") -> Path | None:
-        """
-        Export WeChat chat history using wechat-exporter tool.
-
-        Args:
-            export_dir: Directory to save exported chat history
-
-        Returns:
-            Path to export directory if successful, None otherwise
-        """
-        try:
-            import subprocess
-            import sys
-
-            # Create export directory
-            export_path = Path(export_dir)
-            export_path.mkdir(exist_ok=True)
-
-            print(f"Exporting WeChat chat history to {export_path}...")
-
-            # Check if wechat-exporter directory exists
-            if not self.wechat_exporter_dir.exists():
-                print(f"wechat-exporter directory not found at: {self.wechat_exporter_dir}")
-                return None
-
-            # Install requirements if needed
-            requirements_file = self.wechat_exporter_dir / "requirements.txt"
-            if requirements_file.exists():
-                print("Installing wechat-exporter requirements...")
-                subprocess.run(["uv", "pip", "install", "-r", str(requirements_file)], check=True)
-
-            # Run the export command
-            print("Running wechat-exporter...")
-            result = subprocess.run(
-                [
-                    sys.executable,
-                    str(self.wechat_exporter_dir / "main.py"),
-                    "export-all",
-                    str(export_path),
-                ],
-                capture_output=True,
-                text=True,
-                check=True,
-            )
-
-            print("Export command output:")
-            print(result.stdout)
-            if result.stderr:
-                print("Export errors:")
-                print(result.stderr)
-
-            # Check if export was successful
-            if export_path.exists() and any(export_path.glob("*.json")):
-                json_files = list(export_path.glob("*.json"))
-                print(
-                    f"Successfully exported {len(json_files)} chat history files to {export_path}"
-                )
-                return export_path
-            else:
-                print("Export completed but no JSON files found")
-                return None
-
-        except subprocess.CalledProcessError as e:
-            print(f"Export command failed: {e}")
-            print(f"Command output: {e.stdout}")
-            print(f"Command errors: {e.stderr}")
-            return None
-        except Exception as e:
-            print(f"Export failed: {e}")
-            print("Please ensure WeChat is running and WeChatTweak is installed.")
-            return None
-
-    def find_or_export_wechat_data(self, export_dir: str = "./wechat_export_direct") -> list[Path]:
-        """
-        Find existing WeChat exports or create new ones.
-
-        Args:
-            export_dir: Directory to save exported chat history if needed
-
-        Returns:
-            List of Path objects pointing to WeChat export directories
-        """
-        export_dirs = []
-
-        # Look for existing exports in common locations
-        possible_export_dirs = [
-            Path("./wechat_database_export"),
-            Path("./wechat_export_test"),
-            Path("./wechat_export"),
-            Path("./wechat_export_direct"),
-            Path("./wechat_chat_history"),
-            Path("./chat_export"),
-        ]
-
-        for export_dir_path in possible_export_dirs:
-            if export_dir_path.exists() and any(export_dir_path.glob("*.json")):
-                export_dirs.append(export_dir_path)
-                print(f"Found existing export: {export_dir_path}")
-
-        # If no existing exports, try to export automatically
-        if not export_dirs:
-            print("No existing WeChat exports found. Starting direct export...")
-
-            # Try to export using wechat-exporter
-            exported_path = self.export_wechat_chat_history(export_dir)
-            if exported_path:
-                export_dirs = [exported_path]
-            else:
-                print(
-                    "Failed to export WeChat data. Please ensure WeChat is running and WeChatTweak is installed."
-                )
-
-        return export_dirs

apps/wechat_rag.py

@@ -1,189 +0,0 @@
-"""
-WeChat History RAG example using the unified interface.
-Supports WeChat chat history export and search.
-"""
-
-import subprocess
-import sys
-from pathlib import Path
-
-# Add parent directory to path for imports
-sys.path.insert(0, str(Path(__file__).parent))
-
-from base_rag_example import BaseRAGExample
-
-from .history_data.wechat_history import WeChatHistoryReader
-
-
-class WeChatRAG(BaseRAGExample):
-    """RAG example for WeChat chat history."""
-
-    def __init__(self):
-        # Set default values BEFORE calling super().__init__
-        self.max_items_default = -1  # Match original default
-        self.embedding_model_default = (
-            "sentence-transformers/all-MiniLM-L6-v2"  # Fast 384-dim model
-        )
-
-        super().__init__(
-            name="WeChat History",
-            description="Process and query WeChat chat history with LEANN",
-            default_index_name="wechat_history_magic_test_11Debug_new",
-        )
-
-    def _add_specific_arguments(self, parser):
-        """Add WeChat-specific arguments."""
-        wechat_group = parser.add_argument_group("WeChat Parameters")
-        wechat_group.add_argument(
-            "--export-dir",
-            type=str,
-            default="./wechat_export",
-            help="Directory to store WeChat exports (default: ./wechat_export)",
-        )
-        wechat_group.add_argument(
-            "--force-export",
-            action="store_true",
-            help="Force re-export of WeChat data even if exports exist",
-        )
-        wechat_group.add_argument(
-            "--chunk-size", type=int, default=192, help="Text chunk size (default: 192)"
-        )
-        wechat_group.add_argument(
-            "--chunk-overlap", type=int, default=64, help="Text chunk overlap (default: 64)"
-        )
-
-    def _export_wechat_data(self, export_dir: Path) -> bool:
-        """Export WeChat data using wechattweak-cli."""
-        print("Exporting WeChat data...")
-
-        # Check if WeChat is running
-        try:
-            result = subprocess.run(["pgrep", "WeChat"], capture_output=True, text=True)
-            if result.returncode != 0:
-                print("WeChat is not running. Please start WeChat first.")
-                return False
-        except Exception:
-            pass  # pgrep might not be available on all systems
-
-        # Create export directory
-        export_dir.mkdir(parents=True, exist_ok=True)
-
-        # Run export command
-        cmd = ["packages/wechat-exporter/wechattweak-cli", "export", str(export_dir)]
-
-        try:
-            print(f"Running: {' '.join(cmd)}")
-            result = subprocess.run(cmd, capture_output=True, text=True)
-
-            if result.returncode == 0:
-                print("WeChat data exported successfully!")
-                return True
-            else:
-                print(f"Export failed: {result.stderr}")
-                return False
-
-        except FileNotFoundError:
-            print("\nError: wechattweak-cli not found!")
-            print("Please install it first:")
-            print("  sudo packages/wechat-exporter/wechattweak-cli install")
-            return False
-        except Exception as e:
-            print(f"Export error: {e}")
-            return False
-
-    async def load_data(self, args) -> list[str]:
-        """Load WeChat history and convert to text chunks."""
-        # Initialize WeChat reader with export capabilities
-        reader = WeChatHistoryReader()
-
-        # Find existing exports or create new ones using the centralized method
-        export_dirs = reader.find_or_export_wechat_data(args.export_dir)
-        if not export_dirs:
-            print("Failed to find or export WeChat data. Trying to find any existing exports...")
-            # Try to find any existing exports in common locations
-            export_dirs = reader.find_wechat_export_dirs()
-            if not export_dirs:
-                print("No WeChat data found. Please ensure WeChat exports exist.")
-                return []
-
-        # Load documents from all found export directories
-        all_documents = []
-        total_processed = 0
-
-        for i, export_dir in enumerate(export_dirs):
-            print(f"\nProcessing WeChat export {i + 1}/{len(export_dirs)}: {export_dir}")
-
-            try:
-                # Apply max_items limit per export
-                max_per_export = -1
-                if args.max_items > 0:
-                    remaining = args.max_items - total_processed
-                    if remaining <= 0:
-                        break
-                    max_per_export = remaining
-
-                documents = reader.load_data(
-                    wechat_export_dir=str(export_dir),
-                    max_count=max_per_export,
-                    concatenate_messages=True,  # Enable message concatenation for better context
-                )
-
-                if documents:
-                    print(f"Loaded {len(documents)} chat documents from {export_dir}")
-                    all_documents.extend(documents)
-                    total_processed += len(documents)
-                else:
-                    print(f"No documents loaded from {export_dir}")
-
-            except Exception as e:
-                print(f"Error processing {export_dir}: {e}")
-                continue
-
-        if not all_documents:
-            print("No documents loaded from any source. Exiting.")
-            return []
-
-        print(f"\nTotal loaded {len(all_documents)} chat documents from {len(export_dirs)} exports")
-        print("now starting to split into text chunks ... take some time")
-
-        # Convert to text chunks with contact information
-        all_texts = []
-        for doc in all_documents:
-            # Split the document into chunks
-            from llama_index.core.node_parser import SentenceSplitter
-
-            text_splitter = SentenceSplitter(
-                chunk_size=args.chunk_size, chunk_overlap=args.chunk_overlap
-            )
-            nodes = text_splitter.get_nodes_from_documents([doc])
-
-            for node in nodes:
-                # Add contact information to each chunk
-                contact_name = doc.metadata.get("contact_name", "Unknown")
-                text = f"[Contact] means the message is from: {contact_name}\n" + node.get_content()
-                all_texts.append(text)
-
-        print(f"Created {len(all_texts)} text chunks from {len(all_documents)} documents")
-        return all_texts
-
-
-if __name__ == "__main__":
-    import asyncio
-
-    # Check platform
-    if sys.platform != "darwin":
-        print("\n⚠️  Warning: WeChat export is only supported on macOS")
-        print("   You can still query existing exports on other platforms\n")
-
-    # Example queries for WeChat RAG
-    print("\n💬 WeChat History RAG Example")
-    print("=" * 50)
-    print("\nExample queries you can try:")
-    print("- 'Show me conversations about travel plans'")
-    print("- 'Find group chats about weekend activities'")
-    print("- '我想买魔术师约翰逊的球衣,给我一些对应聊天记录?'")
-    print("- 'What did we discuss about the project last month?'")
-    print("\nNote: WeChat must be running for export to work\n")
-
-    rag = WeChatRAG()
-    asyncio.run(rag.run())

benchmarks/benchmark_embeddings.py

@@ -1,141 +0,0 @@
-import time
-
-import matplotlib.pyplot as plt
-import mlx.core as mx
-import numpy as np
-import torch
-from mlx_lm import load
-from sentence_transformers import SentenceTransformer
-
-# --- Configuration ---
-MODEL_NAME_TORCH = "Qwen/Qwen3-Embedding-0.6B"
-MODEL_NAME_MLX = "mlx-community/Qwen3-Embedding-0.6B-4bit-DWQ"
-BATCH_SIZES = [1, 8, 16, 32, 64, 128]
-NUM_RUNS = 10  # Number of runs to average for each batch size
-WARMUP_RUNS = 2  # Number of warm-up runs
-
-# --- Generate Dummy Data ---
-DUMMY_SENTENCES = ["This is a test sentence for benchmarking." * 5] * max(BATCH_SIZES)
-
-# --- Benchmark Functions ---b
-
-
-def benchmark_torch(model, sentences):
-    start_time = time.time()
-    model.encode(sentences, convert_to_numpy=True)
-    end_time = time.time()
-    return (end_time - start_time) * 1000  # Return time in ms
-
-
-def benchmark_mlx(model, tokenizer, sentences):
-    start_time = time.time()
-
-    # Tokenize sentences using MLX tokenizer
-    tokens = []
-    for sentence in sentences:
-        token_ids = tokenizer.encode(sentence)
-        tokens.append(token_ids)
-
-    # Pad sequences to the same length
-    max_len = max(len(t) for t in tokens)
-    input_ids = []
-    attention_mask = []
-
-    for token_seq in tokens:
-        # Pad sequence
-        padded = token_seq + [tokenizer.eos_token_id] * (max_len - len(token_seq))
-        input_ids.append(padded)
-        # Create attention mask (1 for real tokens, 0 for padding)
-        mask = [1] * len(token_seq) + [0] * (max_len - len(token_seq))
-        attention_mask.append(mask)
-
-    # Convert to MLX arrays
-    input_ids = mx.array(input_ids)
-    attention_mask = mx.array(attention_mask)
-
-    # Get embeddings
-    embeddings = model(input_ids)
-
-    # Mean pooling
-    mask = mx.expand_dims(attention_mask, -1)
-    sum_embeddings = (embeddings * mask).sum(axis=1)
-    sum_mask = mask.sum(axis=1)
-    _ = sum_embeddings / sum_mask
-
-    mx.eval()  # Ensure computation is finished
-    end_time = time.time()
-    return (end_time - start_time) * 1000  # Return time in ms
-
-
-# --- Main Execution ---
-def main():
-    print("--- Initializing Models ---")
-    # Load PyTorch model
-    print(f"Loading PyTorch model: {MODEL_NAME_TORCH}")
-    device = "mps" if torch.backends.mps.is_available() else "cpu"
-    model_torch = SentenceTransformer(MODEL_NAME_TORCH, device=device)
-    print(f"PyTorch model loaded on: {device}")
-
-    # Load MLX model
-    print(f"Loading MLX model: {MODEL_NAME_MLX}")
-    model_mlx, tokenizer_mlx = load(MODEL_NAME_MLX)
-    print("MLX model loaded.")
-
-    # --- Warm-up ---
-    print("\n--- Performing Warm-up Runs ---")
-    for _ in range(WARMUP_RUNS):
-        benchmark_torch(model_torch, DUMMY_SENTENCES[:1])
-        benchmark_mlx(model_mlx, tokenizer_mlx, DUMMY_SENTENCES[:1])
-    print("Warm-up complete.")
-
-    # --- Benchmarking ---
-    print("\n--- Starting Benchmark ---")
-    results_torch = []
-    results_mlx = []
-
-    for batch_size in BATCH_SIZES:
-        print(f"Benchmarking batch size: {batch_size}")
-        sentences_batch = DUMMY_SENTENCES[:batch_size]
-
-        # Benchmark PyTorch
-        torch_times = [benchmark_torch(model_torch, sentences_batch) for _ in range(NUM_RUNS)]
-        results_torch.append(np.mean(torch_times))
-
-        # Benchmark MLX
-        mlx_times = [
-            benchmark_mlx(model_mlx, tokenizer_mlx, sentences_batch) for _ in range(NUM_RUNS)
-        ]
-        results_mlx.append(np.mean(mlx_times))
-
-    print("\n--- Benchmark Results (Average time per batch in ms) ---")
-    print(f"Batch Sizes: {BATCH_SIZES}")
-    print(f"PyTorch (mps): {[f'{t:.2f}' for t in results_torch]}")
-    print(f"MLX:           {[f'{t:.2f}' for t in results_mlx]}")
-
-    # --- Plotting ---
-    print("\n--- Generating Plot ---")
-    plt.figure(figsize=(10, 6))
-    plt.plot(
-        BATCH_SIZES,
-        results_torch,
-        marker="o",
-        linestyle="-",
-        label=f"PyTorch ({device})",
-    )
-    plt.plot(BATCH_SIZES, results_mlx, marker="s", linestyle="-", label="MLX")
-
-    plt.title(f"Embedding Performance: MLX vs PyTorch\nModel: {MODEL_NAME_TORCH}")
-    plt.xlabel("Batch Size")
-    plt.ylabel("Average Time per Batch (ms)")
-    plt.xticks(BATCH_SIZES)
-    plt.grid(True)
-    plt.legend()
-
-    # Save the plot
-    output_filename = "embedding_benchmark.png"
-    plt.savefig(output_filename)
-    print(f"Plot saved to {output_filename}")
-
-
-if __name__ == "__main__":
-    main()

benchmarks/benchmark_no_recompute.py

@@ -1,148 +0,0 @@
-import argparse
-import os
-import time
-from pathlib import Path
-
-from leann import LeannBuilder, LeannSearcher
-
-
-def _meta_exists(index_path: str) -> bool:
-    p = Path(index_path)
-    return (p.parent / f"{p.stem}.meta.json").exists()
-
-
-def ensure_index(index_path: str, backend_name: str, num_docs: int, is_recompute: bool) -> None:
-    # if _meta_exists(index_path):
-    #     return
-    kwargs = {}
-    if backend_name == "hnsw":
-        kwargs["is_compact"] = is_recompute
-    builder = LeannBuilder(
-        backend_name=backend_name,
-        embedding_model=os.getenv("LEANN_EMBED_MODEL", "facebook/contriever"),
-        embedding_mode=os.getenv("LEANN_EMBED_MODE", "sentence-transformers"),
-        graph_degree=32,
-        complexity=64,
-        is_recompute=is_recompute,
-        num_threads=4,
-        **kwargs,
-    )
-    for i in range(num_docs):
-        builder.add_text(
-            f"This is a test document number {i}. It contains some repeated text for benchmarking."
-        )
-    builder.build_index(index_path)
-
-
-def _bench_group(
-    index_path: str,
-    recompute: bool,
-    query: str,
-    repeats: int,
-    complexity: int = 32,
-    top_k: int = 10,
-) -> float:
-    # Independent searcher per group; fixed port when recompute
-    searcher = LeannSearcher(index_path=index_path)
-
-    # Warm-up once
-    _ = searcher.search(
-        query,
-        top_k=top_k,
-        complexity=complexity,
-        recompute_embeddings=recompute,
-    )
-
-    def _once() -> float:
-        t0 = time.time()
-        _ = searcher.search(
-            query,
-            top_k=top_k,
-            complexity=complexity,
-            recompute_embeddings=recompute,
-        )
-        return time.time() - t0
-
-    if repeats <= 1:
-        t = _once()
-    else:
-        vals = [_once() for _ in range(repeats)]
-        vals.sort()
-        t = vals[len(vals) // 2]
-
-    searcher.cleanup()
-    return t
-
-
-def main():
-    parser = argparse.ArgumentParser()
-    parser.add_argument("--num-docs", type=int, default=5000)
-    parser.add_argument("--repeats", type=int, default=3)
-    parser.add_argument("--complexity", type=int, default=32)
-    args = parser.parse_args()
-
-    base = Path.cwd() / ".leann" / "indexes" / f"bench_n{args.num_docs}"
-    base.parent.mkdir(parents=True, exist_ok=True)
-    # ---------- Build HNSW variants ----------
-    hnsw_r = str(base / f"hnsw_recompute_n{args.num_docs}.leann")
-    hnsw_nr = str(base / f"hnsw_norecompute_n{args.num_docs}.leann")
-    ensure_index(hnsw_r, "hnsw", args.num_docs, True)
-    ensure_index(hnsw_nr, "hnsw", args.num_docs, False)
-
-    # ---------- Build DiskANN variants ----------
-    diskann_r = str(base / "diskann_r.leann")
-    diskann_nr = str(base / "diskann_nr.leann")
-    ensure_index(diskann_r, "diskann", args.num_docs, True)
-    ensure_index(diskann_nr, "diskann", args.num_docs, False)
-
-    # ---------- Helpers ----------
-    def _size_for(prefix: str) -> int:
-        p = Path(prefix)
-        base_dir = p.parent
-        stem = p.stem
-        total = 0
-        for f in base_dir.iterdir():
-            if f.is_file() and f.name.startswith(stem):
-                total += f.stat().st_size
-        return total
-
-    # ---------- HNSW benchmark ----------
-    t_hnsw_r = _bench_group(
-        hnsw_r, True, "test document number 42", repeats=args.repeats, complexity=args.complexity
-    )
-    t_hnsw_nr = _bench_group(
-        hnsw_nr, False, "test document number 42", repeats=args.repeats, complexity=args.complexity
-    )
-    size_hnsw_r = _size_for(hnsw_r)
-    size_hnsw_nr = _size_for(hnsw_nr)
-
-    print("Benchmark results (HNSW):")
-    print(f"  recompute=True:  search_time={t_hnsw_r:.3f}s, size={size_hnsw_r / 1024 / 1024:.1f}MB")
-    print(
-        f"  recompute=False: search_time={t_hnsw_nr:.3f}s, size={size_hnsw_nr / 1024 / 1024:.1f}MB"
-    )
-    print("  Expectation: no-recompute should be faster but larger on disk.")
-
-    # ---------- DiskANN benchmark ----------
-    t_diskann_r = _bench_group(
-        diskann_r, True, "DiskANN R test doc 123", repeats=args.repeats, complexity=args.complexity
-    )
-    t_diskann_nr = _bench_group(
-        diskann_nr,
-        False,
-        "DiskANN NR test doc 123",
-        repeats=args.repeats,
-        complexity=args.complexity,
-    )
-    size_diskann_r = _size_for(diskann_r)
-    size_diskann_nr = _size_for(diskann_nr)
-
-    print("\nBenchmark results (DiskANN):")
-    print(f"  build(recompute=True, partition): size={size_diskann_r / 1024 / 1024:.1f}MB")
-    print(f"  build(recompute=False):          size={size_diskann_nr / 1024 / 1024:.1f}MB")
-    print(f"  search recompute=True (final rerank): {t_diskann_r:.3f}s")
-    print(f"  search recompute=False (PQ only):     {t_diskann_nr:.3f}s")
-
-
-if __name__ == "__main__":
-    main()

benchmarks/compare_faiss_vs_leann.py

@@ -1,326 +0,0 @@
-#!/usr/bin/env python3
-"""
-Memory comparison between Faiss HNSW and LEANN HNSW backend
-"""
-
-import gc
-import logging
-import os
-import subprocess
-import sys
-import time
-from pathlib import Path
-
-import psutil
-from llama_index.core.node_parser import SentenceSplitter
-
-# Setup logging
-logging.basicConfig(stream=sys.stdout, level=logging.INFO)
-logger = logging.getLogger(__name__)
-
-
-def get_memory_usage():
-    """Get current memory usage in MB"""
-    process = psutil.Process()
-    return process.memory_info().rss / 1024 / 1024
-
-
-def print_memory_stats(stage: str, start_mem: float):
-    """Print memory statistics"""
-    current_mem = get_memory_usage()
-    diff = current_mem - start_mem
-    print(f"[{stage}] Memory: {current_mem:.1f} MB (+{diff:.1f} MB)")
-    return current_mem
-
-
-class MemoryTracker:
-    def __init__(self, name: str):
-        self.name = name
-        self.start_mem = get_memory_usage()
-        self.stages = []
-
-    def checkpoint(self, stage: str):
-        current_mem = print_memory_stats(f"{self.name} - {stage}", self.start_mem)
-        self.stages.append((stage, current_mem))
-        return current_mem
-
-    def summary(self):
-        print(f"\n=== {self.name} Memory Summary ===")
-        for stage, mem in self.stages:
-            print(f"{stage}: {mem:.1f} MB")
-        peak_mem = max(mem for _, mem in self.stages)
-        print(f"Peak Memory: {peak_mem:.1f} MB")
-        print(f"Total Memory Increase: {peak_mem - self.start_mem:.1f} MB")
-        return peak_mem
-
-
-def test_faiss_hnsw():
-    """Test Faiss HNSW Vector Store in subprocess"""
-    print("\n" + "=" * 50)
-    print("TESTING FAISS HNSW VECTOR STORE")
-    print("=" * 50)
-
-    try:
-        result = subprocess.run(
-            [sys.executable, "benchmarks/faiss_only.py"],
-            capture_output=True,
-            text=True,
-            timeout=300,
-        )
-
-        print(result.stdout)
-        if result.stderr:
-            print("Stderr:", result.stderr)
-
-        if result.returncode != 0:
-            return {
-                "peak_memory": float("inf"),
-                "error": f"Process failed with code {result.returncode}",
-            }
-
-        # Parse peak memory from output
-        lines = result.stdout.split("\n")
-        peak_memory = 0.0
-
-        for line in lines:
-            if "Peak Memory:" in line:
-                peak_memory = float(line.split("Peak Memory:")[1].split("MB")[0].strip())
-
-        return {"peak_memory": peak_memory}
-
-    except Exception as e:
-        return {
-            "peak_memory": float("inf"),
-            "error": str(e),
-        }
-
-
-def test_leann_hnsw():
-    """Test LEANN HNSW Search Memory (load existing index)"""
-    print("\n" + "=" * 50)
-    print("TESTING LEANN HNSW SEARCH MEMORY")
-    print("=" * 50)
-
-    tracker = MemoryTracker("LEANN HNSW Search")
-
-    # Import and setup
-    tracker.checkpoint("Initial")
-
-    from leann.api import LeannSearcher
-
-    tracker.checkpoint("After imports")
-
-    from leann.api import LeannBuilder
-    from llama_index.core import SimpleDirectoryReader
-
-    # Load and parse documents
-    documents = SimpleDirectoryReader(
-        "data",
-        recursive=True,
-        encoding="utf-8",
-        required_exts=[".pdf", ".txt", ".md"],
-    ).load_data()
-
-    tracker.checkpoint("After document loading")
-
-    # Parse into chunks
-    node_parser = SentenceSplitter(
-        chunk_size=256, chunk_overlap=20, separator=" ", paragraph_separator="\n\n"
-    )
-
-    all_texts = []
-    for doc in documents:
-        nodes = node_parser.get_nodes_from_documents([doc])
-        for node in nodes:
-            all_texts.append(node.get_content())
-    print(f"Total number of chunks: {len(all_texts)}")
-
-    tracker.checkpoint("After text chunking")
-
-    # Build LEANN index
-    INDEX_DIR = Path("./test_leann_comparison")
-    INDEX_PATH = str(INDEX_DIR / "comparison.leann")
-
-    # Check if index already exists
-    if os.path.exists(INDEX_PATH + ".meta.json"):
-        print("Loading existing LEANN HNSW index...")
-        tracker.checkpoint("After loading existing index")
-    else:
-        print("Building new LEANN HNSW index...")
-        # Clean up previous index
-        import shutil
-
-        if INDEX_DIR.exists():
-            shutil.rmtree(INDEX_DIR)
-
-        builder = LeannBuilder(
-            backend_name="hnsw",
-            embedding_model="facebook/contriever",
-            graph_degree=32,
-            complexity=64,
-            is_compact=True,
-            is_recompute=True,
-            num_threads=1,
-        )
-
-        tracker.checkpoint("After builder setup")
-
-        print("Building LEANN HNSW index...")
-
-        for chunk_text in all_texts:
-            builder.add_text(chunk_text)
-
-        builder.build_index(INDEX_PATH)
-        del builder
-        gc.collect()
-
-        tracker.checkpoint("After index building")
-
-    # Find existing LEANN index
-    index_paths = [
-        "./test_leann_comparison/comparison.leann",
-    ]
-    index_path = None
-    for path in index_paths:
-        if os.path.exists(path + ".meta.json"):
-            index_path = path
-            break
-
-    if not index_path:
-        print("❌ LEANN index not found. Please build it first")
-        return {"peak_memory": float("inf"), "error": "Index not found"}
-
-    # Measure runtime memory overhead
-    print("\nMeasuring runtime memory overhead...")
-    runtime_start_mem = get_memory_usage()
-    print(f"Before load memory: {runtime_start_mem:.1f} MB")
-    tracker.checkpoint("Before load memory")
-
-    # Load searcher
-    searcher = LeannSearcher(index_path)
-    tracker.checkpoint("After searcher loading")
-
-    print("Running search queries...")
-    queries = [
-        "什么是盘古大模型以及盘古开发过程中遇到了什么阴暗面,任务令一般在什么城市颁发",
-        "What is LEANN and how does it work?",
-        "华为诺亚方舟实验室的主要研究内容",
-    ]
-
-    for i, query in enumerate(queries):
-        start_time = time.time()
-        # Use same parameters as Faiss: top_k=20, ef=120 (complexity parameter)
-        _ = searcher.search(query, top_k=20, ef=120)
-        query_time = time.time() - start_time
-        print(f"Query {i + 1} time: {query_time:.3f}s")
-        tracker.checkpoint(f"After query {i + 1}")
-
-    runtime_end_mem = get_memory_usage()
-    runtime_overhead = runtime_end_mem - runtime_start_mem
-
-    peak_memory = tracker.summary()
-    print(f"Runtime Memory Overhead: {runtime_overhead:.1f} MB")
-
-    # Get storage size before cleanup
-    storage_size = 0
-    INDEX_DIR = Path(index_path).parent
-    if INDEX_DIR.exists():
-        total_size = 0
-        for dirpath, _, filenames in os.walk(str(INDEX_DIR)):
-            for filename in filenames:
-                # Only count actual index files, skip text data and backups
-                if filename.endswith((".old", ".tmp", ".bak", ".jsonl", ".json")):
-                    continue
-                # Count .index, .idx, .map files (actual index structures)
-                if filename.endswith((".index", ".idx", ".map")):
-                    filepath = os.path.join(dirpath, filename)
-                    total_size += os.path.getsize(filepath)
-        storage_size = total_size / (1024 * 1024)  # Convert to MB
-
-    # Clean up
-    del searcher
-    gc.collect()
-
-    return {
-        "peak_memory": peak_memory,
-        "storage_size": storage_size,
-    }
-
-
-def main():
-    """Run comparison tests"""
-    print("Storage + Search Memory Comparison: Faiss HNSW vs LEANN HNSW")
-    print("=" * 60)
-
-    # Test Faiss HNSW
-    faiss_results = test_faiss_hnsw()
-
-    # Force garbage collection
-    gc.collect()
-    time.sleep(2)
-
-    # Test LEANN HNSW
-    leann_results = test_leann_hnsw()
-
-    # Final comparison
-    print("\n" + "=" * 60)
-    print("STORAGE + SEARCH MEMORY COMPARISON")
-    print("=" * 60)
-
-    # Get storage sizes
-    faiss_storage_size = 0
-    leann_storage_size = leann_results.get("storage_size", 0)
-
-    # Get Faiss storage size using Python
-    if os.path.exists("./storage_faiss"):
-        total_size = 0
-        for dirpath, _, filenames in os.walk("./storage_faiss"):
-            for filename in filenames:
-                filepath = os.path.join(dirpath, filename)
-                total_size += os.path.getsize(filepath)
-        faiss_storage_size = total_size / (1024 * 1024)  # Convert to MB
-
-    print("Faiss HNSW:")
-    if "error" in faiss_results:
-        print(f"  ❌ Failed: {faiss_results['error']}")
-    else:
-        print(f"  Search Memory: {faiss_results['peak_memory']:.1f} MB")
-        print(f"  Storage Size: {faiss_storage_size:.1f} MB")
-
-    print("\nLEANN HNSW:")
-    if "error" in leann_results:
-        print(f"  ❌ Failed: {leann_results['error']}")
-    else:
-        print(f"  Search Memory: {leann_results['peak_memory']:.1f} MB")
-        print(f"  Storage Size: {leann_storage_size:.1f} MB")
-
-    # Calculate improvements only if both tests succeeded
-    if "error" not in faiss_results and "error" not in leann_results:
-        memory_ratio = faiss_results["peak_memory"] / leann_results["peak_memory"]
-
-        print("\nLEANN vs Faiss Performance:")
-        memory_saving = faiss_results["peak_memory"] - leann_results["peak_memory"]
-        print(f"  Search Memory: {memory_ratio:.1f}x less ({memory_saving:.1f} MB saved)")
-
-        # Storage comparison
-        if leann_storage_size > faiss_storage_size:
-            storage_ratio = leann_storage_size / faiss_storage_size
-            print(f"  Storage Size: {storage_ratio:.1f}x larger (LEANN uses more storage)")
-        elif faiss_storage_size > leann_storage_size:
-            storage_ratio = faiss_storage_size / leann_storage_size
-            print(f"  Storage Size: {storage_ratio:.1f}x smaller (LEANN uses less storage)")
-        else:
-            print("  Storage Size: similar")
-    else:
-        if "error" not in leann_results:
-            print("\n✅ LEANN HNSW completed successfully!")
-            print(f"📊 Search Memory: {leann_results['peak_memory']:.1f} MB")
-            print(f"📊 Storage Size: {leann_storage_size:.1f} MB")
-        if "error" not in faiss_results:
-            print("\n✅ Faiss HNSW completed successfully!")
-            print(f"📊 Search Memory: {faiss_results['peak_memory']:.1f} MB")
-            print(f"📊 Storage Size: {faiss_storage_size:.1f} MB")
-
-
-if __name__ == "__main__":
-    main()

benchmarks/data/README.md

@@ -1,44 +0,0 @@
----
-license: mit
----
-
-# LEANN-RAG Evaluation Data
-
-This repository contains the necessary data to run the recall evaluation scripts for the [LEANN-RAG](https://huggingface.co/LEANN-RAG) project.
-
-## Dataset Components
-
-This dataset is structured into three main parts:
-
-1.  **Pre-built LEANN Indices**:
-    *   `dpr/`: A pre-built index for the DPR dataset.
-    *   `rpj_wiki/`: A pre-built index for the RPJ-Wiki dataset.
-    These indices were created using the `leann-core` library and are required by the `LeannSearcher`.
-
-2.  **Ground Truth Data**:
-    *   `ground_truth/`: Contains the ground truth files (`flat_results_nq_k3.json`) for both the DPR and RPJ-Wiki datasets. These files map queries to the original passage IDs from the Natural Questions benchmark, evaluated using the Contriever model.
-
-3.  **Queries**:
-    *   `queries/`: Contains the `nq_open.jsonl` file with the Natural Questions queries used for the evaluation.
-
-## Usage
-
-To use this data, you can download it locally using the `huggingface-hub` library. First, install the library:
-
-```bash
-pip install huggingface-hub
-```
-
-Then, you can download the entire dataset to a local directory (e.g., `data/`) with the following Python script:
-
-```python
-from huggingface_hub import snapshot_download
-
-snapshot_download(
-    repo_id="LEANN-RAG/leann-rag-evaluation-data",
-    repo_type="dataset",
-    local_dir="data"
-)
-```
-
-This will download all the necessary files into a local `data` folder, preserving the repository structure. The evaluation scripts in the main [LEANN-RAG Space](https://huggingface.co/LEANN-RAG) are configured to work with this data structure.

benchmarks/diskann_vs_hnsw_speed_comparison.py

@@ -1,286 +0,0 @@
-#!/usr/bin/env python3
-"""
-DiskANN vs HNSW Search Performance Comparison
-
-This benchmark compares search performance between DiskANN and HNSW backends:
-- DiskANN: With graph partitioning enabled (is_recompute=True)
-- HNSW: With recompute enabled (is_recompute=True)
-- Tests performance across different dataset sizes
-- Measures search latency, recall, and index size
-"""
-
-import gc
-import multiprocessing as mp
-import tempfile
-import time
-from pathlib import Path
-from typing import Any
-
-import numpy as np
-
-# Prefer 'fork' start method to avoid POSIX semaphore leaks on macOS
-try:
-    mp.set_start_method("fork", force=True)
-except Exception:
-    pass
-
-
-def create_test_texts(n_docs: int) -> list[str]:
-    """Create synthetic test documents for benchmarking."""
-    np.random.seed(42)
-    topics = [
-        "machine learning and artificial intelligence",
-        "natural language processing and text analysis",
-        "computer vision and image recognition",
-        "data science and statistical analysis",
-        "deep learning and neural networks",
-        "information retrieval and search engines",
-        "database systems and data management",
-        "software engineering and programming",
-        "cybersecurity and network protection",
-        "cloud computing and distributed systems",
-    ]
-
-    texts = []
-    for i in range(n_docs):
-        topic = topics[i % len(topics)]
-        variation = np.random.randint(1, 100)
-        text = (
-            f"This is document {i} about {topic}. Content variation {variation}. "
-            f"Additional information about {topic} with details and examples. "
-            f"Technical discussion of {topic} including implementation aspects."
-        )
-        texts.append(text)
-
-    return texts
-
-
-def benchmark_backend(
-    backend_name: str, texts: list[str], test_queries: list[str], backend_kwargs: dict[str, Any]
-) -> dict[str, float]:
-    """Benchmark a specific backend with the given configuration."""
-    from leann.api import LeannBuilder, LeannSearcher
-
-    print(f"\n🔧 Testing {backend_name.upper()} backend...")
-
-    with tempfile.TemporaryDirectory() as temp_dir:
-        index_path = str(Path(temp_dir) / f"benchmark_{backend_name}.leann")
-
-        # Build index
-        print(f"📦 Building {backend_name} index with {len(texts)} documents...")
-        start_time = time.time()
-
-        builder = LeannBuilder(
-            backend_name=backend_name,
-            embedding_model="facebook/contriever",
-            embedding_mode="sentence-transformers",
-            **backend_kwargs,
-        )
-
-        for text in texts:
-            builder.add_text(text)
-
-        builder.build_index(index_path)
-        build_time = time.time() - start_time
-
-        # Measure index size
-        index_dir = Path(index_path).parent
-        index_files = list(index_dir.glob(f"{Path(index_path).stem}.*"))
-        total_size = sum(f.stat().st_size for f in index_files if f.is_file())
-        size_mb = total_size / (1024 * 1024)
-
-        print(f"   ✅ Build completed in {build_time:.2f}s, index size: {size_mb:.1f}MB")
-
-        # Search benchmark
-        print("🔍 Running search benchmark...")
-        searcher = LeannSearcher(index_path)
-
-        search_times = []
-        all_results = []
-
-        for query in test_queries:
-            start_time = time.time()
-            results = searcher.search(query, top_k=5)
-            search_time = time.time() - start_time
-            search_times.append(search_time)
-            all_results.append(results)
-
-        avg_search_time = np.mean(search_times) * 1000  # Convert to ms
-        print(f"   ✅ Average search time: {avg_search_time:.1f}ms")
-
-        # Check for valid scores (detect -inf issues)
-        all_scores = [
-            result.score
-            for results in all_results
-            for result in results
-            if result.score is not None
-        ]
-        valid_scores = [
-            score for score in all_scores if score != float("-inf") and score != float("inf")
-        ]
-        score_validity_rate = len(valid_scores) / len(all_scores) if all_scores else 0
-
-        # Clean up (ensure embedding server shutdown and object GC)
-        try:
-            if hasattr(searcher, "cleanup"):
-                searcher.cleanup()
-            del searcher
-            del builder
-            gc.collect()
-        except Exception as e:
-            print(f"⚠️  Warning: Resource cleanup error: {e}")
-
-        return {
-            "build_time": build_time,
-            "avg_search_time_ms": avg_search_time,
-            "index_size_mb": size_mb,
-            "score_validity_rate": score_validity_rate,
-        }
-
-
-def run_comparison(n_docs: int = 500, n_queries: int = 10):
-    """Run performance comparison between DiskANN and HNSW."""
-    print("🚀 Starting DiskANN vs HNSW Performance Comparison")
-    print(f"📊 Dataset: {n_docs} documents, {n_queries} test queries")
-
-    # Create test data
-    texts = create_test_texts(n_docs)
-    test_queries = [
-        "machine learning algorithms",
-        "natural language processing",
-        "computer vision techniques",
-        "data analysis methods",
-        "neural network architectures",
-        "database query optimization",
-        "software development practices",
-        "security vulnerabilities",
-        "cloud infrastructure",
-        "distributed computing",
-    ][:n_queries]
-
-    # HNSW benchmark
-    hnsw_results = benchmark_backend(
-        backend_name="hnsw",
-        texts=texts,
-        test_queries=test_queries,
-        backend_kwargs={
-            "is_recompute": True,  # Enable recompute for fair comparison
-            "M": 16,
-            "efConstruction": 200,
-        },
-    )
-
-    # DiskANN benchmark
-    diskann_results = benchmark_backend(
-        backend_name="diskann",
-        texts=texts,
-        test_queries=test_queries,
-        backend_kwargs={
-            "is_recompute": True,  # Enable graph partitioning
-            "num_neighbors": 32,
-            "search_list_size": 50,
-        },
-    )
-
-    # Performance comparison
-    print("\n📈 Performance Comparison Results")
-    print(f"{'=' * 60}")
-    print(f"{'Metric':<25} {'HNSW':<15} {'DiskANN':<15} {'Speedup':<10}")
-    print(f"{'-' * 60}")
-
-    # Build time comparison
-    build_speedup = hnsw_results["build_time"] / diskann_results["build_time"]
-    print(
-        f"{'Build Time (s)':<25} {hnsw_results['build_time']:<15.2f} {diskann_results['build_time']:<15.2f} {build_speedup:<10.2f}x"
-    )
-
-    # Search time comparison
-    search_speedup = hnsw_results["avg_search_time_ms"] / diskann_results["avg_search_time_ms"]
-    print(
-        f"{'Search Time (ms)':<25} {hnsw_results['avg_search_time_ms']:<15.1f} {diskann_results['avg_search_time_ms']:<15.1f} {search_speedup:<10.2f}x"
-    )
-
-    # Index size comparison
-    size_ratio = diskann_results["index_size_mb"] / hnsw_results["index_size_mb"]
-    print(
-        f"{'Index Size (MB)':<25} {hnsw_results['index_size_mb']:<15.1f} {diskann_results['index_size_mb']:<15.1f} {size_ratio:<10.2f}x"
-    )
-
-    # Score validity
-    print(
-        f"{'Score Validity (%)':<25} {hnsw_results['score_validity_rate'] * 100:<15.1f} {diskann_results['score_validity_rate'] * 100:<15.1f}"
-    )
-
-    print(f"{'=' * 60}")
-    print("\n🎯 Summary:")
-    if search_speedup > 1:
-        print(f"   DiskANN is {search_speedup:.2f}x faster than HNSW for search")
-    else:
-        print(f"   HNSW is {1 / search_speedup:.2f}x faster than DiskANN for search")
-
-    if size_ratio > 1:
-        print(f"   DiskANN uses {size_ratio:.2f}x more storage than HNSW")
-    else:
-        print(f"   DiskANN uses {1 / size_ratio:.2f}x less storage than HNSW")
-
-    print(
-        f"   Both backends achieved {min(hnsw_results['score_validity_rate'], diskann_results['score_validity_rate']) * 100:.1f}% score validity"
-    )
-
-
-if __name__ == "__main__":
-    import sys
-
-    try:
-        # Handle help request
-        if len(sys.argv) > 1 and sys.argv[1] in ["-h", "--help", "help"]:
-            print("DiskANN vs HNSW Performance Comparison")
-            print("=" * 50)
-            print(f"Usage: python {sys.argv[0]} [n_docs] [n_queries]")
-            print()
-            print("Arguments:")
-            print("  n_docs      Number of documents to index (default: 500)")
-            print("  n_queries   Number of test queries to run (default: 10)")
-            print()
-            print("Examples:")
-            print("  python benchmarks/diskann_vs_hnsw_speed_comparison.py")
-            print("  python benchmarks/diskann_vs_hnsw_speed_comparison.py 1000")
-            print("  python benchmarks/diskann_vs_hnsw_speed_comparison.py 2000 20")
-            sys.exit(0)
-
-        # Parse command line arguments
-        n_docs = int(sys.argv[1]) if len(sys.argv) > 1 else 500
-        n_queries = int(sys.argv[2]) if len(sys.argv) > 2 else 10
-
-        print("DiskANN vs HNSW Performance Comparison")
-        print("=" * 50)
-        print(f"Dataset: {n_docs} documents, {n_queries} queries")
-        print()
-
-        run_comparison(n_docs=n_docs, n_queries=n_queries)
-
-    except KeyboardInterrupt:
-        print("\n⚠️  Benchmark interrupted by user")
-        sys.exit(130)
-    except Exception as e:
-        print(f"\n❌ Benchmark failed: {e}")
-        sys.exit(1)
-    finally:
-        # Ensure clean exit (forceful to prevent rare hangs from atexit/threads)
-        try:
-            gc.collect()
-            print("\n🧹 Cleanup completed")
-            # Flush stdio to ensure message is visible before hard-exit
-            try:
-                import sys as _sys
-
-                _sys.stdout.flush()
-                _sys.stderr.flush()
-            except Exception:
-                pass
-        except Exception:
-            pass
-        # Use os._exit to bypass atexit handlers that may hang in rare cases
-        import os as _os
-
-        _os._exit(0)

benchmarks/faiss_only.py

@@ -1,151 +0,0 @@
-#!/usr/bin/env python3
-"""Test only Faiss HNSW"""
-
-import os
-import sys
-import time
-
-import psutil
-
-
-def get_memory_usage():
-    process = psutil.Process()
-    return process.memory_info().rss / 1024 / 1024
-
-
-class MemoryTracker:
-    def __init__(self, name: str):
-        self.name = name
-        self.start_mem = get_memory_usage()
-        self.stages = []
-
-    def checkpoint(self, stage: str):
-        current_mem = get_memory_usage()
-        diff = current_mem - self.start_mem
-        print(f"[{self.name} - {stage}] Memory: {current_mem:.1f} MB (+{diff:.1f} MB)")
-        self.stages.append((stage, current_mem))
-        return current_mem
-
-    def summary(self):
-        peak_mem = max(mem for _, mem in self.stages)
-        print(f"Peak Memory: {peak_mem:.1f} MB")
-        return peak_mem
-
-
-def main():
-    try:
-        import faiss
-    except ImportError:
-        print("Faiss is not installed.")
-        print(
-            "Please install it with `uv pip install faiss-cpu` and you can  then run this script again"
-        )
-        sys.exit(1)
-
-    from llama_index.core import (
-        Settings,
-        SimpleDirectoryReader,
-        StorageContext,
-        VectorStoreIndex,
-    )
-    from llama_index.core.node_parser import SentenceSplitter
-    from llama_index.embeddings.huggingface import HuggingFaceEmbedding
-    from llama_index.vector_stores.faiss import FaissVectorStore
-
-    tracker = MemoryTracker("Faiss HNSW")
-    tracker.checkpoint("Initial")
-
-    embed_model = HuggingFaceEmbedding(model_name="facebook/contriever")
-    Settings.embed_model = embed_model
-    tracker.checkpoint("After embedding model setup")
-
-    d = 768
-    faiss_index = faiss.IndexHNSWFlat(d, 32)
-    faiss_index.hnsw.efConstruction = 64
-    tracker.checkpoint("After Faiss index creation")
-
-    documents = SimpleDirectoryReader(
-        "data",
-        recursive=True,
-        encoding="utf-8",
-        required_exts=[".pdf", ".txt", ".md"],
-    ).load_data()
-    tracker.checkpoint("After document loading")
-
-    # Parse into chunks using the same splitter as LEANN
-    node_parser = SentenceSplitter(
-        chunk_size=256, chunk_overlap=20, separator=" ", paragraph_separator="\n\n"
-    )
-
-    tracker.checkpoint("After text splitter setup")
-
-    # Check if index already exists and try to load it
-    index_loaded = False
-    if os.path.exists("./storage_faiss"):
-        print("Loading existing Faiss HNSW index...")
-        try:
-            # Use the correct Faiss loading pattern from the example
-            vector_store = FaissVectorStore.from_persist_dir("./storage_faiss")
-            storage_context = StorageContext.from_defaults(
-                vector_store=vector_store, persist_dir="./storage_faiss"
-            )
-            from llama_index.core import load_index_from_storage
-
-            index = load_index_from_storage(storage_context=storage_context)
-            print("Index loaded from ./storage_faiss")
-            tracker.checkpoint("After loading existing index")
-            index_loaded = True
-        except Exception as e:
-            print(f"Failed to load existing index: {e}")
-            print("Cleaning up corrupted index and building new one...")
-            # Clean up corrupted index
-            import shutil
-
-            if os.path.exists("./storage_faiss"):
-                shutil.rmtree("./storage_faiss")
-
-    if not index_loaded:
-        print("Building new Faiss HNSW index...")
-
-        # Use the correct Faiss building pattern from the example
-        vector_store = FaissVectorStore(faiss_index=faiss_index)
-        storage_context = StorageContext.from_defaults(vector_store=vector_store)
-        index = VectorStoreIndex.from_documents(
-            documents, storage_context=storage_context, transformations=[node_parser]
-        )
-        tracker.checkpoint("After index building")
-
-        # Save index to disk using the correct pattern
-        index.storage_context.persist(persist_dir="./storage_faiss")
-        tracker.checkpoint("After index saving")
-
-    # Measure runtime memory overhead
-    print("\nMeasuring runtime memory overhead...")
-    runtime_start_mem = get_memory_usage()
-    print(f"Before load memory: {runtime_start_mem:.1f} MB")
-    tracker.checkpoint("Before load memory")
-
-    query_engine = index.as_query_engine(similarity_top_k=20)
-    queries = [
-        "什么是盘古大模型以及盘古开发过程中遇到了什么阴暗面,任务令一般在什么城市颁发",
-        "What is LEANN and how does it work?",
-        "华为诺亚方舟实验室的主要研究内容",
-    ]
-
-    for i, query in enumerate(queries):
-        start_time = time.time()
-        _ = query_engine.query(query)
-        query_time = time.time() - start_time
-        print(f"Query {i + 1} time: {query_time:.3f}s")
-        tracker.checkpoint(f"After query {i + 1}")
-
-    runtime_end_mem = get_memory_usage()
-    runtime_overhead = runtime_end_mem - runtime_start_mem
-
-    peak_memory = tracker.summary()
-    print(f"Peak Memory: {peak_memory:.1f} MB")
-    print(f"Runtime Memory Overhead: {runtime_overhead:.1f} MB")
-
-
-if __name__ == "__main__":
-    main()

benchmarks/run_evaluation.py

@@ -1,393 +0,0 @@
-#!/usr/bin/env python3
-"""
-This script runs a recall evaluation on a given LEANN index.
-It correctly compares results by fetching the text content for both the new search
-results and the golden standard results, making the comparison robust to ID changes.
-"""
-
-import argparse
-import json
-import sys
-import time
-from pathlib import Path
-
-import numpy as np
-from leann.api import LeannBuilder, LeannChat, LeannSearcher
-
-
-def download_data_if_needed(data_root: Path, download_embeddings: bool = False):
-    """Checks if the data directory exists, and if not, downloads it from HF Hub."""
-    if not data_root.exists():
-        print(f"Data directory '{data_root}' not found.")
-        print("Downloading evaluation data from Hugging Face Hub... (this may take a moment)")
-        try:
-            from huggingface_hub import snapshot_download
-
-            if download_embeddings:
-                # Download everything including embeddings (large files)
-                snapshot_download(
-                    repo_id="LEANN-RAG/leann-rag-evaluation-data",
-                    repo_type="dataset",
-                    local_dir=data_root,
-                    local_dir_use_symlinks=False,
-                )
-                print("Data download complete (including embeddings)!")
-            else:
-                # Download only specific folders, excluding embeddings
-                allow_patterns = [
-                    "ground_truth/**",
-                    "indices/**",
-                    "queries/**",
-                    "*.md",
-                    "*.txt",
-                ]
-                snapshot_download(
-                    repo_id="LEANN-RAG/leann-rag-evaluation-data",
-                    repo_type="dataset",
-                    local_dir=data_root,
-                    local_dir_use_symlinks=False,
-                    allow_patterns=allow_patterns,
-                )
-                print("Data download complete (excluding embeddings)!")
-        except ImportError:
-            print(
-                "Error: huggingface_hub is not installed. Please install it to download the data:"
-            )
-            print("uv pip install -e '.[dev]'")
-            sys.exit(1)
-        except Exception as e:
-            print(f"An error occurred during data download: {e}")
-            sys.exit(1)
-
-
-def download_embeddings_if_needed(data_root: Path, dataset_type: str | None = None):
-    """Download embeddings files specifically."""
-    embeddings_dir = data_root / "embeddings"
-
-    if dataset_type:
-        # Check if specific dataset embeddings exist
-        target_file = embeddings_dir / dataset_type / "passages_00.pkl"
-        if target_file.exists():
-            print(f"Embeddings for {dataset_type} already exist")
-            return str(target_file)
-
-    print("Downloading embeddings from HuggingFace Hub...")
-    try:
-        from huggingface_hub import snapshot_download
-
-        # Download only embeddings folder
-        snapshot_download(
-            repo_id="LEANN-RAG/leann-rag-evaluation-data",
-            repo_type="dataset",
-            local_dir=data_root,
-            local_dir_use_symlinks=False,
-            allow_patterns=["embeddings/**/*.pkl"],
-        )
-        print("Embeddings download complete!")
-
-        if dataset_type:
-            target_file = embeddings_dir / dataset_type / "passages_00.pkl"
-            if target_file.exists():
-                return str(target_file)
-
-        return str(embeddings_dir)
-
-    except Exception as e:
-        print(f"Error downloading embeddings: {e}")
-        sys.exit(1)
-
-
-# --- Helper Function to get Golden Passages ---
-def get_golden_texts(searcher: LeannSearcher, golden_ids: list[int]) -> set:
-    """
-    Retrieves the text for golden passage IDs directly from the LeannSearcher's
-    passage manager.
-    """
-    golden_texts = set()
-    for gid in golden_ids:
-        try:
-            # PassageManager uses string IDs
-            passage_data = searcher.passage_manager.get_passage(str(gid))
-            golden_texts.add(passage_data["text"])
-        except KeyError:
-            print(f"Warning: Golden passage ID '{gid}' not found in the index's passage data.")
-    return golden_texts
-
-
-def load_queries(file_path: Path) -> list[str]:
-    queries = []
-    with open(file_path, encoding="utf-8") as f:
-        for line in f:
-            data = json.loads(line)
-            queries.append(data["query"])
-    return queries
-
-
-def build_index_from_embeddings(embeddings_file: str, output_path: str, backend: str = "hnsw"):
-    """
-    Build a LEANN index from pre-computed embeddings.
-
-    Args:
-        embeddings_file: Path to pickle file with (ids, embeddings) tuple
-        output_path: Path where to save the index
-        backend: Backend to use ("hnsw" or "diskann")
-    """
-    print(f"Building {backend} index from embeddings: {embeddings_file}")
-
-    # Create builder with appropriate parameters
-    if backend == "hnsw":
-        builder_kwargs = {
-            "M": 32,  # Graph degree
-            "efConstruction": 256,  # Construction complexity
-            "is_compact": True,  # Use compact storage
-            "is_recompute": True,  # Enable pruning for better recall
-        }
-    elif backend == "diskann":
-        builder_kwargs = {
-            "complexity": 64,
-            "graph_degree": 32,
-            "search_memory_maximum": 8.0,  # GB
-            "build_memory_maximum": 16.0,  # GB
-        }
-    else:
-        builder_kwargs = {}
-
-    builder = LeannBuilder(
-        backend_name=backend,
-        embedding_model="facebook/contriever-msmarco",  # Model used to create embeddings
-        dimensions=768,  # Will be auto-detected from embeddings
-        **builder_kwargs,
-    )
-
-    # Build index from precomputed embeddings
-    builder.build_index_from_embeddings(output_path, embeddings_file)
-    print(f"Index saved to: {output_path}")
-    return output_path
-
-
-def main():
-    parser = argparse.ArgumentParser(description="Run recall evaluation on a LEANN index.")
-    parser.add_argument(
-        "index_path",
-        type=str,
-        nargs="?",
-        help="Path to the LEANN index to evaluate or build (optional).",
-    )
-    parser.add_argument(
-        "--mode",
-        choices=["evaluate", "build"],
-        default="evaluate",
-        help="Mode: 'evaluate' existing index or 'build' from embeddings",
-    )
-    parser.add_argument(
-        "--embeddings-file",
-        type=str,
-        help="Path to embeddings pickle file (optional for build mode)",
-    )
-    parser.add_argument(
-        "--backend",
-        choices=["hnsw", "diskann"],
-        default="hnsw",
-        help="Backend to use for building index (default: hnsw)",
-    )
-    parser.add_argument(
-        "--num-queries", type=int, default=10, help="Number of queries to evaluate."
-    )
-    parser.add_argument("--top-k", type=int, default=3, help="The 'k' value for recall@k.")
-    parser.add_argument(
-        "--ef-search", type=int, default=120, help="The 'efSearch' parameter for HNSW."
-    )
-    parser.add_argument(
-        "--batch-size",
-        type=int,
-        default=0,
-        help="Batch size for HNSW batched search (0 disables batching)",
-    )
-    parser.add_argument(
-        "--llm-type",
-        type=str,
-        choices=["ollama", "hf", "openai", "gemini", "simulated"],
-        default="ollama",
-        help="LLM backend type to optionally query during evaluation (default: ollama)",
-    )
-    parser.add_argument(
-        "--llm-model",
-        type=str,
-        default="qwen3:1.7b",
-        help="LLM model identifier for the chosen backend (default: qwen3:1.7b)",
-    )
-    args = parser.parse_args()
-
-    # --- Path Configuration ---
-    # Assumes a project structure where the script is in 'benchmarks/'
-    # and evaluation data is in 'benchmarks/data/'.
-    script_dir = Path(__file__).resolve().parent
-    data_root = script_dir / "data"
-
-    # Download data based on mode
-    if args.mode == "build":
-        # For building mode, we need embeddings
-        download_data_if_needed(data_root, download_embeddings=False)  # Basic data first
-
-        # Auto-detect dataset type and download embeddings
-        if args.embeddings_file:
-            embeddings_file = args.embeddings_file
-            # Try to detect dataset type from embeddings file path
-            if "rpj_wiki" in str(embeddings_file):
-                dataset_type = "rpj_wiki"
-            elif "dpr" in str(embeddings_file):
-                dataset_type = "dpr"
-            else:
-                dataset_type = "dpr"  # Default
-        else:
-            # Auto-detect from index path if provided, otherwise default to DPR
-            if args.index_path:
-                index_path_str = str(args.index_path)
-                if "rpj_wiki" in index_path_str:
-                    dataset_type = "rpj_wiki"
-                elif "dpr" in index_path_str:
-                    dataset_type = "dpr"
-                else:
-                    dataset_type = "dpr"  # Default to DPR
-            else:
-                dataset_type = "dpr"  # Default to DPR
-
-            embeddings_file = download_embeddings_if_needed(data_root, dataset_type)
-
-        # Auto-generate index path if not provided
-        if not args.index_path:
-            indices_dir = data_root / "indices" / dataset_type
-            indices_dir.mkdir(parents=True, exist_ok=True)
-            args.index_path = str(indices_dir / f"{dataset_type}_from_embeddings")
-            print(f"Auto-generated index path: {args.index_path}")
-
-        print(f"Building index from embeddings: {embeddings_file}")
-        built_index_path = build_index_from_embeddings(
-            embeddings_file, args.index_path, args.backend
-        )
-        print(f"Index built successfully: {built_index_path}")
-
-        # Ask if user wants to run evaluation
-        eval_response = input("Run evaluation on the built index? (y/n): ").strip().lower()
-        if eval_response != "y":
-            print("Index building complete. Exiting.")
-            return
-    else:
-        # For evaluation mode, don't need embeddings
-        download_data_if_needed(data_root, download_embeddings=False)
-
-        # Auto-detect index path if not provided
-        if not args.index_path:
-            # Default to using downloaded indices
-            indices_dir = data_root / "indices"
-
-            # Try common datasets in order of preference
-            for dataset in ["dpr", "rpj_wiki"]:
-                dataset_dir = indices_dir / dataset
-                if dataset_dir.exists():
-                    # Look for index files
-                    index_files = list(dataset_dir.glob("*.index")) + list(
-                        dataset_dir.glob("*_disk.index")
-                    )
-                    if index_files:
-                        args.index_path = str(
-                            index_files[0].with_suffix("")
-                        )  # Remove .index extension
-                        print(f"Using index: {args.index_path}")
-                        break
-
-            if not args.index_path:
-                print("No indices found. The data download should have included pre-built indices.")
-                print(
-                    "Please check the benchmarks/data/indices/ directory or provide --index-path manually."
-                )
-                sys.exit(1)
-
-    # Detect dataset type from index path to select the correct ground truth
-    index_path_str = str(args.index_path)
-    if "rpj_wiki" in index_path_str:
-        dataset_type = "rpj_wiki"
-    elif "dpr" in index_path_str:
-        dataset_type = "dpr"
-    else:
-        # Fallback: try to infer from the index directory name
-        dataset_type = Path(args.index_path).name
-        print(f"WARNING: Could not detect dataset type from path, inferred '{dataset_type}'.")
-
-    queries_file = data_root / "queries" / "nq_open.jsonl"
-    golden_results_file = data_root / "ground_truth" / dataset_type / "flat_results_nq_k3.json"
-
-    print(f"INFO: Detected dataset type: {dataset_type}")
-    print(f"INFO: Using queries file: {queries_file}")
-    print(f"INFO: Using ground truth file: {golden_results_file}")
-
-    try:
-        searcher = LeannSearcher(args.index_path)
-        queries = load_queries(queries_file)
-
-        with open(golden_results_file) as f:
-            golden_results_data = json.load(f)
-
-        num_eval_queries = min(args.num_queries, len(queries))
-        queries = queries[:num_eval_queries]
-
-        print(f"\nRunning evaluation on {num_eval_queries} queries...")
-        recall_scores = []
-        search_times = []
-
-        for i in range(num_eval_queries):
-            start_time = time.time()
-            new_results = searcher.search(
-                queries[i],
-                top_k=args.top_k,
-                complexity=args.ef_search,
-                batch_size=args.batch_size,
-            )
-            search_times.append(time.time() - start_time)
-
-            # Optional: also call the LLM with configurable backend/model (does not affect recall)
-            llm_config = {"type": args.llm_type, "model": args.llm_model}
-            chat = LeannChat(args.index_path, llm_config=llm_config, searcher=searcher)
-            answer = chat.ask(
-                queries[i],
-                top_k=args.top_k,
-                complexity=args.ef_search,
-                batch_size=args.batch_size,
-            )
-            print(f"Answer: {answer}")
-            # Correct Recall Calculation: Based on TEXT content
-            new_texts = {result.text for result in new_results}
-
-            # Get golden texts directly from the searcher's passage manager
-            golden_ids = golden_results_data["indices"][i][: args.top_k]
-            golden_texts = get_golden_texts(searcher, golden_ids)
-
-            overlap = len(new_texts & golden_texts)
-            recall = overlap / len(golden_texts) if golden_texts else 0
-            recall_scores.append(recall)
-
-            print("\n--- EVALUATION RESULTS ---")
-            print(f"Query: {queries[i]}")
-            print(f"New Results: {new_texts}")
-            print(f"Golden Results: {golden_texts}")
-            print(f"Overlap: {overlap}")
-            print(f"Recall: {recall}")
-            print(f"Search Time: {search_times[-1]:.4f}s")
-            print("--------------------------------")
-
-        avg_recall = np.mean(recall_scores) if recall_scores else 0
-        avg_time = np.mean(search_times) if search_times else 0
-
-        print("\n🎉 --- Evaluation Complete ---")
-        print(f"Avg. Recall@{args.top_k} (efSearch={args.ef_search}): {avg_recall:.4f}")
-        print(f"Avg. Search Time: {avg_time:.4f}s")
-
-    except Exception as e:
-        print(f"\n❌ An error occurred during evaluation: {e}")
-        import traceback
-
-        traceback.print_exc()
-
-
-if __name__ == "__main__":
-    main()

benchmarks/simple_mac_tpt_test.py

@@ -1,317 +0,0 @@
-import time
-from dataclasses import dataclass
-
-import numpy as np
-import torch
-from torch import nn
-from tqdm import tqdm
-from transformers import AutoModel
-
-# Add MLX imports
-try:
-    import mlx.core as mx
-    from mlx_lm.utils import load
-
-    MLX_AVAILABLE = True
-except ImportError:
-    print("MLX not available. Install with: uv pip install mlx mlx-lm")
-    MLX_AVAILABLE = False
-
-
-@dataclass
-class BenchmarkConfig:
-    model_path: str = "facebook/contriever-msmarco"
-    batch_sizes: list[int] = None
-    seq_length: int = 256
-    num_runs: int = 5
-    use_fp16: bool = True
-    use_int4: bool = False
-    use_int8: bool = False
-    use_cuda_graphs: bool = False
-    use_flash_attention: bool = False
-    use_linear8bitlt: bool = False
-    use_mlx: bool = False  # New flag for MLX testing
-
-    def __post_init__(self):
-        if self.batch_sizes is None:
-            self.batch_sizes = [1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024]
-
-
-class MLXBenchmark:
-    """MLX-specific benchmark for embedding models"""
-
-    def __init__(self, config: BenchmarkConfig):
-        self.config = config
-        self.model, self.tokenizer = self._load_model()
-
-    def _load_model(self):
-        """Load MLX model and tokenizer following the API pattern"""
-        print(f"Loading MLX model from {self.config.model_path}...")
-        try:
-            model, tokenizer = load(self.config.model_path)
-            print("MLX model loaded successfully")
-            return model, tokenizer
-        except Exception as e:
-            print(f"Error loading MLX model: {e}")
-            raise
-
-    def _create_random_batch(self, batch_size: int):
-        """Create random input batches for MLX testing - same as PyTorch"""
-        return torch.randint(0, 1000, (batch_size, self.config.seq_length), dtype=torch.long)
-
-    def _run_inference(self, input_ids: torch.Tensor) -> float:
-        """Run MLX inference with same input as PyTorch"""
-        start_time = time.time()
-        try:
-            # Convert PyTorch tensor to MLX array
-            input_ids_mlx = mx.array(input_ids.numpy())
-
-            # Get embeddings
-            embeddings = self.model(input_ids_mlx)
-
-            # Mean pooling (following the API pattern)
-            pooled = embeddings.mean(axis=1)
-
-            # Convert to numpy (following the API pattern)
-            pooled_numpy = np.array(pooled.tolist(), dtype=np.float32)
-
-            # Force computation
-            _ = pooled_numpy.shape
-
-        except Exception as e:
-            print(f"MLX inference error: {e}")
-            return float("inf")
-        end_time = time.time()
-
-        return end_time - start_time
-
-    def run(self) -> dict[int, dict[str, float]]:
-        """Run the MLX benchmark across all batch sizes"""
-        results = {}
-
-        print(f"Starting MLX benchmark with model: {self.config.model_path}")
-        print(f"Testing batch sizes: {self.config.batch_sizes}")
-
-        for batch_size in self.config.batch_sizes:
-            print(f"\n=== Testing MLX batch size: {batch_size} ===")
-            times = []
-
-            # Create input batch (same as PyTorch)
-            input_ids = self._create_random_batch(batch_size)
-
-            # Warm up
-            print("Warming up...")
-            for _ in range(3):
-                try:
-                    self._run_inference(input_ids[:2])  # Warm up with smaller batch
-                except Exception as e:
-                    print(f"Warmup error: {e}")
-                    break
-
-            # Run benchmark
-            for _i in tqdm(range(self.config.num_runs), desc=f"MLX Batch size {batch_size}"):
-                try:
-                    elapsed_time = self._run_inference(input_ids)
-                    if elapsed_time != float("inf"):
-                        times.append(elapsed_time)
-                except Exception as e:
-                    print(f"Error during MLX inference: {e}")
-                    break
-
-            if not times:
-                print(f"Skipping batch size {batch_size} due to errors")
-                continue
-
-            # Calculate statistics
-            avg_time = np.mean(times)
-            std_time = np.std(times)
-            throughput = batch_size / avg_time
-
-            results[batch_size] = {
-                "avg_time": avg_time,
-                "std_time": std_time,
-                "throughput": throughput,
-                "min_time": np.min(times),
-                "max_time": np.max(times),
-            }
-
-            print(f"MLX Results for batch size {batch_size}:")
-            print(f"  Avg Time: {avg_time:.4f}s ± {std_time:.4f}s")
-            print(f"  Min Time: {np.min(times):.4f}s")
-            print(f"  Max Time: {np.max(times):.4f}s")
-            print(f"  Throughput: {throughput:.2f} sequences/second")
-
-        return results
-
-
-class Benchmark:
-    def __init__(self, config: BenchmarkConfig):
-        self.config = config
-        self.device = (
-            "cuda"
-            if torch.cuda.is_available()
-            else "mps"
-            if torch.backends.mps.is_available()
-            else "cpu"
-        )
-        self.model = self._load_model()
-
-    def _load_model(self) -> nn.Module:
-        print(f"Loading model from {self.config.model_path}...")
-
-        model = AutoModel.from_pretrained(self.config.model_path)
-        if self.config.use_fp16:
-            model = model.half()
-        model = torch.compile(model)
-        model = model.to(self.device)
-
-        model.eval()
-        return model
-
-    def _create_random_batch(self, batch_size: int) -> torch.Tensor:
-        return torch.randint(
-            0,
-            1000,
-            (batch_size, self.config.seq_length),
-            device=self.device,
-            dtype=torch.long,
-        )
-
-    def _run_inference(self, input_ids: torch.Tensor) -> float:
-        attention_mask = torch.ones_like(input_ids)
-        # print shape of input_ids and attention_mask
-        print(f"input_ids shape: {input_ids.shape}")
-        print(f"attention_mask shape: {attention_mask.shape}")
-        start_time = time.time()
-        with torch.no_grad():
-            self.model(input_ids=input_ids, attention_mask=attention_mask)
-        if torch.cuda.is_available():
-            torch.cuda.synchronize()
-        if torch.backends.mps.is_available():
-            torch.mps.synchronize()
-        end_time = time.time()
-
-        return end_time - start_time
-
-    def run(self) -> dict[int, dict[str, float]]:
-        results = {}
-
-        if torch.cuda.is_available():
-            torch.cuda.reset_peak_memory_stats()
-
-        for batch_size in self.config.batch_sizes:
-            print(f"\nTesting batch size: {batch_size}")
-            times = []
-
-            input_ids = self._create_random_batch(batch_size)
-
-            for _i in tqdm(range(self.config.num_runs), desc=f"Batch size {batch_size}"):
-                try:
-                    elapsed_time = self._run_inference(input_ids)
-                    times.append(elapsed_time)
-                except Exception as e:
-                    print(f"Error during inference: {e}")
-                    break
-
-            if not times:
-                continue
-
-            avg_time = np.mean(times)
-            std_time = np.std(times)
-            throughput = batch_size / avg_time
-
-            results[batch_size] = {
-                "avg_time": avg_time,
-                "std_time": std_time,
-                "throughput": throughput,
-            }
-
-            print(f"Avg Time: {avg_time:.4f}s ± {std_time:.4f}s")
-            print(f"Throughput: {throughput:.2f} sequences/second")
-
-        if torch.cuda.is_available():
-            peak_memory_gb = torch.cuda.max_memory_allocated() / (1024**3)
-        else:
-            peak_memory_gb = 0.0
-
-        for batch_size in results:
-            results[batch_size]["peak_memory_gb"] = peak_memory_gb
-
-        return results
-
-
-def run_benchmark():
-    """Main function to run the benchmark with optimized parameters."""
-    config = BenchmarkConfig()
-
-    try:
-        benchmark = Benchmark(config)
-        results = benchmark.run()
-
-        max_throughput = max(results[batch_size]["throughput"] for batch_size in results)
-        avg_throughput = np.mean([results[batch_size]["throughput"] for batch_size in results])
-
-        return {
-            "max_throughput": max_throughput,
-            "avg_throughput": avg_throughput,
-            "results": results,
-        }
-
-    except Exception as e:
-        print(f"Benchmark failed: {e}")
-        return {"max_throughput": 0.0, "avg_throughput": 0.0, "error": str(e)}
-
-
-def run_mlx_benchmark():
-    """Run MLX-specific benchmark"""
-    if not MLX_AVAILABLE:
-        print("MLX not available, skipping MLX benchmark")
-        return {
-            "max_throughput": 0.0,
-            "avg_throughput": 0.0,
-            "error": "MLX not available",
-        }
-
-    config = BenchmarkConfig(model_path="mlx-community/all-MiniLM-L6-v2-4bit", use_mlx=True)
-
-    try:
-        benchmark = MLXBenchmark(config)
-        results = benchmark.run()
-
-        if not results:
-            return {
-                "max_throughput": 0.0,
-                "avg_throughput": 0.0,
-                "error": "No valid results",
-            }
-
-        max_throughput = max(results[batch_size]["throughput"] for batch_size in results)
-        avg_throughput = np.mean([results[batch_size]["throughput"] for batch_size in results])
-
-        return {
-            "max_throughput": max_throughput,
-            "avg_throughput": avg_throughput,
-            "results": results,
-        }
-
-    except Exception as e:
-        print(f"MLX benchmark failed: {e}")
-        return {"max_throughput": 0.0, "avg_throughput": 0.0, "error": str(e)}
-
-
-if __name__ == "__main__":
-    print("=== PyTorch Benchmark ===")
-    pytorch_result = run_benchmark()
-    print(f"PyTorch Max throughput: {pytorch_result['max_throughput']:.2f} sequences/second")
-    print(f"PyTorch Average throughput: {pytorch_result['avg_throughput']:.2f} sequences/second")
-
-    print("\n=== MLX Benchmark ===")
-    mlx_result = run_mlx_benchmark()
-    print(f"MLX Max throughput: {mlx_result['max_throughput']:.2f} sequences/second")
-    print(f"MLX Average throughput: {mlx_result['avg_throughput']:.2f} sequences/second")
-
-    # Compare results
-    if pytorch_result["max_throughput"] > 0 and mlx_result["max_throughput"] > 0:
-        speedup = mlx_result["max_throughput"] / pytorch_result["max_throughput"]
-        print("\n=== Comparison ===")
-        print(f"MLX is {speedup:.2f}x {'faster' if speedup > 1 else 'slower'} than PyTorch")

data/huawei_pangu.md

@@ -1,82 +0,0 @@
-# 盘古之殇：华为诺亚盘古大模型研发历程的心酸与黑暗
-
-各位好，
-
-我是一名盘古大模型团队，华为诺亚方舟实验室的员工。
-
-首先为自证身份，列举一些细节：
-
-1. 现诺亚主任，前算法应用部部长，后改名为小模型实验室的主任王云鹤。前诺亚主任：姚骏（大家称姚老师）。几个实验室主任：唐睿明（明哥，明队，已离职），尚利峰，张维（维哥），郝建业（郝老师），刘武龙（称呼为武龙所）等。其他骨干成员和专家陆续有很多人离职。
-2. 我们隶属于“四野”这个组织。四野下属有许多纵队，基础语言大模型是四纵。王云鹤的小模型是十六纵队。我们参加过苏州的集结，有各种月份的时间节点。在苏州攻关会颁发任务令，需要在节点前达成目标。苏州集结会把各地的人员都集中在苏州研究所，平常住宾馆，比如在甪直的酒店，与家人孩子天各一方。
-3. 在苏州集结的时候周六默认上班，非常辛苦，不过周六有下午茶，有一次还有小龙虾。在苏州研究所的工位搬迁过一次，从一栋楼换到了另一栋。苏州研究所楼栋都是欧式装修，门口有大坡，里面景色很不错。去苏州集结一般至少要去一周，甚至更久，多的人甚至一两个月都回不了家。
-4. 诺亚曾经传说是研究型的，但是来了之后因为在四野做大模型项目，项目成员完全变成了交付型的，且充满了例会，评审，汇报。很多时候做实验都要申请。团队需要对接终端小艺，华为云，ICT等诸多业务线，交付压力不小。
-5. 诺亚研发的盘古模型早期内部代号叫做“盘古智子”，一开始只有内部需要申请试用的网页版，到后续迫于压力在welink上接入和公测开放。
-
-这些天发生关于质疑盘古大模型抄袭千问的事情闹的沸沸扬扬。作为一个盘古团队的成员，我最近夜夜辗转反侧，难以入眠。盘古的品牌受到如此大的影响，一方面，我自私的为我的职业发展担忧，也为自己过去的努力工作感到不值。另一方面，由于有人开始揭露这些事情我内心又感到大快人心。在多少个日日夜夜，我们对内部某些人一次次靠着造假而又获得了无数利益的行为咬牙切齿而又无能为力。这种压抑和羞辱也逐渐消磨了我对华为的感情，让我在这里的时日逐渐浑浑噩噩，迷茫无措，时常怀疑自己的人生和自我价值。
-
-我承认我是一个懦弱的人，作为一个小小的打工人，我不仅不敢和王云鹤等内部手眼通天的人做对，更不敢和华为这样的庞然大物做对。我很怕失去我的工作，毕竟我也有家人和孩子，所以我打心眼里很佩服揭露者。但是，看到内部还在试图洗地掩盖事实，蒙蔽公众的时候，我实在不能容忍了。我也希望勇敢一次，顺从自己本心。就算自损八百，我也希望能伤敌一千。我决定把我在这里的所见所闻（部分来自于同事口述）公布出来，关于盘古大模型的“传奇故事”：
-
-华为确实主要在昇腾卡上训练大模型（小模型实验室有不少英伟达的卡，他们之前也会用来训练，后面转移到昇腾）。曾经我被华为“打造世界第二选择”的决心而折服，我本身也曾经对华为有深厚的感情。我们陪着昇腾一步步摸爬滚打，从充满bug到现在能训出模型，付出了巨大的心血和代价。
-
-最初我们的算力非常有限，在910A上训练模型。那会只支持fp16，训练的稳定性远不如bf16。盘古的moe开始很早，23年就主要是训练38Bmoe模型和后续的71B dense模型。71B的dense模型通过扩增变成了第一代的135Bdense模型，后面主力模型也逐渐在910B上训练。
-
-71B和135B模型都有一个巨大的硬伤就是tokenizer。当时使用的tokenizer编码效率极低，每个单个的符号，数字，空格，乃至汉字都会占用一个token。可想而知这会非常浪费算力，且使得模型的效果很差。这时候小模型实验室正好有个自己训的词表。姚老师当时怀疑是不是模型的tokenizer不好（虽然事后来看，他的怀疑是无疑正确的），于是就决定，让71B和135B换tokenizer，因为小模型实验室曾经尝试过。团队缝合了两个tokenizer，开始了tokenizer的更换。71B模型的更换失败了，而135B因为采用了更精细的embedding初始化策略，续训了至少1T的数据后词表总算更换成功，但可想而知，效果并不会变好。
-
-于此同期，阿里和智谱等国内其他公司在GPU上训练，且已经摸索出了正确的方法，盘古和竞品的差距越来越大。内部一个230B从头训练的dense模型又因为各种原因训练失败，导致项目的状况几乎陷入绝境。面临几个节点的压力以及内部对盘古的强烈质疑时，团队的士气低迷到了极点。团队在算力极其有限的时候，做出了很多努力和挣扎。比如，团队偶然发现当时的38B moe并没有预期moe的效果。于是去掉了moe参数，还原为了13B的dense模型。由于38B的moe源自很早的pangu alpha 13B，架构相对落后，团队进行了一系列的操作，比如切换绝对位置编码到rope，去掉bias，切换为rmsnorm。同时鉴于tokenizer的一些失败和换词表的经验，这个模型的词表也更换为了王云鹤的小模型实验室7B模型所使用的词表。后面这个13B模型进行了扩增续训，变成了第二代38B dense模型（在几个月内这个模型都是主要的盘古中档位模型），曾经具有一定的竞争力。但是，由于更大的135B模型架构落后，且更换词表模型损伤巨大（后续分析发现当时更换的缝合词表有更严重的bug），续训后也与千问等当时国内领先模型存在很大差距。这时由于内部的质疑声和领导的压力也越来越大。团队的状态几乎陷入了绝境。
-
-在这种情况下，王云鹤和他的小模型实验室出手了。他们声称是从旧的135B参数继承改造而来，通过训练短短的几百B数据，各项指标平均提升了十个点左右。实际上，这就是他们套壳应用到大模型的第一次杰作。华为的外行领导内行，使得领导完全对于这种扯淡的事情没有概念，他们只会觉得肯定是有什么算法创新。经过内部的分析，他们实际上是使用Qwen 1.5 110B续训而来，通过加层，扩增ffn维度，添加盘古pi论文的一些机制得来，凑够了大概135B的参数。实际上，旧的135B有107层，而这个模型只有82层，各种配置也都不一样。新的来路不明的135B训练完很多参数的分布也和Qwen 110B几乎一模一样。连模型代码的类名当时都是Qwen，甚至懒得改名。后续这个模型就是所谓的135B V2。而这个模型当时也提供给了很多下游，甚至包括外部客户。
-
-这件事对于我们这些认真诚实做事的同事们带来了巨大的冲击，内部很多人其实都知道这件事，甚至包括终端和华为云。我们都戏称以后别叫盘古模型了，叫千古吧。当时团队成员就想向bcg举报了，毕竟这已经是重大的业务造假了。但是后面据说被领导拦了下来，因为更高级别的领导（比如姚老师，以及可能熊总和查老）其实后面也知道了，但是并不管，因为通过套壳拿出好的结果，对他们也是有利的。这件事使得当时团队几位最强的同事开始心灰意冷，离职跑路也逐渐成为挂在嘴边的事。
-
-此时，盘古似乎迎来了转机。由于前面所述的这些盘古模型基本都是续训和改造而来，当时诺亚完全没有掌握从头训练的技术，何况还是在昇腾的NPU上进行训练。在当时团队的核心成员的极力争取下，盘古开始了第三代模型的训练，付出了巨大的努力后，在数据架构和训练算法方面都与业界逐渐接轨，而这其中的艰辛和小模型实验室的人一点关系都没有。
-
-一开始团队成员毫无信心，只从一个13B的模型开始训练，但是后面发现效果还不错，于是这个模型后续再次进行了一次参数扩增，变成了第三代的38B，代号38B V3。想必很多产品线的兄弟都对这个模型很熟悉。当时这个模型的tokenizer是基于llama的词表进行扩展的（也是业界常见的做法）。而当时王云鹤的实验室做出来了另一个词表（也就是后续pangu系列的词表）。当时两个词表还被迫进行了一次赛马，最终没有明显的好坏结论。于是，领导当即决定，应该统一词表，使用王云鹤他们的。于是，在后续从头训练的135B V3（也就是对外的Pangu Ultra），便是采用了这个tokenizer。这也解释了很多使用我们模型的兄弟的疑惑，为什么当时同为V3代的两个不同档位的模型，会使用不同的tokenizer。
-
-
-我们打心眼里觉得，135B V3是我们四纵团队当时的骄傲。这是第一个真正意义上的，华为全栈自研，正经从头训练的千亿级别的模型，且效果与24年同期竞品可比的。写到这里我已经热泪盈眶，太不容易了。当时为了稳定训练，团队做了大量实验对比，并且多次在模型梯度出现异常的时候进行及时回退重启。这个模型真正做到了后面技术报告所说的训练全程没有一个loss spike。我们克服了不知道多少困难，我们做到了，我们愿用生命和荣誉保证这个模型训练的真实性。多少个凌晨，我们为了它的训练而不眠。在被内部心声骂的一文不值的时候，我们有多么不甘，有多少的委屈，我们挺住了。
-
-我们这帮人是真的在为打磨国产算力底座燃烧自己的青春啊……客居他乡，我们放弃了家庭，放弃了假期，放弃了健康，放弃了娱乐，抛头颅洒热血，其中的艰辛与困苦，寥寥数笔不足以概括其万一。在各种动员大会上，当时口号中喊出的盘古必胜，华为必胜，我们心里是真的深深被感动。
-
-然而，我们的所有辛苦的成果，经常被小模型实验室轻飘飘的拿走了。数据，直接要走。代码，直接要走，还要求我们配合适配到能一键运行。我们当时戏称小模型实验室为点鼠标实验室。我们付出辛苦，他们取得荣耀。果然应了那句话，你在负重前行是因为有人替你岁月静好。在这种情况下，越来越多的战友再也坚持不下去了，选择了离开。看到身边那些优秀的同事一个个离职，我的内心又感叹又难过。在这种作战一样的环境下，我们比起同事来说更像是战友。他们在技术上也有无数值得我学习的地方，堪称良师。看到他们去了诸如字节Seed，Deepseek，月之暗面，腾讯和快手等等很多出色的团队，我打心眼里为他们高兴和祝福，脱离了这个辛苦却肮脏的地方。我至今还对一位离职同事的话记忆犹新，ta说：“来这里是我技术生涯中的耻辱，在这里再呆每一天都是浪费生命”。话虽难听却让我无言以对。我担心我自己技术方面的积累不足，以及没法适应互联网公司高淘汰的环境，让我多次想离职的心始终没有迈出这一步。
-
-盘古除了dense模型，后续也启动了moe的探索。一开始训练的是一个224B的moe模型。而与之平行的，小模型实验室也开启了第二次主要的套壳行动（次要的插曲可能还包括一些别的模型，比如math模型），即这次流传甚广的pangu pro moe 72B。这个模型内部自称是从小模型实验室的7B扩增上来的（就算如此，这也与技术报告不符，何况是套壳qwen 2.5的14b续训）。还记得他们训了没几天，内部的评测就立刻追上了当时的38B V3。AI系统实验室很多兄弟因为需要适配模型，都知道他们的套壳行动，只是迫于各种原因，无法伸张正义。实际上，对于后续训了很久很久的这个模型，Honestagi能够分析出这个量级的相似性我已经很诧异了，因为这个模型为了续训洗参数，所付出的算力甚至早就足够从头训一个同档位的模型了。听同事说他们为了洗掉千问的水印，采取了不少办法，甚至包括故意训了脏数据。这也为学术界研究模型血缘提供了一个前所未有的特殊模范吧。以后新的血缘方法提出可以拿出来溜溜。
-
-24年底和25年初，在Deepseek v3和r1发布之后，由于其惊艳的技术水平，团队受到了巨大的冲击，也受到了更大的质疑。于是为了紧跟潮流，盘古模仿Deepseek的模型尺寸，开启了718B moe的训练。这个时候，小模型实验室再次出手了。他们选择了套壳Deepseekv3续训。他们通过冻住Deepseek加载的参数，进行训练。连任务加载ckpt的目录都是deepseekv3，改都不改，何其嚣张？与之相反，一些有真正技术信仰的同事，在从头训练另一个718B的moe。但其中出现了各种各样的问题。但是很显然，这个模型怎么可能比直接套壳的好呢？如果不是团队leader坚持，早就被叫停了。
-
-华为的流程管理之繁重，严重拖累了大模型的研发节奏，例如版本管理，模型血缘，各种流程化，各种可追溯。讽刺的是，小模型实验室的模型似乎从来不受这些流程的约束，想套壳就套壳，想续训就续训，算力源源不断的伸手拿走。这种强烈到近乎魔幻的对比，说明了当前流程管理的情况：只许州官放火，不许百姓点灯。何其可笑？何其可悲？何其可恶？何其可耻！
-
-HonestAGI的事情出来后，内部让大家不停的研讨分析，如何公关和“回应”。诚然，这个原文的分析也许不够有力，给了王云鹤与小模型实验室他们狡辩和颠倒黑白的机会。为此，这两天我内心感到作呕，时时怀疑自己的人生意义以及苍天无眼。我不奉陪了，我要离职了，同时我也在申请从盘古部分技术报告的作者名单中移除。曾经在这些技术报告上署名是我一生都无法抹除的污点。当时我没想到，他们竟然猖狂到敢开源。我没想到，他们敢如此愚弄世人，大肆宣发。当时，我也许是存了侥幸心理，没有拒绝署名。我相信很多扎实做事的战友，也只是被迫上了贼船，或者不知情。但这件事已经无法挽回，我希望我的余生能够坚持扎实做真正有意义的事，为我当时的软弱和不坚定赎罪。
-
-深夜写到这里，我已经泪流满面，泣不成声。还记得一些出色的同事离职时，我苦笑问他们要不要发个长长的心声惯例帖，揭露一下现状。对方说：不了，浪费时间，而且我也怕揭露出来你们过的更糟。我当时一下黯然神伤，因为曾经共同为了理想奋斗过的战友已经彻底对华为彻底灰心了。当时大家调侃，我们用着当年共产党的小米加步枪，组织却有着堪比当年国民党的作风。
-
-曾几何时，我为我们用着小米加步枪打败洋枪洋炮而自豪。
-
-现在，我累了，我想投降。
-
-其实时至今日，我还是真心希望华为能认真吸取教训，能做好盘古，把盘古做到世界一流，把昇腾变成英伟达的水平。内部的劣币驱逐良币，使得诺亚乃至华为在短时间内急剧流失了大量出色的大模型人才。相信他们也正在如Deepseek等各个团队闪耀着，施展着他们的抱负才华，为中美在AI的激烈竞赛中奉献力量。我时常感叹，华为不是没有人才，而是根本不知道怎么留住人才。如果给这些人合适的环境，合适的资源，更少的枷锁，更少的政治斗争，盘古何愁不成？
-
-最后：我以生命，人格和荣誉发誓，我写的以上所有内容均为真实（至少在我有限的认知范围内）。我没有那么高的技术水平以及机会去做详尽扎实的分析，也不敢直接用内部记录举证，怕因为信息安全抓到。但是我相信我很多曾经的战友，会为我作证。在华为内部的兄弟，包括我们曾经服务过的产品线兄弟们，相信本文的无数细节能和你们的印象对照，印证我的说法。你们可能也曾经被蒙骗，但这些残酷的真相不会被尘封。我们奋战过的痕迹，也不应该被扭曲和埋葬。
-
-写了这么多，某些人肯定想把我找出来，抹杀掉。公司搞不好也想让我噤声乃至追责。如果真的这样，我，乃至我的家人的人身乃至生命安全可能都会受到威胁。为了自我保护，我近期每天会跟大家报平安。
-
-如果我消失了，就当是我为了真理和理想，为了华为乃至中国能够更好地发展算力和AI而牺牲了吧，我愿埋葬于那片曾经奋斗过的地方。
-
-诺亚，再见
-
-2025年7月6日凌晨      写于深圳
-
----
-
-各位好，
-
-感谢大家的关心与祝福。我目前暂时安全，但公司应该在进行排查与某些名单收集，后续情况未知。
-
-我补充一些细节，以免某些人继续颠倒黑白。
-
-关于135B V2，小模型实验室在迅速地完成套壳并拿完所有套壳带来的好处后（比如任务令表彰和及时激励），因为不想继续支撑下游应用和模型迭代，又把这个烫手山芋甩给了四纵。确实技高一筹，直接把四纵的兄弟们拉下水。同事提供过去一个老旧的模型，最终拿回了一个当时一个魔改的先进的千问。做大模型的人，自己做的模型就像自己孩子一样熟悉，不要把别人都当傻子。就像自家儿子出门一趟，回来个别人家孩子。
-
-盘古report的署名是不符合学术规范的。例如，135B V3有不少有技术贡献的人，因为作者名额数量限制，劳动成果没有得到应有的回报，团队内曾经有不小的意见。这个模型当时是大家智慧和汗水的结晶，甚至是团队当时的精神支柱，支撑着不少兄弟们继续留在诺亚。所谓的名额限制，以及挂名了一些毫无技术贡献的人（如一些小模型实验室的人），让兄弟们何其心寒。
-
----
-
-暂时平安。另外，支持我勇于说出真相的战友们 https://github.com/HW-whistleblower/True-Story-of-Pangu/issues/317

demo.ipynb

@@ -1,116 +1,362 @@
 {
  "cells": [
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "# Quick Start \n",
-    "\n",
-    "**Home GitHub Repository:** [LEANN on GitHub](https://github.com/yichuan-w/LEANN)\n",
-    "\n",
-    "**Important for Colab users:** Set your runtime type to T4 GPU for optimal performance. Go to Runtime → Change runtime type → Hardware accelerator → T4 GPU."
-   ]
-  },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 1,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Initializing leann-backend-diskann...\n",
+      "INFO: Registering backend 'diskann'\n",
+      "INFO: DiskANN backend loaded successfully\n",
+      "INFO: LeannBuilder initialized with 'diskann' backend.\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/home/ubuntu/LEANN_clean/leann/.venv/lib/python3.11/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
+      "  from .autonotebook import tqdm as notebook_tqdm\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "INFO: Computing embeddings for 6 chunks using 'sentence-transformers/all-mpnet-base-v2'...\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Batches: 100%|██████████| 1/1 [00:00<00:00,  2.91it/s]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "INFO: Building DiskANN index for 6 vectors with metric Metric.INNER_PRODUCT...\n",
+      "Using Inner Product search, so need to pre-process base data into temp file. Please ensure there is additional (n*(d+1)*4) bytes for storing pre-processed base vectors, apart from the interim indices created by DiskANN and the final index.\n",
+      "Pre-processing base file by adding extra coordinate\n",
+      "✅ DiskANN index built successfully at 'knowledge'\n",
+      "Writing bin: knowledge_disk.index_max_base_norm.bin\n",
+      "bin: #pts = 1, #dims = 1, size = 12B\n",
+      "Finished writing bin.\n",
+      "Time for preprocessing data for inner product: 0.000172 seconds\n",
+      "Reading max_norm_of_base from knowledge_disk.index_max_base_norm.bin\n",
+      "Reading bin file knowledge_disk.index_max_base_norm.bin ...\n",
+      "Opening bin file knowledge_disk.index_max_base_norm.bin... \n",
+      "Metadata: #pts = 1, #dims = 1...\n",
+      "done.\n",
+      "max_norm_of_base: 1\n",
+      "! Using prepped_base file at knowledge_prepped_base.bin\n",
+      "Starting index build: R=32 L=64 Query RAM budget: 4.02653e+09 Indexing ram budget: 8 T: 8\n",
+      "getting bin metadata\n",
+      "Time for getting bin metadata: 0.000019 seconds\n",
+      "Compressing 769-dimensional data into 512 bytes per vector.\n",
+      "Opened: knowledge_prepped_base.bin, size: 18464, cache_size: 18464\n",
+      "Training data with 6 samples loaded.\n",
+      "Reading bin file knowledge_pq_pivots.bin ...\n",
+      "Opening bin file knowledge_pq_pivots.bin... \n",
+      "Metadata: #pts = 256, #dims = 769...\n",
+      "done.\n",
+      "PQ pivot file exists. Not generating again\n",
+      "Opened: knowledge_prepped_base.bin, size: 18464, cache_size: 18464\n",
+      "Reading bin file knowledge_pq_pivots.bin ...\n",
+      "Opening bin file knowledge_pq_pivots.bin... \n",
+      "Metadata: #pts = 4, #dims = 1...\n",
+      "done.\n",
+      "Reading bin file knowledge_pq_pivots.bin ...\n",
+      "Opening bin file knowledge_pq_pivots.bin... \n",
+      "Metadata: #pts = 256, #dims = 769...\n",
+      "done.\n",
+      "Reading bin file knowledge_pq_pivots.bin ...\n",
+      "Opening bin file knowledge_pq_pivots.bin... \n",
+      "Metadata: #pts = 769, #dims = 1...\n",
+      "done.\n",
+      "Reading bin file knowledge_pq_pivots.bin ...\n",
+      "Opening bin file knowledge_pq_pivots.bin... \n",
+      "Metadata: #pts = 513, #dims = 1...\n",
+      "done.\n",
+      "Loaded PQ pivot information\n",
+      "Processing points  [0, 6)...done.\n",
+      "Time for generating quantized data: 0.055587 seconds\n",
+      "Full index fits in RAM budget, should consume at most 2.03973e-05GiBs, so building in one shot\n",
+      "L2: Using AVX2 distance computation DistanceL2Float\n",
+      "Passed, empty search_params while creating index config\n",
+      "Using only first 6 from file.. \n",
+      "Starting index build with 6 points... \n",
+      "0% of index build completed.Starting final cleanup..done. Link time: 0.00011s\n",
+      "Index built with degree: max:5  avg:5  min:5  count(deg<2):0\n",
+      "Not saving tags as they are not enabled.\n",
+      "Time taken for save: 0.000148s.\n",
+      "Time for building merged vamana index: 0.000836 seconds\n",
+      "Opened: knowledge_prepped_base.bin, size: 18464, cache_size: 18464\n",
+      "Vamana index file size=168\n",
+      "Opened: knowledge_disk.index, cache_size: 67108864\n",
+      "medoid: 0B\n",
+      "max_node_len: 3100B\n",
+      "nnodes_per_sector: 1B\n",
+      "# sectors: 6\n",
+      "Sector #0written\n",
+      "Finished writing 28672B\n",
+      "Writing bin: knowledge_disk.index\n",
+      "bin: #pts = 9, #dims = 1, size = 80B\n",
+      "Finished writing bin.\n",
+      "Output disk index file written to knowledge_disk.index\n",
+      "Finished writing 28672B\n",
+      "Time for generating disk layout: 0.040268 seconds\n",
+      "Opened: knowledge_prepped_base.bin, size: 18464, cache_size: 18464\n",
+      "Loading base knowledge_prepped_base.bin. #points: 6. #dim: 769.\n",
+      "Wrote 1 points to sample file: knowledge_sample_data.bin\n",
+      "Indexing time: 0.0970594\n",
+      "INFO: Leann metadata saved to knowledge.leann.meta.json\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Opened file : knowledge_disk.index\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "✅ DiskANN index loaded successfully.\n",
+      "INFO: LeannSearcher initialized with 'diskann' backend using index 'knowledge.leann'.\n",
+      "Since data is floating point, we assume that it has been appropriately pre-processed (normalization for cosine, and convert-to-l2 by adding extra dimension for MIPS). So we shall invoke an l2 distance function.\n",
+      "L2: Using AVX2 distance computation DistanceL2Float\n",
+      "L2: Using AVX2 distance computation DistanceL2Float\n",
+      "Before index load\n",
+      "Reading bin file knowledge_pq_compressed.bin ...\n",
+      "Opening bin file knowledge_pq_compressed.bin... \n",
+      "Metadata: #pts = 6, #dims = 512...\n",
+      "done.\n",
+      "Reading bin file knowledge_pq_pivots.bin ...\n",
+      "Opening bin file knowledge_pq_pivots.bin... \n",
+      "Metadata: #pts = 4, #dims = 1...\n",
+      "done.\n",
+      "Offsets: 4096 791560 794644 796704\n",
+      "Reading bin file knowledge_pq_pivots.bin ...\n",
+      "Opening bin file knowledge_pq_pivots.bin... \n",
+      "Metadata: #pts = 256, #dims = 769...\n",
+      "done.\n",
+      "Reading bin file knowledge_pq_pivots.bin ...\n",
+      "Opening bin file knowledge_pq_pivots.bin... \n",
+      "Metadata: #pts = 769, #dims = 1...\n",
+      "done.\n",
+      "Reading bin file knowledge_pq_pivots.bin ...\n",
+      "Opening bin file knowledge_pq_pivots.bin... \n",
+      "Metadata: #pts = 513, #dims = 1...\n",
+      "done.\n",
+      "Loaded PQ Pivots: #ctrs: 256, #dims: 769, #chunks: 512\n",
+      "Loaded PQ centroids and in-memory compressed vectors. #points: 6 #dim: 769 #aligned_dim: 776 #chunks: 512\n",
+      "Loading index metadata from knowledge_disk.index\n",
+      "Disk-Index File Meta-data: # nodes per sector: 1, max node len (bytes): 3100, max node degree: 5\n",
+      "Disk-Index Meta: nodes per sector: 1, max node len: 3100, max node degree: 5\n",
+      "Setting up thread-specific contexts for nthreads: 8\n",
+      "allocating ctx: 0x7a33f7204000 to thread-id:134367072315200\n",
+      "allocating ctx: 0x7a33f6805000 to thread-id:134355206802368\n",
+      "allocating ctx: 0x7a33f5e72000 to thread-id:134355217288000\n",
+      "allocating ctx: 0x7a33f5e61000 to thread-id:134355227773632\n",
+      "allocating ctx: 0x7a33f5e50000 to thread-id:134355196316736\n",
+      "allocating ctx: 0x7a33f5e3f000 to thread-id:134355164859840\n",
+      "allocating ctx: 0x7a33f5e2e000 to thread-id:134355175345472\n",
+      "allocating ctx: 0x7a33f5e1d000 to thread-id:134355185831104\n",
+      "Loading centroid data from medoids vector data of 1 medoid(s)\n",
+      "Reading bin file knowledge_disk.index_max_base_norm.bin ...\n",
+      "Opening bin file knowledge_disk.index_max_base_norm.bin... \n",
+      "Metadata: #pts = 1, #dims = 1...\n",
+      "done.\n",
+      "Setting re-scaling factor of base vectors to 1\n",
+      "load_from_separate_paths done.\n",
+      "Reading (with alignment) bin file knowledge_sample_data.bin ...Metadata: #pts = 1, #dims = 769, aligned_dim = 776... allocating aligned memory of 3104 bytes... done. Copying data to mem_aligned buffer... done.\n",
+      "reserve ratio: 1\n",
+      "Graph traversal completed, hops: 3\n",
+      "Loading the cache list into memory....done.\n",
+      "After index load\n",
+      "INFO: Computing embeddings for 1 chunks using 'sentence-transformers/all-mpnet-base-v2'...\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Batches: 100%|██████████| 1/1 [00:00<00:00, 60.54it/s]"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "INFO: DiskANN ZMQ mode enabled - ensuring embedding server is running\n",
+      "INFO: Starting session-level embedding server as a background process...\n",
+      "INFO: Running command from project root: /home/ubuntu/LEANN_clean/leann\n",
+      "INFO: Server process started with PID: 424761\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "✅ Embedding server is up and ready for this session.\n",
+      "[EmbeddingServer LOG]: Initializing leann-backend-diskann...\n",
+      "[EmbeddingServer LOG]: WARNING: Could not import DiskANN backend: cannot import name '_diskannpy' from partially initialized module 'packages.leann-backend-diskann.leann_backend_diskann' (most likely due to a circular import) (/home/ubuntu/LEANN_clean/leann/packages/leann-backend-diskann/leann_backend_diskann/__init__.py)\n",
+      "[EmbeddingServer LOG]: INFO: Initializing embedding server thread on port 5555\n",
+      "[EmbeddingServer LOG]: INFO: Using CUDA device\n",
+      "[EmbeddingServer LOG]: INFO: Loading model sentence-transformers/all-mpnet-base-v2\n",
+      "[EmbeddingServer LOG]: INFO: Using FP16 precision with model: sentence-transformers/all-mpnet-base-v2\n",
+      "[EmbeddingServer LOG]: INFO: Loaded 6 demo documents\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ ROUTER server listening on port 5555\n",
+      "[EmbeddingServer LOG]: INFO: Embedding server ready to serve requests\n",
+      "[EmbeddingServer LOG]: INFO: Received ZMQ request from client 006b8b45, size 3 bytes\n",
+      "[EmbeddingServer LOG]: INFO: Request for 1 node embeddings: [0]\n",
+      "[EmbeddingServer LOG]: DEBUG: Node ID range: 0 to 0\n",
+      "[EmbeddingServer LOG]: Time taken for text lookup: 0.000028 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Total batch size: 1, max_batch_size: 128\n",
+      "[EmbeddingServer LOG]: INFO: Processing batch of size 1\n",
+      "[EmbeddingServer LOG]: Time taken for tokenization (batch): 0.019294 seconds\n",
+      "[EmbeddingServer LOG]: Batch size: 1, Sequence length: 256\n",
+      "[EmbeddingServer LOG]: Time taken for transfer to device (batch): 0.000210 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for embedding (batch): 3.065444 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for mean pooling (batch): 0.041810 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Serialize time: 0.000194 seconds\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ E2E time: 3.128073 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Received ZMQ request from client 006b8b45, size 7 bytes\n",
+      "[EmbeddingServer LOG]: INFO: Request for 5 node embeddings: [1, 2, 3, 4, 5]\n",
+      "[EmbeddingServer LOG]: DEBUG: Node ID range: 1 to 5\n",
+      "[EmbeddingServer LOG]: Time taken for text lookup: 0.000042 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Total batch size: 5, max_batch_size: 128\n",
+      "[EmbeddingServer LOG]: INFO: Processing batch of size 5\n",
+      "[EmbeddingServer LOG]: Time taken for tokenization (batch): 0.001791 seconds\n",
+      "[EmbeddingServer LOG]: Batch size: 5, Sequence length: 256\n",
+      "[EmbeddingServer LOG]: Time taken for transfer to device (batch): 0.000112 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for embedding (batch): 3.674183 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for mean pooling (batch): 0.000372 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Serialize time: 0.000177 seconds\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ E2E time: 3.677425 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Received ZMQ request from client 006b8b45, size 7 bytes\n",
+      "[EmbeddingServer LOG]: INFO: Request for 5 node embeddings: [3, 4, 2, 1, 0]\n",
+      "[EmbeddingServer LOG]: DEBUG: Node ID range: 0 to 4\n",
+      "[EmbeddingServer LOG]: Time taken for text lookup: 0.000030 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Total batch size: 5, max_batch_size: 128\n",
+      "[EmbeddingServer LOG]: INFO: Processing batch of size 5\n",
+      "[EmbeddingServer LOG]: Time taken for tokenization (batch): 0.001550 seconds\n",
+      "[EmbeddingServer LOG]: Batch size: 5, Sequence length: 256\n",
+      "[EmbeddingServer LOG]: Time taken for transfer to device (batch): 0.000097 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for embedding (batch): 0.009335 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for mean pooling (batch): 0.000154 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Serialize time: 0.000073 seconds\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ E2E time: 0.011773 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Received ZMQ request from client 006b8b45, size 7 bytes\n",
+      "[EmbeddingServer LOG]: INFO: Request for 5 node embeddings: [0, 1, 2, 4, 5]\n",
+      "[EmbeddingServer LOG]: DEBUG: Node ID range: 0 to 5\n",
+      "[EmbeddingServer LOG]: Time taken for text lookup: 0.000020 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Total batch size: 5, max_batch_size: 128\n",
+      "[EmbeddingServer LOG]: INFO: Processing batch of size 5\n",
+      "[EmbeddingServer LOG]: Time taken for tokenization (batch): 0.001041 seconds\n",
+      "[EmbeddingServer LOG]: Batch size: 5, Sequence length: 256\n",
+      "[EmbeddingServer LOG]: Time taken for transfer to device (batch): 0.000125 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for embedding (batch): 0.008972 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for mean pooling (batch): 0.000151 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Serialize time: 0.000048 seconds\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ E2E time: 0.010853 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Received ZMQ request from client 006b8b45, size 7 bytes\n",
+      "[EmbeddingServer LOG]: INFO: Request for 5 node embeddings: [3, 1, 0, 2, 5]\n",
+      "[EmbeddingServer LOG]: DEBUG: Node ID range: 0 to 5\n",
+      "[EmbeddingServer LOG]: Time taken for text lookup: 0.000020 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Total batch size: 5, max_batch_size: 128\n",
+      "[EmbeddingServer LOG]: INFO: Processing batch of size 5\n",
+      "[EmbeddingServer LOG]: Time taken for tokenization (batch): 0.001350 seconds\n",
+      "[EmbeddingServer LOG]: Batch size: 5, Sequence length: 256\n",
+      "[EmbeddingServer LOG]: Time taken for transfer to device (batch): 0.000088 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for embedding (batch): 0.008869 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for mean pooling (batch): 0.000146 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Serialize time: 0.000063 seconds\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ E2E time: 0.011054 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Received ZMQ request from client 006b8b45, size 7 bytes\n",
+      "[EmbeddingServer LOG]: INFO: Request for 5 node embeddings: [0, 2, 3, 4, 5]\n",
+      "[EmbeddingServer LOG]: DEBUG: Node ID range: 0 to 5\n",
+      "[EmbeddingServer LOG]: Time taken for text lookup: 0.000022 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Total batch size: 5, max_batch_size: 128\n",
+      "[EmbeddingServer LOG]: INFO: Processing batch of size 5\n",
+      "[EmbeddingServer LOG]: Time taken for tokenization (batch): 0.001195 seconds\n",
+      "[EmbeddingServer LOG]: Batch size: 5, Sequence length: 256\n",
+      "[EmbeddingServer LOG]: Time taken for transfer to device (batch): 0.000087 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for embedding (batch): 0.008903 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for mean pooling (batch): 0.000145 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Serialize time: 0.000060 seconds\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ E2E time: 0.010921 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Received ZMQ request from client 006b8b45, size 7 bytes\n",
+      "[EmbeddingServer LOG]: INFO: Request for 5 node embeddings: [1, 0, 3, 4, 5]\n",
+      "[EmbeddingServer LOG]: DEBUG: Node ID range: 0 to 5\n",
+      "[EmbeddingServer LOG]: Time taken for text lookup: 0.000020 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Total batch size: 5, max_batch_size: 128\n",
+      "[EmbeddingServer LOG]: INFO: Processing batch of size 5\n",
+      "[EmbeddingServer LOG]: Time taken for tokenization (batch): 0.001188 seconds\n",
+      "[EmbeddingServer LOG]: Batch size: 5, Sequence length: 256\n",
+      "[EmbeddingServer LOG]: Time taken for transfer to device (batch): 0.000087 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for embedding (batch): 0.008858 seconds\n",
+      "[EmbeddingServer LOG]: Time taken for mean pooling (batch): 0.000153 seconds\n",
+      "[EmbeddingServer LOG]: INFO: Serialize time: 0.000052 seconds\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ E2E time: 0.010886 seconds\n",
+      "reserve ratio: Score: -0.481 - C++ is a powerful programming language1\n",
+      "Graph traversal completed, hops: 3\n",
+      "\n",
+      "Score: -1.049 - Java is a powerful programming language\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "[EmbeddingServer LOG]: INFO: ZMQ socket timeout, continuing to listen\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ socket timeout, continuing to listen\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ socket timeout, continuing to listen\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ socket timeout, continuing to listen\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ socket timeout, continuing to listen\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ socket timeout, continuing to listen\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ socket timeout, continuing to listen\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ socket timeout, continuing to listen\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ socket timeout, continuing to listen\n",
+      "[EmbeddingServer LOG]: INFO: ZMQ socket timeout, continuing to listen\n"
+     ]
+    }
+   ],
    "source": [
-    "# install this if you are using colab\n",
+    "from leann.api import LeannBuilder, LeannSearcher\n",
-    "! uv pip install leann-core leann-backend-hnsw --no-deps\n",
+    "import leann_backend_diskann\n",
-    "! uv pip install leann --no-deps\n",
+    "# 1. Build index (no embeddings stored!)\n",
-    "# For Colab environment, we need to set some environment variables\n",
+    "builder = LeannBuilder(backend_name=\"diskann\")\n",
-    "import os\n",
+    "builder.add_text(\"Python is a powerful programming language\")\n",
-    "\n",
+    "builder.add_text(\"Machine learning transforms industries\")  \n",
-    "os.environ[\"LEANN_LOG_LEVEL\"] = \"INFO\"  # Enable more detailed logging"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from pathlib import Path\n",
-    "\n",
-    "INDEX_DIR = Path(\"./\").resolve()\n",
-    "INDEX_PATH = str(INDEX_DIR / \"demo.leann\")"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Build the index"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from leann.api import LeannBuilder\n",
-    "\n",
-    "builder = LeannBuilder(backend_name=\"hnsw\")\n",
-    "builder.add_text(\"C# is a powerful programming language and it is good at game development\")\n",
-    "builder.add_text(\n",
-    "    \"Python is a powerful programming language and it is good at machine learning tasks\"\n",
-    ")\n",
-    "builder.add_text(\"Machine learning transforms industries\")\n",
     "builder.add_text(\"Neural networks process complex data\")\n",
-    "builder.add_text(\"Leann is a great storage saving engine for RAG on your MacBook\")\n",
+    "builder.add_text(\"Java is a powerful programming language\")\n",
-    "builder.build_index(INDEX_PATH)"
+    "builder.add_text(\"C++ is a powerful programming language\")\n",
-   ]
+    "builder.add_text(\"C# is a powerful programming language\")\n",
-  },
+    "builder.build_index(\"knowledge.leann\")\n",
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Search with real-time embeddings"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from leann.api import LeannSearcher\n",
     "\n",
-    "searcher = LeannSearcher(INDEX_PATH)\n",
+    "# 2. Search with real-time embeddings\n",
-    "results = searcher.search(\"programming languages\", top_k=2)\n",
+    "searcher = LeannSearcher(\"knowledge.leann\")\n",
-    "results"
+    "results = searcher.search(\"C++ programming languages\", top_k=2,recompute_beighbor_embeddings=True)\n",
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Chat with LEANN using retrieved results"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from leann.api import LeannChat\n",
     "\n",
-    "llm_config = {\n",
+    "for result in results:\n",
-    "    \"type\": \"hf\",\n",
+    "    print(f\"Score: {result['score']:.3f} - {result['text']}\")"
-    "    \"model\": \"Qwen/Qwen3-0.6B\",\n",
-    "}\n",
-    "\n",
-    "chat = LeannChat(index_path=INDEX_PATH, llm_config=llm_config)\n",
-    "response = chat.ask(\n",
-    "    \"Compare the two retrieved programming languages and tell me their advantages.\",\n",
-    "    top_k=2,\n",
-    "    llm_kwargs={\"max_tokens\": 128},\n",
-    ")\n",
-    "response"
    ]
   }
  ],
@@ -130,7 +376,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.11.12"
+   "version": "3.11.11"
   }
  },
  "nbformat": 4,

docs/CONTRIBUTING.md

@@ -1,220 +0,0 @@
-# 🤝 Contributing
-
-We welcome contributions! Leann is built by the community, for the community.
-
-## Ways to Contribute
-
-- 🐛 **Bug Reports**: Found an issue? Let us know!
-- 💡 **Feature Requests**: Have an idea? We'd love to hear it!
-- 🔧 **Code Contributions**: PRs welcome for all skill levels
-- 📖 **Documentation**: Help make Leann more accessible
-- 🧪 **Benchmarks**: Share your performance results
-
-## 🚀 Development Setup
-
-### Prerequisites
-
-1. **Install uv** (fast Python package installer):
-   ```bash
-   curl -LsSf https://astral.sh/uv/install.sh | sh
-   ```
-
-2. **Clone the repository**:
-   ```bash
-   git clone https://github.com/LEANN-RAG/LEANN-RAG.git
-   cd LEANN-RAG
-   ```
-
-3. **Install system dependencies**:
-
-   **macOS:**
-   ```bash
-   brew install llvm libomp boost protobuf zeromq pkgconf
-   ```
-
-   **Ubuntu/Debian:**
-   ```bash
-   sudo apt-get install libomp-dev libboost-all-dev protobuf-compiler \
-                        libabsl-dev libmkl-full-dev libaio-dev libzmq3-dev
-   ```
-
-4. **Build from source**:
-   ```bash
-   # macOS
-   CC=$(brew --prefix llvm)/bin/clang CXX=$(brew --prefix llvm)/bin/clang++ uv sync
-
-   # Ubuntu/Debian
-   uv sync
-   ```
-
-## 🔨 Pre-commit Hooks
-
-We use pre-commit hooks to ensure code quality and consistency. This runs automatically before each commit.
-
-### Setup Pre-commit
-
-1. **Install pre-commit** (already included when you run `uv sync`):
-   ```bash
-   uv pip install pre-commit
-   ```
-
-2. **Install the git hooks**:
-   ```bash
-   pre-commit install
-   ```
-
-3. **Run pre-commit manually** (optional):
-   ```bash
-   pre-commit run --all-files
-   ```
-
-### Pre-commit Checks
-
-Our pre-commit configuration includes:
-- **Trailing whitespace removal**
-- **End-of-file fixing**
-- **YAML validation**
-- **Large file prevention**
-- **Merge conflict detection**
-- **Debug statement detection**
-- **Code formatting with ruff**
-- **Code linting with ruff**
-
-## 🧪 Testing
-
-### Running Tests
-
-```bash
-# Run all tests
-uv run pytest
-
-# Run specific test file
-uv run pytest test/test_filename.py
-
-# Run with coverage
-uv run pytest --cov=leann
-```
-
-### Writing Tests
-
-- Place tests in the `test/` directory
-- Follow the naming convention `test_*.py`
-- Use descriptive test names that explain what's being tested
-- Include both positive and negative test cases
-
-## 📝 Code Style
-
-We use `ruff` for both linting and formatting to ensure consistent code style.
-
-### Format Your Code
-
-```bash
-# Format all files
-ruff format
-
-# Check formatting without changing files
-ruff format --check
-```
-
-### Lint Your Code
-
-```bash
-# Run linter with auto-fix
-ruff check --fix
-
-# Just check without fixing
-ruff check
-```
-
-### Style Guidelines
-
-- Follow PEP 8 conventions
-- Use descriptive variable names
-- Add type hints where appropriate
-- Write docstrings for all public functions and classes
-- Keep functions focused and single-purpose
-
-## 🚦 CI/CD
-
-Our CI pipeline runs automatically on all pull requests. It includes:
-
-1. **Linting and Formatting**: Ensures code follows our style guidelines
-2. **Multi-platform builds**: Tests on Ubuntu and macOS
-3. **Python version matrix**: Tests on Python 3.9-3.13
-4. **Wheel building**: Ensures packages can be built and distributed
-
-### CI Commands
-
-The CI uses the same commands as pre-commit to ensure consistency:
-```bash
-# Linting
-ruff check .
-
-# Format checking
-ruff format --check .
-```
-
-Make sure your code passes these checks locally before pushing!
-
-## 🔄 Pull Request Process
-
-1. **Fork the repository** and create your branch from `main`:
-   ```bash
-   git checkout -b feature/your-feature-name
-   ```
-
-2. **Make your changes**:
-   - Write clean, documented code
-   - Add tests for new functionality
-   - Update documentation as needed
-
-3. **Run pre-commit checks**:
-   ```bash
-   pre-commit run --all-files
-   ```
-
-4. **Test your changes**:
-   ```bash
-   uv run pytest
-   ```
-
-5. **Commit with descriptive messages**:
-   ```bash
-   git commit -m "feat: add new search algorithm"
-   ```
-
-   Follow [Conventional Commits](https://www.conventionalcommits.org/):
-   - `feat:` for new features
-   - `fix:` for bug fixes
-   - `docs:` for documentation changes
-   - `test:` for test additions/changes
-   - `refactor:` for code refactoring
-   - `perf:` for performance improvements
-
-6. **Push and create a pull request**:
-   - Provide a clear description of your changes
-   - Reference any related issues
-   - Include examples or screenshots if applicable
-
-## 📚 Documentation
-
-When adding new features or making significant changes:
-
-1. Update relevant documentation in `/docs`
-2. Add docstrings to new functions/classes
-3. Update README.md if needed
-4. Include usage examples
-
-## 🤔 Getting Help
-
-- **Discord**: Join our community for discussions
-- **Issues**: Check existing issues or create a new one
-- **Discussions**: For general questions and ideas
-
-## 📄 License
-
-By contributing, you agree that your contributions will be licensed under the same license as the project (MIT).
-
----
-
-Thank you for contributing to LEANN! Every contribution, no matter how small, helps make the project better for everyone. 🌟

docs/RELEASE.md

@@ -1,22 +0,0 @@
-# Release Guide
-
-## Setup (One-time)
-
-Add `PYPI_API_TOKEN` to GitHub Secrets:
-1. Get token: https://pypi.org/manage/account/token/
-2. Add to secrets: Settings → Secrets → Actions → `PYPI_API_TOKEN`
-
-## Release (One-click)
-
-1. Go to: https://github.com/yichuan-w/LEANN/actions/workflows/release-manual.yml
-2. Click "Run workflow"
-3. Enter version: `0.1.2`
-4. Click green "Run workflow" button
-
-That's it! The workflow will automatically:
-- ✅ Update version in all packages
-- ✅ Build all packages
-- ✅ Publish to PyPI
-- ✅ Create GitHub tag and release
-
-Check progress: https://github.com/yichuan-w/LEANN/actions

docs/THINKING_BUDGET_FEATURE.md

@@ -1,123 +0,0 @@
-# Thinking Budget Feature Implementation
-
-## Overview
-
-This document describes the implementation of the **thinking budget** feature for LEANN, which allows users to control the computational effort for reasoning models like GPT-Oss:20b.
-
-## Feature Description
-
-The thinking budget feature provides three levels of computational effort for reasoning models:
-- **`low`**: Fast responses, basic reasoning (default for simple queries)
-- **`medium`**: Balanced speed and reasoning depth
-- **`high`**: Maximum reasoning effort, best for complex analytical questions
-
-## Implementation Details
-
-### 1. Command Line Interface
-
-Added `--thinking-budget` parameter to both CLI and RAG examples:
-
-```bash
-# LEANN CLI
-leann ask my-index --llm ollama --model gpt-oss:20b --thinking-budget high
-
-# RAG Examples
-python apps/email_rag.py --llm ollama --llm-model gpt-oss:20b --thinking-budget high
-python apps/document_rag.py --llm openai --llm-model o3 --thinking-budget medium
-```
-
-### 2. LLM Backend Support
-
-#### Ollama Backend (`packages/leann-core/src/leann/chat.py`)
-
-```python
-def ask(self, prompt: str, **kwargs) -> str:
-    # Handle thinking budget for reasoning models
-    options = kwargs.copy()
-    thinking_budget = kwargs.get("thinking_budget")
-    if thinking_budget:
-        options.pop("thinking_budget", None)
-        if thinking_budget in ["low", "medium", "high"]:
-            options["reasoning"] = {"effort": thinking_budget, "exclude": False}
-```
-
-**API Format**: Uses Ollama's `reasoning` parameter with `effort` and `exclude` fields.
-
-#### OpenAI Backend (`packages/leann-core/src/leann/chat.py`)
-
-```python
-def ask(self, prompt: str, **kwargs) -> str:
-    # Handle thinking budget for reasoning models
-    thinking_budget = kwargs.get("thinking_budget")
-    if thinking_budget and thinking_budget in ["low", "medium", "high"]:
-        # Check if this is an o-series model
-        o_series_models = ["o3", "o3-mini", "o4-mini", "o1", "o3-pro", "o3-deep-research"]
-        if any(model in self.model for model in o_series_models):
-            params["reasoning_effort"] = thinking_budget
-```
-
-**API Format**: Uses OpenAI's `reasoning_effort` parameter for o-series models.
-
-### 3. Parameter Propagation
-
-The thinking budget parameter is properly propagated through the LEANN architecture:
-
-1. **CLI** (`packages/leann-core/src/leann/cli.py`): Captures `--thinking-budget` argument
-2. **Base RAG** (`apps/base_rag_example.py`): Adds parameter to argument parser
-3. **LeannChat** (`packages/leann-core/src/leann/api.py`): Passes `llm_kwargs` to LLM
-4. **LLM Interface**: Handles the parameter in backend-specific implementations
-
-## Files Modified
-
-### Core Implementation
-- `packages/leann-core/src/leann/chat.py`: Added thinking budget support to OllamaChat and OpenAIChat
-- `packages/leann-core/src/leann/cli.py`: Added `--thinking-budget` argument
-- `apps/base_rag_example.py`: Added thinking budget parameter to RAG examples
-
-### Documentation
-- `README.md`: Added thinking budget parameter to usage examples
-- `docs/configuration-guide.md`: Added detailed documentation and usage guidelines
-
-### Examples
-- `examples/thinking_budget_demo.py`: Comprehensive demo script with usage examples
-
-## Usage Examples
-
-### Basic Usage
-```bash
-# High reasoning effort for complex questions
-leann ask my-index --llm ollama --model gpt-oss:20b --thinking-budget high
-
-# Medium reasoning for balanced performance
-leann ask my-index --llm openai --model gpt-4o --thinking-budget medium
-
-# Low reasoning for fast responses
-leann ask my-index --llm ollama --model gpt-oss:20b --thinking-budget low
-```
-
-### RAG Examples
-```bash
-# Email RAG with high reasoning
-python apps/email_rag.py --llm ollama --llm-model gpt-oss:20b --thinking-budget high
-
-# Document RAG with medium reasoning
-python apps/document_rag.py --llm openai --llm-model gpt-4o --thinking-budget medium
-```
-
-## Supported Models
-
-### Ollama Models
-- **GPT-Oss:20b**: Primary target model with reasoning capabilities
-- **Other reasoning models**: Any Ollama model that supports the `reasoning` parameter
-
-### OpenAI Models
-- **o3, o3-mini, o4-mini, o1**: o-series reasoning models with `reasoning_effort` parameter
-- **GPT-OSS models**: Models that support reasoning capabilities
-
-## Testing
-
-The implementation includes comprehensive testing:
-- Parameter handling verification
-- Backend-specific API format validation
-- CLI argument parsing tests
-- Integration with existing LEANN architecture

docs/ast_chunking_guide.md

@@ -1,143 +0,0 @@
-# AST-Aware Code chunking guide
-
-## Overview
-
-This guide covers best practices for using AST-aware code chunking in LEANN. AST chunking provides better semantic understanding of code structure compared to traditional text-based chunking.
-
-## Quick Start
-
-### Basic Usage
-
-```bash
-# Enable AST chunking for mixed content (code + docs)
-python -m apps.document_rag --enable-code-chunking --data-dir ./my_project
-
-# Specialized code repository indexing
-python -m apps.code_rag --repo-dir ./my_codebase
-
-# Global CLI with AST support
-leann build my-code-index --docs ./src --use-ast-chunking
-```
-
-### Installation
-
-```bash
-# Install LEANN with AST chunking support
-uv pip install -e "."
-```
-
-#### For normal users (PyPI install)
-- Use `pip install leann` or `uv pip install leann`.
-- `astchunk` is pulled automatically from PyPI as a dependency; no extra steps.
-
-#### For developers (from source, editable)
-```bash
-git clone https://github.com/yichuan-w/LEANN.git leann
-cd leann
-git submodule update --init --recursive
-uv sync
-```
-- This repo vendors `astchunk` as a git submodule at `packages/astchunk-leann` (our fork).
-- `[tool.uv.sources]` maps the `astchunk` package to that path in editable mode.
-- You can edit code under `packages/astchunk-leann` and Python will use your changes immediately (no separate `pip install astchunk` needed).
-
-## Best Practices
-
-### When to Use AST Chunking
-
-✅ **Recommended for:**
-- Code repositories with multiple languages
-- Mixed documentation and code content
-- Complex codebases with deep function/class hierarchies
-- When working with Claude Code for code assistance
-
-❌ **Not recommended for:**
-- Pure text documents
-- Very large files (>1MB)
-- Languages not supported by tree-sitter
-
-### Optimal Configuration
-
-```bash
-# Recommended settings for most codebases
-python -m apps.code_rag \
-    --repo-dir ./src \
-    --ast-chunk-size 768 \
-    --ast-chunk-overlap 96 \
-    --exclude-dirs .git __pycache__ node_modules build dist
-```
-
-### Supported Languages
-
-| Extension | Language | Status |
-|-----------|----------|--------|
-| `.py` | Python | ✅ Full support |
-| `.java` | Java | ✅ Full support |
-| `.cs` | C# | ✅ Full support |
-| `.ts`, `.tsx` | TypeScript | ✅ Full support |
-| `.js`, `.jsx` | JavaScript | ✅ Via TypeScript parser |
-
-## Integration Examples
-
-### Document RAG with Code Support
-
-```python
-# Enable code chunking in document RAG
-python -m apps.document_rag \
-    --enable-code-chunking \
-    --data-dir ./project \
-    --query "How does authentication work in the codebase?"
-```
-
-### Claude Code Integration
-
-When using with Claude Code MCP server, AST chunking provides better context for:
-- Code completion and suggestions
-- Bug analysis and debugging
-- Architecture understanding
-- Refactoring assistance
-
-## Troubleshooting
-
-### Common Issues
-
-1. **Fallback to Traditional Chunking**
-   - Normal behavior for unsupported languages
-   - Check logs for specific language support
-
-2. **Performance with Large Files**
-   - Adjust `--max-file-size` parameter
-   - Use `--exclude-dirs` to skip unnecessary directories
-
-3. **Quality Issues**
-   - Try different `--ast-chunk-size` values (512, 768, 1024)
-   - Adjust overlap for better context preservation
-
-### Debug Mode
-
-```bash
-export LEANN_LOG_LEVEL=DEBUG
-python -m apps.code_rag --repo-dir ./my_code
-```
-
-## Migration from Traditional Chunking
-
-Existing workflows continue to work without changes. To enable AST chunking:
-
-```bash
-# Before
-python -m apps.document_rag --chunk-size 256
-
-# After (maintains traditional chunking for non-code files)
-python -m apps.document_rag --enable-code-chunking --chunk-size 256 --ast-chunk-size 768
-```
-
-## References
-
-- [astchunk GitHub Repository](https://github.com/yilinjz/astchunk)
-- [LEANN MCP Integration](../packages/leann-mcp/README.md)
-- [Research Paper](https://arxiv.org/html/2506.15655v1)
-
----
-
-**Note**: AST chunking maintains full backward compatibility while enhancing code understanding capabilities.

docs/code/embedding_model_compare.py

@@ -1,98 +0,0 @@
-"""
-Comparison between Sentence Transformers and OpenAI embeddings
-
-This example shows how different embedding models handle complex queries
-and demonstrates the differences between local and API-based embeddings.
-"""
-
-import numpy as np
-from leann.embedding_compute import compute_embeddings
-
-# OpenAI API key should be set as environment variable
-# export OPENAI_API_KEY="your-api-key-here"
-
-# Test data
-conference_text = "[Title]: COLING 2025 Conference\n[URL]: https://coling2025.org/"
-browser_text = "[Title]: Browser Use Tool\n[URL]: https://github.com/browser-use"
-
-# Two queries with same intent but different wording
-query1 = "Tell me my browser history about some conference i often visit"
-query2 = "browser history about conference I often visit"
-
-texts = [query1, query2, conference_text, browser_text]
-
-
-def cosine_similarity(a, b):
-    return np.dot(a, b)  # Already normalized
-
-
-def analyze_embeddings(embeddings, model_name):
-    print(f"\n=== {model_name} Results ===")
-
-    # Results for Query 1
-    sim1_conf = cosine_similarity(embeddings[0], embeddings[2])
-    sim1_browser = cosine_similarity(embeddings[0], embeddings[3])
-
-    print(f"Query 1: '{query1}'")
-    print(f"  → Conference similarity: {sim1_conf:.4f} {'✓' if sim1_conf > sim1_browser else ''}")
-    print(
-        f"  → Browser similarity:    {sim1_browser:.4f} {'✓' if sim1_browser > sim1_conf else ''}"
-    )
-    print(f"  Winner: {'Conference' if sim1_conf > sim1_browser else 'Browser'}")
-
-    # Results for Query 2
-    sim2_conf = cosine_similarity(embeddings[1], embeddings[2])
-    sim2_browser = cosine_similarity(embeddings[1], embeddings[3])
-
-    print(f"\nQuery 2: '{query2}'")
-    print(f"  → Conference similarity: {sim2_conf:.4f} {'✓' if sim2_conf > sim2_browser else ''}")
-    print(
-        f"  → Browser similarity:    {sim2_browser:.4f} {'✓' if sim2_browser > sim2_conf else ''}"
-    )
-    print(f"  Winner: {'Conference' if sim2_conf > sim2_browser else 'Browser'}")
-
-    # Show the impact
-    print("\n=== Impact Analysis ===")
-    print(f"Conference similarity change: {sim2_conf - sim1_conf:+.4f}")
-    print(f"Browser similarity change:    {sim2_browser - sim1_browser:+.4f}")
-
-    if sim1_conf > sim1_browser and sim2_browser > sim2_conf:
-        print("❌ FLIP: Adding 'browser history' flips winner from Conference to Browser!")
-    elif sim1_conf > sim1_browser and sim2_conf > sim2_browser:
-        print("✅ STABLE: Conference remains winner in both queries")
-    elif sim1_browser > sim1_conf and sim2_browser > sim2_conf:
-        print("✅ STABLE: Browser remains winner in both queries")
-    else:
-        print("🔄 MIXED: Results vary between queries")
-
-    return {
-        "query1_conf": sim1_conf,
-        "query1_browser": sim1_browser,
-        "query2_conf": sim2_conf,
-        "query2_browser": sim2_browser,
-    }
-
-
-# Test Sentence Transformers
-print("Testing Sentence Transformers (facebook/contriever)...")
-try:
-    st_embeddings = compute_embeddings(texts, "facebook/contriever", mode="sentence-transformers")
-    st_results = analyze_embeddings(st_embeddings, "Sentence Transformers (facebook/contriever)")
-except Exception as e:
-    print(f"❌ Sentence Transformers failed: {e}")
-    st_results = None
-
-# Test OpenAI
-print("\n" + "=" * 60)
-print("Testing OpenAI (text-embedding-3-small)...")
-try:
-    openai_embeddings = compute_embeddings(texts, "text-embedding-3-small", mode="openai")
-    openai_results = analyze_embeddings(openai_embeddings, "OpenAI (text-embedding-3-small)")
-except Exception as e:
-    print(f"❌ OpenAI failed: {e}")
-    openai_results = None
-
-# Compare results
-if st_results and openai_results:
-    print("\n" + "=" * 60)
-    print("=== COMPARISON SUMMARY ===")

docs/configuration-guide.md

@@ -1,459 +0,0 @@
-# LEANN Configuration Guide
-
-This guide helps you optimize LEANN for different use cases and understand the trade-offs between various configuration options.
-
-## Getting Started: Simple is Better
-
-When first trying LEANN, start with a small dataset to quickly validate your approach:
-
-**For document RAG**: The default `data/` directory works perfectly - includes 2 AI research papers, Pride and Prejudice literature, and a technical report
-```bash
-python -m apps.document_rag --query "What techniques does LEANN use?"
-```
-
-**For other data sources**: Limit the dataset size for quick testing
-```bash
-# WeChat: Test with recent messages only
-python -m apps.wechat_rag --max-items 100 --query "What did we discuss about the project timeline?"
-
-# Browser history: Last few days
-python -m apps.browser_rag --max-items 500 --query "Find documentation about vector databases"
-
-# Email: Recent inbox
-python -m apps.email_rag --max-items 200 --query "Who sent updates about the deployment status?"
-```
-
-Once validated, scale up gradually:
-- 100 documents → 1,000 → 10,000 → full dataset (`--max-items -1`)
-- This helps identify issues early before committing to long processing times
-
-## Embedding Model Selection: Understanding the Trade-offs
-
-Based on our experience developing LEANN, embedding models fall into three categories:
-
-### Small Models (< 100M parameters)
-**Example**: `sentence-transformers/all-MiniLM-L6-v2` (22M params)
-- **Pros**: Lightweight, fast for both indexing and inference
-- **Cons**: Lower semantic understanding, may miss nuanced relationships
-- **Use when**: Speed is critical, handling simple queries, interactive mode, or just experimenting with LEANN. If time is not a constraint, consider using a larger/better embedding model
-
-### Medium Models (100M-500M parameters)
-**Example**: `facebook/contriever` (110M params), `BAAI/bge-base-en-v1.5` (110M params)
-- **Pros**: Balanced performance, good multilingual support, reasonable speed
-- **Cons**: Requires more compute than small models
-- **Use when**: Need quality results without extreme compute requirements, general-purpose RAG applications
-
-### Large Models (500M+ parameters)
-**Example**: `Qwen/Qwen3-Embedding-0.6B` (600M params), `intfloat/multilingual-e5-large` (560M params)
-- **Pros**: Best semantic understanding, captures complex relationships, excellent multilingual support. **Qwen3-Embedding-0.6B achieves nearly OpenAI API performance!**
-- **Cons**: Slower inference, longer index build times
-- **Use when**: Quality is paramount and you have sufficient compute resources. **Highly recommended** for production use
-
-### Quick Start: Cloud and Local Embedding Options
-
-**OpenAI Embeddings (Fastest Setup)**
-For immediate testing without local model downloads(also if you [do not have GPU](https://github.com/yichuan-w/LEANN/issues/43) and do not care that much about your document leak, you should use this, we compute the embedding and recompute using openai API):
-```bash
-# Set OpenAI embeddings (requires OPENAI_API_KEY)
---embedding-mode openai --embedding-model text-embedding-3-small
-```
-
-**Ollama Embeddings (Privacy-Focused)**
-For local embeddings with complete privacy:
-```bash
-# First, pull an embedding model
-ollama pull nomic-embed-text
-
-# Use Ollama embeddings
---embedding-mode ollama --embedding-model nomic-embed-text
-```
-
-<details>
-<summary><strong>Cloud vs Local Trade-offs</strong></summary>
-
-**OpenAI Embeddings** (`text-embedding-3-small/large`)
-- **Pros**: No local compute needed, consistently fast, high quality
-- **Cons**: Requires API key, costs money, data leaves your system, [known limitations with certain languages](https://yichuan-w.github.io/blog/lessons_learned_in_dev_leann/)
-- **When to use**: Prototyping, non-sensitive data, need immediate results
-
-**Local Embeddings**
-- **Pros**: Complete privacy, no ongoing costs, full control, can sometimes outperform OpenAI embeddings
-- **Cons**: Slower than cloud APIs, requires local compute resources
-- **When to use**: Production systems, sensitive data, cost-sensitive applications
-
-</details>
-
-## Local & Remote Inference Endpoints
-
-> Applies to both LLMs (`leann ask`) and embeddings (`leann build`).
-
-LEANN now treats Ollama, LM Studio, and other OpenAI-compatible runtimes as first-class providers. You can point LEANN at any compatible endpoint – either on the same machine or across the network – with a couple of flags or environment variables.
-
-### One-Time Environment Setup
-
-```bash
-# Works for OpenAI-compatible runtimes such as LM Studio, vLLM, SGLang, llamafile, etc.
-export OPENAI_API_KEY="your-key"            # or leave unset for local servers that do not check keys
-export OPENAI_BASE_URL="http://localhost:1234/v1"
-
-# Ollama-compatible runtimes (Ollama, Ollama on another host, llamacpp-server, etc.)
-export LEANN_OLLAMA_HOST="http://localhost:11434"   # falls back to OLLAMA_HOST or LOCAL_LLM_ENDPOINT
-```
-
-LEANN also recognises `LEANN_LOCAL_LLM_HOST` (highest priority), `LEANN_OPENAI_BASE_URL`, and `LOCAL_OPENAI_BASE_URL`, so existing scripts continue to work.
-
-### Passing Hosts Per Command
-
-```bash
-# Build an index with a remote embedding server
-leann build my-notes \
-  --docs ./notes \
-  --embedding-mode openai \
-  --embedding-model text-embedding-qwen3-embedding-0.6b \
-  --embedding-api-base http://192.168.1.50:1234/v1 \
-  --embedding-api-key local-dev-key
-
-# Query using a local LM Studio instance via OpenAI-compatible API
-leann ask my-notes \
-  --llm openai \
-  --llm-model qwen3-8b \
-  --api-base http://localhost:1234/v1 \
-  --api-key local-dev-key
-
-# Query an Ollama instance running on another box
-leann ask my-notes \
-  --llm ollama \
-  --llm-model qwen3:14b \
-  --host http://192.168.1.101:11434
-```
-
-⚠️ **Make sure the endpoint is reachable**: when your inference server runs on a home/workstation and the index/search job runs in the cloud, the server must be able to reach the host you configured. Typical options include:
-
-- Expose a public IP (and open the relevant port) on the machine that hosts LM Studio/Ollama.
-- Configure router or cloud provider port forwarding.
-- Tunnel traffic through tools like `tailscale`, `cloudflared`, or `ssh -R`.
-
-When you set these options while building an index, LEANN stores them in `meta.json`. Any subsequent `leann ask` or searcher process automatically reuses the same provider settings – even when we spawn background embedding servers. This makes the “server without GPU talking to my local workstation” workflow from [issue #80](https://github.com/yichuan-w/LEANN/issues/80#issuecomment-2287230548) work out-of-the-box.
-
-**Tip:** If your runtime does not require an API key (many local stacks don’t), leave `--api-key` unset. LEANN will skip injecting credentials.
-
-### Python API Usage
-
-You can pass the same configuration from Python:
-
-```python
-from leann.api import LeannBuilder
-
-builder = LeannBuilder(
-    backend_name="hnsw",
-    embedding_mode="openai",
-    embedding_model="text-embedding-qwen3-embedding-0.6b",
-    embedding_options={
-        "base_url": "http://192.168.1.50:1234/v1",
-        "api_key": "local-dev-key",
-    },
-)
-builder.build_index("./indexes/my-notes", chunks)
-```
-
-`embedding_options` is persisted to the index `meta.json`, so subsequent `LeannSearcher` or `LeannChat` sessions automatically reuse the same provider settings (the embedding server manager forwards them to the provider for you).
-
-## Index Selection: Matching Your Scale
-
-### HNSW (Hierarchical Navigable Small World)
-**Best for**: Small to medium datasets (< 10M vectors) - **Default and recommended for extreme low storage**
-- Full recomputation required
-- High memory usage during build phase
-- Excellent recall (95%+)
-
-```bash
-# Optimal for most use cases
---backend-name hnsw --graph-degree 32 --build-complexity 64
-```
-
-### DiskANN
-**Best for**: Large datasets, especially when you want `recompute=True`.
-
-**Key advantages:**
-- **Faster search** on large datasets (3x+ speedup vs HNSW in many cases)
-- **Smart storage**: `recompute=True` enables automatic graph partitioning for smaller indexes
-- **Better scaling**: Designed for 100k+ documents
-
-**Recompute behavior:**
-- `recompute=True` (recommended): Pure PQ traversal + final reranking - faster and enables partitioning
-- `recompute=False`: PQ + partial real distances during traversal - slower but higher accuracy
-
-```bash
-# Recommended for most use cases
---backend-name diskann --graph-degree 32 --build-complexity 64
-```
-
-**Performance Benchmark**: Run `uv run benchmarks/diskann_vs_hnsw_speed_comparison.py` to compare DiskANN and HNSW on your system.
-
-## LLM Selection: Engine and Model Comparison
-
-### LLM Engines
-
-**OpenAI** (`--llm openai`)
-- **Pros**: Best quality, consistent performance, no local resources needed
-- **Cons**: Costs money ($0.15-2.5 per million tokens), requires internet, data privacy concerns
-- **Models**: `gpt-4o-mini` (fast, cheap), `gpt-4o` (best quality), `o3` (reasoning), `o3-mini` (reasoning, cheaper)
-- **Thinking Budget**: Use `--thinking-budget low/medium/high` for o-series reasoning models (o3, o3-mini, o4-mini)
-- **Note**: Our current default, but we recommend switching to Ollama for most use cases
-
-**Ollama** (`--llm ollama`)
-- **Pros**: Fully local, free, privacy-preserving, good model variety
-- **Cons**: Requires local GPU/CPU resources, slower than cloud APIs, need to install extra [ollama app](https://github.com/ollama/ollama?tab=readme-ov-file#ollama) and pre-download models by `ollama pull`
-- **Models**: `qwen3:0.6b` (ultra-fast), `qwen3:1.7b` (balanced), `qwen3:4b` (good quality), `qwen3:7b` (high quality), `deepseek-r1:1.5b` (reasoning)
-- **Thinking Budget**: Use `--thinking-budget low/medium/high` for reasoning models like GPT-Oss:20b
-
-**HuggingFace** (`--llm hf`)
-- **Pros**: Free tier available, huge model selection, direct model loading (vs Ollama's server-based approach)
-- **Cons**: More complex initial setup
-- **Models**: `Qwen/Qwen3-1.7B-FP8`
-
-## Parameter Tuning Guide
-
-### Search Complexity Parameters
-
-**`--build-complexity`** (index building)
-- Controls thoroughness during index construction
-- Higher = better recall but slower build
-- Recommendations:
-  - 32: Quick prototyping
-  - 64: Balanced (default)
-  - 128: Production systems
-  - 256: Maximum quality
-
-**`--search-complexity`** (query time)
-- Controls search thoroughness
-- Higher = better results but slower
-- Recommendations:
-  - 16: Fast/Interactive search
-  - 32: High quality with diversity
-  - 64+: Maximum accuracy
-
-### Top-K Selection
-
-**`--top-k`** (number of retrieved chunks)
-- More chunks = better context but slower LLM processing
-- Should be always smaller than `--search-complexity`
-- Guidelines:
-  - 10-20: General questions (default: 20)
-  - 30+: Complex multi-hop reasoning requiring comprehensive context
-
-**Trade-off formula**:
-- Retrieval time ∝ log(n) × search_complexity
-- LLM processing time ∝ top_k × chunk_size
-- Total context = top_k × chunk_size tokens
-
-### Thinking Budget for Reasoning Models
-
-**`--thinking-budget`** (reasoning effort level)
-- Controls the computational effort for reasoning models
-- Options: `low`, `medium`, `high`
-- Guidelines:
-  - `low`: Fast responses, basic reasoning (default for simple queries)
-  - `medium`: Balanced speed and reasoning depth
-  - `high`: Maximum reasoning effort, best for complex analytical questions
-- **Supported Models**:
-  - **Ollama**: `gpt-oss:20b`, `gpt-oss:120b`
-  - **OpenAI**: `o3`, `o3-mini`, `o4-mini`, `o1` (o-series reasoning models)
-- **Note**: Models without reasoning support will show a warning and proceed without reasoning parameters
-- **Example**: `--thinking-budget high` for complex analytical questions
-
-**📖 For detailed usage examples and implementation details, check out [Thinking Budget Documentation](THINKING_BUDGET_FEATURE.md)**
-
-**💡 Quick Examples:**
-```bash
-# OpenAI o-series reasoning model
-python apps/document_rag.py --query "What are the main techniques LEANN explores?" \
-  --index-dir hnswbuild --backend hnsw \
-  --llm openai --llm-model o3 --thinking-budget medium
-
-# Ollama reasoning model
-python apps/document_rag.py --query "What are the main techniques LEANN explores?" \
-  --index-dir hnswbuild --backend hnsw \
-  --llm ollama --llm-model gpt-oss:20b --thinking-budget high
-```
-
-### Graph Degree (HNSW/DiskANN)
-
-**`--graph-degree`**
-- Number of connections per node in the graph
-- Higher = better recall but more memory
-- HNSW: 16-32 (default: 32)
-- DiskANN: 32-128 (default: 64)
-
-
-## Performance Optimization Checklist
-
-### If Embedding is Too Slow
-
-1. **Switch to smaller model**:
-   ```bash
-   # From large model
-   --embedding-model Qwen/Qwen3-Embedding-0.6B
-   # To small model
-   --embedding-model sentence-transformers/all-MiniLM-L6-v2
-   ```
-
-2. **Limit dataset size for testing**:
-   ```bash
-   --max-items 1000  # Process first 1k items only
-   ```
-
-3. **Use MLX on Apple Silicon** (optional optimization):
-   ```bash
-   --embedding-mode mlx --embedding-model mlx-community/Qwen3-Embedding-0.6B-8bit
-   ```
-    MLX might not be the best choice, as we tested and found that it only offers 1.3x acceleration compared to HF, so maybe using ollama is a better choice for embedding generation
-
-4. **Use Ollama**
-   ```bash
-   --embedding-mode ollama --embedding-model nomic-embed-text
-   ```
-   To discover additional embedding models in ollama, check out https://ollama.com/search?c=embedding or read more about embedding models at https://ollama.com/blog/embedding-models, please do check the model size that works best for you
-### If Search Quality is Poor
-
-1. **Increase retrieval count**:
-   ```bash
-   --top-k 30  # Retrieve more candidates
-   ```
-
-2. **Upgrade embedding model**:
-   ```bash
-   # For English
-   --embedding-model BAAI/bge-base-en-v1.5
-   # For multilingual
-   --embedding-model intfloat/multilingual-e5-large
-   ```
-
-## Understanding the Trade-offs
-
-Every configuration choice involves trade-offs:
-
-| Factor | Small/Fast | Large/Quality |
-|--------|------------|---------------|
-| Embedding Model | `all-MiniLM-L6-v2` | `Qwen/Qwen3-Embedding-0.6B` |
-| Chunk Size | 512 tokens | 128 tokens |
-| Index Type | HNSW | DiskANN |
-| LLM | `qwen3:1.7b` | `gpt-4o` |
-
-The key is finding the right balance for your specific use case. Start small and simple, measure performance, then scale up only where needed.
-
-## Low-resource setups
-
-If you don’t have a local GPU or builds/searches are too slow, use one or more of the options below.
-
-### 1) Use OpenAI embeddings (no local compute)
-
-Fastest path with zero local GPU requirements. Set your API key and use OpenAI embeddings during build and search:
-
-```bash
-export OPENAI_API_KEY=sk-...
-
-# Build with OpenAI embeddings
-leann build my-index \
-  --embedding-mode openai \
-  --embedding-model text-embedding-3-small
-
-# Search with OpenAI embeddings (recompute at query time)
-leann search my-index "your query" \
-  --recompute
-```
-
-### 2) Run remote builds with SkyPilot (cloud GPU)
-
-Offload embedding generation and index building to a GPU VM using [SkyPilot](https://skypilot.readthedocs.io/en/latest/). A template is provided at `sky/leann-build.yaml`.
-
-```bash
-# One-time: install and configure SkyPilot
-pip install skypilot
-
-# Launch with defaults (L4:1) and mount ./data to ~/leann-data; the build runs automatically
-sky launch -c leann-gpu sky/leann-build.yaml
-
-# Override parameters via -e key=value (optional)
-sky launch -c leann-gpu sky/leann-build.yaml \
-  -e index_name=my-index \
-  -e backend=hnsw \
-  -e embedding_mode=sentence-transformers \
-  -e embedding_model=Qwen/Qwen3-Embedding-0.6B
-
-# Copy the built index back to your local .leann (use rsync)
-rsync -Pavz leann-gpu:~/.leann/indexes/my-index ./.leann/indexes/
-```
-
-### 3) Disable recomputation to trade storage for speed
-
-If you need lower latency and have more storage/memory, disable recomputation. This stores full embeddings and avoids recomputing at search time.
-
-```bash
-# Build without recomputation (HNSW requires non-compact in this mode)
-leann build my-index --no-recompute --no-compact
-
-# Search without recomputation
-leann search my-index "your query" --no-recompute
-```
-
-When to use:
-- Extreme low latency requirements (high QPS, interactive assistants)
-- Read-heavy workloads where storage is cheaper than latency
-- No always-available GPU
-
-Constraints:
-- HNSW: when `--no-recompute` is set, LEANN automatically disables compact mode during build
-- DiskANN: supported; `--no-recompute` skips selective recompute during search
-
-Storage impact:
-- Storing N embeddings of dimension D with float32 requires approximately N × D × 4 bytes
-- Example: 1,000,000 chunks × 768 dims × 4 bytes ≈ 2.86 GB (plus graph/metadata)
-
-Converting an existing index (rebuild required):
-```bash
-# Rebuild in-place (ensure you still have original docs or can regenerate chunks)
-leann build my-index --force --no-recompute --no-compact
-```
-
-Python API usage:
-```python
-from leann import LeannSearcher
-
-searcher = LeannSearcher("/path/to/my-index.leann")
-results = searcher.search("your query", top_k=10, recompute_embeddings=False)
-```
-
-Trade-offs:
-- Lower latency and fewer network hops at query time
-- Significantly higher storage (10–100× vs selective recomputation)
-- Slightly larger memory footprint during build and search
-
-Quick benchmark results (`benchmarks/benchmark_no_recompute.py` with 5k texts, complexity=32):
-
-- HNSW
-
-  ```text
-  recompute=True:  search_time=0.818s, size=1.1MB
-  recompute=False: search_time=0.012s, size=16.6MB
-  ```
-
-- DiskANN
-
-  ```text
-  recompute=True:  search_time=0.041s, size=5.9MB
-  recompute=False: search_time=0.013s, size=24.6MB
-  ```
-
-Conclusion:
-- **HNSW**: `no-recompute` is significantly faster (no embedding recomputation) but requires much more storage (stores all embeddings)
-- **DiskANN**: `no-recompute` uses PQ + partial real distances during traversal (slower but higher accuracy), while `recompute=True` uses pure PQ traversal + final reranking (faster traversal, enables build-time partitioning for smaller storage)
-
-
-
-## Further Reading
-
-- [Lessons Learned Developing LEANN](https://yichuan-w.github.io/blog/lessons_learned_in_dev_leann/)
-- [LEANN Technical Paper](https://arxiv.org/abs/2506.08276)
-- [DiskANN Original Paper](https://papers.nips.cc/paper/2019/file/09853c7fb1d3f8ee67a61b6bf4a7f8e6-Paper.pdf)
-- [SSD-based Graph Partitioning](https://github.com/SonglinLife/SSD_BASED_PLAN)

docs/faq.md

@@ -1,10 +0,0 @@
-# FAQ
-
-## 1. My building time seems long
-
-You can speed up the process by using a lightweight embedding model. Add this to your arguments:
-
-```bash
---embedding-model sentence-transformers/all-MiniLM-L6-v2
-```
-**Model sizes:** `all-MiniLM-L6-v2` (30M parameters), `facebook/contriever` (~100M parameters), `Qwen3-0.6B` (600M parameters)

docs/features.md

@@ -1,23 +0,0 @@
-# ✨ Detailed Features
-
-## 🔥 Core Features
-
-- **🔄 Real-time Embeddings** - Eliminate heavy embedding storage with dynamic computation using optimized ZMQ servers and highly optimized search paradigm (overlapping and batching) with highly optimized embedding engine
-- **🧠 AST-Aware Code Chunking** - Intelligent code chunking that preserves semantic boundaries (functions, classes, methods) for Python, Java, C#, and TypeScript files
-- **📈 Scalable Architecture** - Handles millions of documents on consumer hardware; the larger your dataset, the more LEANN can save
-- **🎯 Graph Pruning** - Advanced techniques to minimize the storage overhead of vector search to a limited footprint
-- **🏗️ Pluggable Backends** - HNSW/FAISS (default), with optional DiskANN for large-scale deployments
-
-## 🛠️ Technical Highlights
-- **🔄 Recompute Mode** - Highest accuracy scenarios while eliminating vector storage overhead
-- **⚡ Zero-copy Operations** - Minimize IPC overhead by transferring distances instead of embeddings
-- **🚀 High-throughput Embedding Pipeline** - Optimized batched processing for maximum efficiency
-- **🎯 Two-level Search** - Novel coarse-to-fine search overlap for accelerated query processing (optional)
-- **💾 Memory-mapped Indices** - Fast startup with raw text mapping to reduce memory overhead
-- **🚀 MLX Support** - Ultra-fast recompute/build with quantized embedding models, accelerating building and search ([minimal example](../examples/mlx_demo.py))
-
-## 🎨 Developer Experience
-
-- **Simple Python API** - Get started in minutes
-- **Extensible backend system** - Easy to add new algorithms
-- **Comprehensive examples** - From basic usage to production deployment

docs/grep_search.md

@@ -1,149 +0,0 @@
-# LEANN Grep Search Usage Guide
-
-## Overview
-
-LEANN's grep search functionality provides exact text matching for finding specific code patterns, error messages, function names, or exact phrases in your indexed documents.
-
-## Basic Usage
-
-### Simple Grep Search
-
-```python
-from leann.api import LeannSearcher
-
-searcher = LeannSearcher("your_index_path")
-
-# Exact text search
-results = searcher.search("def authenticate_user", use_grep=True, top_k=5)
-
-for result in results:
-    print(f"Score: {result.score}")
-    print(f"Text: {result.text[:100]}...")
-    print("-" * 40)
-```
-
-### Comparison: Semantic vs Grep Search
-
-```python
-# Semantic search - finds conceptually similar content
-semantic_results = searcher.search("machine learning algorithms", top_k=3)
-
-# Grep search - finds exact text matches
-grep_results = searcher.search("def train_model", use_grep=True, top_k=3)
-```
-
-## When to Use Grep Search
-
-### Use Cases
-
-- **Code Search**: Finding specific function definitions, class names, or variable references
-- **Error Debugging**: Locating exact error messages or stack traces
-- **Documentation**: Finding specific API endpoints or exact terminology
-
-### Examples
-
-```python
-# Find function definitions
-functions = searcher.search("def __init__", use_grep=True)
-
-# Find import statements
-imports = searcher.search("from sklearn import", use_grep=True)
-
-# Find specific error types
-errors = searcher.search("FileNotFoundError", use_grep=True)
-
-# Find TODO comments
-todos = searcher.search("TODO:", use_grep=True)
-
-# Find configuration entries
-configs = searcher.search("server_port=", use_grep=True)
-```
-
-## Technical Details
-
-### How It Works
-
-1. **File Location**: Grep search operates on the raw text stored in `.jsonl` files
-2. **Command Execution**: Uses the system `grep` command with case-insensitive search
-3. **Result Processing**: Parses JSON lines and extracts text and metadata
-4. **Scoring**: Simple frequency-based scoring based on query term occurrences
-
-### Search Process
-
-```
-Query: "def train_model"
-  ↓
-grep -i -n "def train_model" documents.leann.passages.jsonl
-  ↓
-Parse matching JSON lines
-  ↓
-Calculate scores based on term frequency
-  ↓
-Return top_k results
-```
-
-### Scoring Algorithm
-
-```python
-# Term frequency in document
-score = text.lower().count(query.lower())
-```
-
-Results are ranked by score (highest first), with higher scores indicating more occurrences of the search term.
-
-## Error Handling
-
-### Common Issues
-
-#### Grep Command Not Found
-```
-RuntimeError: grep command not found. Please install grep or use semantic search.
-```
-
-**Solution**: Install grep on your system:
-- **Ubuntu/Debian**: `sudo apt-get install grep`
-- **macOS**: grep is pre-installed
-- **Windows**: Use WSL or install grep via Git Bash/MSYS2
-
-#### No Results Found
-```python
-# Check if your query exists in the raw data
-results = searcher.search("your_query", use_grep=True)
-if not results:
-    print("No exact matches found. Try:")
-    print("1. Check spelling and case")
-    print("2. Use partial terms")
-    print("3. Switch to semantic search")
-```
-
-## Complete Example
-
-```python
-#!/usr/bin/env python3
-"""
-Grep Search Example
-Demonstrates grep search for exact text matching.
-"""
-
-from leann.api import LeannSearcher
-
-def demonstrate_grep_search():
-    # Initialize searcher
-    searcher = LeannSearcher("my_index")
-
-    print("=== Function Search ===")
-    functions = searcher.search("def __init__", use_grep=True, top_k=5)
-    for i, result in enumerate(functions, 1):
-        print(f"{i}. Score: {result.score}")
-        print(f"   Preview: {result.text[:60]}...")
-        print()
-
-    print("=== Error Search ===")
-    errors = searcher.search("FileNotFoundError", use_grep=True, top_k=3)
-    for result in errors:
-        print(f"Content: {result.text.strip()}")
-        print("-" * 40)
-
-if __name__ == "__main__":
-    demonstrate_grep_search()
-```

docs/metadata_filtering.md

@@ -1,300 +0,0 @@
-# LEANN Metadata Filtering Usage Guide
-
-## Overview
-
-Leann possesses metadata filtering capabilities that allow you to filter search results based on arbitrary metadata fields set during chunking. This feature enables use cases like spoiler-free book search, document filtering by date/type, code search by file type, and potentially much more.
-
-## Basic Usage
-
-### Adding Metadata to Your Documents
-
-When building your index, add metadata to each text chunk:
-
-```python
-from leann.api import LeannBuilder
-
-builder = LeannBuilder("hnsw")
-
-# Add text with metadata
-builder.add_text(
-    text="Chapter 1: Alice falls down the rabbit hole",
-    metadata={
-        "chapter": 1,
-        "character": "Alice",
-        "themes": ["adventure", "curiosity"],
-        "word_count": 150
-    }
-)
-
-builder.build_index("alice_in_wonderland_index")
-```
-
-### Searching with Metadata Filters
-
-Use the `metadata_filters` parameter in search calls:
-
-```python
-from leann.api import LeannSearcher
-
-searcher = LeannSearcher("alice_in_wonderland_index")
-
-# Search with filters
-results = searcher.search(
-    query="What happens to Alice?",
-    top_k=10,
-    metadata_filters={
-        "chapter": {"<=": 5},           # Only chapters 1-5
-        "spoiler_level": {"!=": "high"} # No high spoilers
-    }
-)
-```
-
-## Filter Syntax
-
-### Basic Structure
-
-```python
-metadata_filters = {
-    "field_name": {"operator": value},
-    "another_field": {"operator": value}
-}
-```
-
-### Supported Operators
-
-#### Comparison Operators
-- `"=="`: Equal to
-- `"!="`: Not equal to
-- `"<"`: Less than
-- `"<="`: Less than or equal
-- `">"`: Greater than
-- `">="`: Greater than or equal
-
-```python
-# Examples
-{"chapter": {"==": 1}}           # Exactly chapter 1
-{"page": {">": 100}}            # Pages after 100
-{"rating": {">=": 4.0}}         # Rating 4.0 or higher
-{"word_count": {"<": 500}}      # Short passages
-```
-
-#### Membership Operators
-- `"in"`: Value is in list
-- `"not_in"`: Value is not in list
-
-```python
-# Examples
-{"character": {"in": ["Alice", "Bob"]}}      # Alice OR Bob
-{"genre": {"not_in": ["horror", "thriller"]}} # Exclude genres
-{"tags": {"in": ["fiction", "adventure"]}}   # Any of these tags
-```
-
-#### String Operators
-- `"contains"`: String contains substring
-- `"starts_with"`: String starts with prefix
-- `"ends_with"`: String ends with suffix
-
-```python
-# Examples
-{"title": {"contains": "alice"}}        # Title contains "alice"
-{"filename": {"ends_with": ".py"}}      # Python files
-{"author": {"starts_with": "Dr."}}      # Authors with "Dr." prefix
-```
-
-#### Boolean Operators
-- `"is_true"`: Field is truthy
-- `"is_false"`: Field is falsy
-
-```python
-# Examples
-{"is_published": {"is_true": True}}     # Published content
-{"is_draft": {"is_false": False}}       # Not drafts
-```
-
-### Multiple Operators on Same Field
-
-You can apply multiple operators to the same field (AND logic):
-
-```python
-metadata_filters = {
-    "word_count": {
-        ">=": 100,    # At least 100 words
-        "<=": 500     # At most 500 words
-    }
-}
-```
-
-### Compound Filters
-
-Multiple fields are combined with AND logic:
-
-```python
-metadata_filters = {
-    "chapter": {"<=": 10},              # Up to chapter 10
-    "character": {"==": "Alice"},       # About Alice
-    "spoiler_level": {"!=": "high"}     # No major spoilers
-}
-```
-
-## Use Case Examples
-
-### 1. Spoiler-Free Book Search
-
-```python
-# Reader has only read up to chapter 5
-def search_spoiler_free(query, max_chapter):
-    return searcher.search(
-        query=query,
-        metadata_filters={
-            "chapter": {"<=": max_chapter},
-            "spoiler_level": {"in": ["none", "low"]}
-        }
-    )
-
-results = search_spoiler_free("What happens to Alice?", max_chapter=5)
-```
-
-### 2. Document Management by Date
-
-```python
-# Find recent documents
-recent_docs = searcher.search(
-    query="project updates",
-    metadata_filters={
-        "date": {">=": "2024-01-01"},
-        "document_type": {"==": "report"}
-    }
-)
-```
-
-### 3. Code Search by File Type
-
-```python
-# Search only Python files
-python_code = searcher.search(
-    query="authentication function",
-    metadata_filters={
-        "file_extension": {"==": ".py"},
-        "lines_of_code": {"<": 100}
-    }
-)
-```
-
-### 4. Content Filtering by Audience
-
-```python
-# Age-appropriate content
-family_content = searcher.search(
-    query="adventure stories",
-    metadata_filters={
-        "age_rating": {"in": ["G", "PG"]},
-        "content_warnings": {"not_in": ["violence", "adult_themes"]}
-    }
-)
-```
-
-### 5. Multi-Book Series Management
-
-```python
-# Search across first 3 books only
-early_series = searcher.search(
-    query="character development",
-    metadata_filters={
-        "series": {"==": "Harry Potter"},
-        "book_number": {"<=": 3}
-    }
-)
-```
-
-## Running the Example
-
-You can see metadata filtering in action with our spoiler-free book RAG example:
-
-```bash
-# Don't forget to set up the environment
-uv venv
-source .venv/bin/activate
-
-# Set your OpenAI API key (required for embeddings, but you can update the example locally and use ollama instead)
-export OPENAI_API_KEY="your-api-key-here"
-
-# Run the spoiler-free book RAG example
-uv run examples/spoiler_free_book_rag.py
-```
-
-This example demonstrates:
-- Building an index with metadata (chapter numbers, characters, themes, locations)
-- Searching with filters to avoid spoilers (e.g., only show results up to chapter 5)
-- Different scenarios for readers at various points in the book
-
-The example uses Alice's Adventures in Wonderland as sample data and shows how you can search for information without revealing plot points from later chapters.
-
-## Advanced Patterns
-
-### Custom Chunking with metadata
-
-```python
-def chunk_book_with_metadata(book_text, book_info):
-    chunks = []
-
-    for chapter_num, chapter_text in parse_chapters(book_text):
-        # Extract entities, themes, etc.
-        characters = extract_characters(chapter_text)
-        themes = classify_themes(chapter_text)
-        spoiler_level = assess_spoiler_level(chapter_text, chapter_num)
-
-        # Create chunks with rich metadata
-        for paragraph in split_paragraphs(chapter_text):
-            chunks.append({
-                "text": paragraph,
-                "metadata": {
-                    "book_title": book_info["title"],
-                    "chapter": chapter_num,
-                    "characters": characters,
-                    "themes": themes,
-                    "spoiler_level": spoiler_level,
-                    "word_count": len(paragraph.split()),
-                    "reading_level": calculate_reading_level(paragraph)
-                }
-            })
-
-    return chunks
-```
-
-## Performance Considerations
-
-### Efficient Filtering Strategies
-
-1. **Post-search filtering**: Applies filters after vector search, which should be efficient for typical result sets (10-100 results).
-
-2. **Metadata design**: Keep metadata fields simple and avoid deeply nested structures.
-
-### Best Practices
-
-1. **Consistent metadata schema**: Use consistent field names and value types across your documents.
-
-2. **Reasonable metadata size**: Keep metadata reasonably sized to avoid storage overhead.
-
-3. **Type consistency**: Use consistent data types for the same fields (e.g., always integers for chapter numbers).
-
-4. **Index multiple granularities**: Consider chunking at different levels (paragraph, section, chapter) with appropriate metadata.
-
-### Adding Metadata to Existing Indices
-
-To add metadata filtering to existing indices, you'll need to rebuild them with metadata:
-
-```python
-# Read existing passages and add metadata
-def add_metadata_to_existing_chunks(chunks):
-    for chunk in chunks:
-        # Extract or assign metadata based on content
-        chunk["metadata"] = extract_metadata(chunk["text"])
-    return chunks
-
-# Rebuild index with metadata
-enhanced_chunks = add_metadata_to_existing_chunks(existing_chunks)
-builder = LeannBuilder("hnsw")
-for chunk in enhanced_chunks:
-    builder.add_text(chunk["text"], chunk["metadata"])
-builder.build_index("enhanced_index")
-```

docs/normalized_embeddings.md

@@ -1,75 +0,0 @@
-# Normalized Embeddings Support in LEANN
-
-LEANN now automatically detects normalized embedding models and sets the appropriate distance metric for optimal performance.
-
-## What are Normalized Embeddings?
-
-Normalized embeddings are vectors with L2 norm = 1 (unit vectors). These embeddings are optimized for cosine similarity rather than Maximum Inner Product Search (MIPS).
-
-## Automatic Detection
-
-When you create a `LeannBuilder` instance with a normalized embedding model, LEANN will:
-
-1. **Automatically set `distance_metric="cosine"`** if not specified
-2. **Show a warning** if you manually specify a different distance metric
-3. **Provide optimal search performance** with the correct metric
-
-## Supported Normalized Embedding Models
-
-### OpenAI
-All OpenAI text embedding models are normalized:
-- `text-embedding-ada-002`
-- `text-embedding-3-small`
-- `text-embedding-3-large`
-
-### Voyage AI
-All Voyage AI embedding models are normalized:
-- `voyage-2`
-- `voyage-3`
-- `voyage-large-2`
-- `voyage-multilingual-2`
-- `voyage-code-2`
-
-### Cohere
-All Cohere embedding models are normalized:
-- `embed-english-v3.0`
-- `embed-multilingual-v3.0`
-- `embed-english-light-v3.0`
-- `embed-multilingual-light-v3.0`
-
-## Example Usage
-
-```python
-from leann.api import LeannBuilder
-
-# Automatic detection - will use cosine distance
-builder = LeannBuilder(
-    backend_name="hnsw",
-    embedding_model="text-embedding-3-small",
-    embedding_mode="openai"
-)
-# Warning: Detected normalized embeddings model 'text-embedding-3-small'...
-# Automatically setting distance_metric='cosine'
-
-# Manual override (not recommended)
-builder = LeannBuilder(
-    backend_name="hnsw",
-    embedding_model="text-embedding-3-small",
-    embedding_mode="openai",
-    distance_metric="mips"  # Will show warning
-)
-# Warning: Using 'mips' distance metric with normalized embeddings...
-```
-
-## Non-Normalized Embeddings
-
-Models like `facebook/contriever` and other sentence-transformers models that are not normalized will continue to use MIPS by default, which is optimal for them.
-
-## Why This Matters
-
-Using the wrong distance metric with normalized embeddings can lead to:
-- **Poor search quality** due to HNSW's early termination with narrow score ranges
-- **Incorrect ranking** of search results
-- **Suboptimal performance** compared to using the correct metric
-
-For more details on why this happens, see our analysis in the [embedding detection code](../packages/leann-core/src/leann/api.py) which automatically handles normalized embeddings and MIPS distance metric issues.

docs/roadmap.md

@@ -1,21 +0,0 @@
-# 📈 Roadmap
-
-## 🎯 Q2 2025
-
-- [X] HNSW backend integration
-- [X] DiskANN backend with MIPS/L2/Cosine support
-- [X] Real-time embedding pipeline
-- [X] Memory-efficient graph pruning
-
-## 🚀 Q3 2025
-
-- [ ] Advanced caching strategies
-- [ ] Add contextual-retrieval https://www.anthropic.com/news/contextual-retrieval
-- [ ] Add sleep-time-compute and summarize agent! to summarilze the file on computer!
-- [ ] Add OpenAI recompute API
-
-## 🌟 Q4 2025
-
-- [ ] Integration with LangChain/LlamaIndex
-- [ ] Visual similarity search
-- [ ] Query rewrtiting, rerank and expansion

examples/LEANN_email_reader.py

@@ -0,0 +1,130 @@
+import os
+import email
+from pathlib import Path
+from typing import List, Any
+from llama_index.core import Document
+from llama_index.core.readers.base import BaseReader
+
+class EmlxReader(BaseReader):
+    """
+    Apple Mail .emlx file reader with embedded metadata.
+    
+    Reads individual .emlx files from Apple Mail's storage format.
+    """
+    
+    def __init__(self) -> None:
+        """Initialize."""
+        pass
+    
+    def load_data(self, input_dir: str, **load_kwargs: Any) -> List[Document]:
+        """
+        Load data from the input directory containing .emlx files.
+        
+        Args:
+            input_dir: Directory containing .emlx files
+            **load_kwargs:
+                max_count (int): Maximum amount of messages to read.
+        """
+        docs: List[Document] = []
+        max_count = load_kwargs.get('max_count', 1000)
+        count = 0
+        
+        # Walk through the directory recursively
+        for dirpath, dirnames, filenames in os.walk(input_dir):
+            # Skip hidden directories
+            dirnames[:] = [d for d in dirnames if not d.startswith(".")]
+            
+            for filename in filenames:
+                if count >= max_count:
+                    break
+                    
+                if filename.endswith(".emlx"):
+                    filepath = os.path.join(dirpath, filename)
+                    try:
+                        # Read the .emlx file
+                        with open(filepath, 'r', encoding='utf-8', errors='ignore') as f:
+                            content = f.read()
+                        
+                        # .emlx files have a length prefix followed by the email content
+                        # The first line contains the length, followed by the email
+                        lines = content.split('\n', 1)
+                        if len(lines) >= 2:
+                            email_content = lines[1]
+                            
+                            # Parse the email using Python's email module
+                            try:
+                                msg = email.message_from_string(email_content)
+                                
+                                # Extract email metadata
+                                subject = msg.get('Subject', 'No Subject')
+                                from_addr = msg.get('From', 'Unknown')
+                                to_addr = msg.get('To', 'Unknown')
+                                date = msg.get('Date', 'Unknown')
+                                
+                                # Extract email body
+                                body = ""
+                                if msg.is_multipart():
+                                    for part in msg.walk():
+                                        if part.get_content_type() == "text/plain" or part.get_content_type() == "text/html":
+                                            # if part.get_content_type() == "text/html":
+                                            #     continue
+                                            body += part.get_payload(decode=True).decode('utf-8', errors='ignore')
+                                            # break
+                                else:
+                                    body = msg.get_payload(decode=True).decode('utf-8', errors='ignore')
+                                
+                                # Create document content with metadata embedded in text
+                                doc_content = f"""
+[EMAIL METADATA]
+File: {filename}
+From: {from_addr}
+To: {to_addr}
+Subject: {subject}
+Date: {date}
+[END METADATA]
+
+{body}
+"""
+                                
+                                # No separate metadata - everything is in the text
+                                doc = Document(text=doc_content, metadata={})
+                                docs.append(doc)
+                                count += 1
+                                
+                            except Exception as e:
+                                print(f"Error parsing email from {filepath}: {e}")
+                                continue
+                                
+                    except Exception as e:
+                        print(f"Error reading file {filepath}: {e}")
+                        continue
+        
+        print(f"Loaded {len(docs)} email documents")
+        return docs
+
+    @staticmethod
+    def find_all_messages_directories(base_path: str) -> List[Path]:
+        """
+        Find all Messages directories under the given base path.
+        
+        Args:
+            base_path: Base path to search for Messages directories
+            
+        Returns:
+            List of Path objects pointing to Messages directories
+        """
+        base_path_obj = Path(base_path)
+        messages_dirs = []
+        
+        if not base_path_obj.exists():
+            print(f"Base path {base_path} does not exist")
+            return messages_dirs
+        
+        # Find all Messages directories recursively
+        for messages_dir in base_path_obj.rglob("Messages"):
+            if messages_dir.is_dir():
+                messages_dirs.append(messages_dir)
+                print(f"Found Messages directory: {messages_dir}")
+        
+        print(f"Found {len(messages_dirs)} Messages directories")
+        return messages_dirs 

examples/document_search.py

@@ -0,0 +1,146 @@
+#!/usr/bin/env python3
+"""
+Document search demo with recompute mode
+"""
+
+import os
+from pathlib import Path
+import shutil
+import time
+
+# Import backend packages to trigger plugin registration
+try:
+    import leann_backend_diskann
+    import leann_backend_hnsw
+    print("INFO: Backend packages imported successfully.")
+except ImportError as e:
+    print(f"WARNING: Could not import backend packages. Error: {e}")
+
+# Import upper-level API from leann-core
+from leann.api import LeannBuilder, LeannSearcher, LeannChat
+
+
+def load_sample_documents():
+    """Create sample documents for demonstration"""
+    docs = [
+        {"title": "Intro to Python", "content": "Python is a high-level, interpreted language known for simplicity."},
+        {"title": "ML Basics", "content": "Machine learning builds systems that learn from data."},
+        {"title": "Data Structures", "content": "Data structures like arrays, lists, and graphs organize data."},
+    ]
+    return docs
+
+def main():
+    print("==========================================================")
+    print("=== Leann Document Search Demo (DiskANN + Recompute) ===")
+    print("==========================================================")
+    
+    INDEX_DIR = Path("./test_indices")
+    INDEX_PATH = str(INDEX_DIR / "documents.diskann")
+    BACKEND_TO_TEST = "diskann"
+
+    if INDEX_DIR.exists():
+        print(f"--- Cleaning up old index directory: {INDEX_DIR} ---")
+        shutil.rmtree(INDEX_DIR)
+
+    # --- 1. Build index ---
+    print(f"\n[PHASE 1] Building index using '{BACKEND_TO_TEST}' backend...")
+    
+    builder = LeannBuilder(
+        backend_name=BACKEND_TO_TEST, 
+        graph_degree=32, 
+        complexity=64
+    )
+    
+    documents = load_sample_documents()
+    print(f"Loaded {len(documents)} sample documents.")
+    for doc in documents:
+        builder.add_text(doc["content"], metadata={"title": doc["title"]})
+        
+    builder.build_index(INDEX_PATH)
+    print(f"\nIndex built!")
+
+    # --- 2. Basic search demo ---
+    print(f"\n[PHASE 2] Basic search using '{BACKEND_TO_TEST}' backend...")
+    searcher = LeannSearcher(index_path=INDEX_PATH)
+    
+    query = "What is machine learning?"
+    print(f"\nQuery: '{query}'")
+    
+    print("\n--- Basic search mode (PQ computation) ---")
+    start_time = time.time()
+    results = searcher.search(query, top_k=2)
+    basic_time = time.time() - start_time
+    
+    print(f"⏱️  Basic search time: {basic_time:.3f} seconds")
+    print(">>> Basic search results <<<")
+    for i, res in enumerate(results, 1):
+        print(f"  {i}. ID: {res.id}, Score: {res.score:.4f}, Text: '{res.text}', Metadata: {res.metadata}")
+
+    # --- 3. Recompute search demo ---
+    print(f"\n[PHASE 3] Recompute search using embedding server...")
+    
+    print("\n--- Recompute search mode (get real embeddings via network) ---")
+    
+    # Configure recompute parameters
+    recompute_params = {
+        "recompute_beighbor_embeddings": True,  # Enable network recomputation
+        "USE_DEFERRED_FETCH": False,           # Don't use deferred fetch
+        "skip_search_reorder": True,           # Skip search reordering
+        "dedup_node_dis": True,               # Enable node distance deduplication
+        "prune_ratio": 0.1,                   # Pruning ratio 10%
+        "batch_recompute": False,             # Don't use batch recomputation
+        "global_pruning": False,              # Don't use global pruning
+        "zmq_port": 5555,                     # ZMQ port
+        "embedding_model": "sentence-transformers/all-mpnet-base-v2"
+    }
+    
+    print("Recompute parameter configuration:")
+    for key, value in recompute_params.items():
+        print(f"  {key}: {value}")
+    
+    print(f"\n🔄 Executing Recompute search...")
+    try:
+        start_time = time.time()
+        recompute_results = searcher.search(query, top_k=2, **recompute_params)
+        recompute_time = time.time() - start_time
+        
+        print(f"⏱️  Recompute search time: {recompute_time:.3f} seconds")
+        print(">>> Recompute search results <<<")
+        for i, res in enumerate(recompute_results, 1):
+            print(f"  {i}. ID: {res.id}, Score: {res.score:.4f}, Text: '{res.text}', Metadata: {res.metadata}")
+        
+        # Compare results
+        print(f"\n--- Result comparison ---")
+        print(f"Basic search time: {basic_time:.3f} seconds")
+        print(f"Recompute time: {recompute_time:.3f} seconds")
+        
+        print("\nBasic search vs Recompute results:")
+        for i in range(min(len(results), len(recompute_results))):
+            basic_score = results[i].score
+            recompute_score = recompute_results[i].score
+            score_diff = abs(basic_score - recompute_score)
+            print(f"  Position {i+1}: PQ={basic_score:.4f}, Recompute={recompute_score:.4f}, Difference={score_diff:.4f}")
+        
+        if recompute_time > basic_time:
+            print(f"✅ Recompute mode working correctly (more accurate but slower)")
+        else:
+            print(f"ℹ️  Recompute time is unusually fast, network recomputation may not be enabled")
+            
+    except Exception as e:
+        print(f"❌ Recompute search failed: {e}")
+        print("This usually indicates an embedding server connection issue")
+
+    # --- 4. Chat demo ---
+    print(f"\n[PHASE 4] Starting chat session...")
+    chat = LeannChat(index_path=INDEX_PATH)
+    chat_response = chat.ask(query)
+    print(f"You: {query}")
+    print(f"Leann: {chat_response}")
+
+    print("\n==========================================================")
+    print("✅ Demo finished successfully!")
+    print("==========================================================")
+
+
+if __name__ == "__main__":
+    main()

examples/dynamic_update_no_recompute.py

@@ -1,404 +0,0 @@
-"""Dynamic HNSW update demo without compact storage.
-
-This script reproduces the minimal scenario we used while debugging on-the-fly
-recompute:
-
-1. Build a non-compact HNSW index from the first few paragraphs of a text file.
-2. Print the top results with `recompute_embeddings=True`.
-3. Append additional paragraphs with :meth:`LeannBuilder.update_index`.
-4. Run the same query again to show the newly inserted passages.
-
-Run it with ``uv`` (optionally pointing LEANN_HNSW_LOG_PATH at a file to inspect
-ZMQ activity)::
-
-    LEANN_HNSW_LOG_PATH=embedding_fetch.log \
-    uv run -m examples.dynamic_update_no_recompute \
-      --index-path .leann/examples/leann-demo.leann
-
-By default the script builds an index from ``data/2501.14312v1 (1).pdf`` and
-then updates it with LEANN-related material from ``data/2506.08276v1.pdf``.
-It issues the query "What's LEANN?" before and after the update to show how the
-new passages become immediately searchable. The script uses the
-``sentence-transformers/all-MiniLM-L6-v2`` model with ``is_recompute=True`` so
-Faiss pulls existing vectors on demand via the ZMQ embedding server, while
-freshly added passages are embedded locally just like the initial build.
-
-To make storage comparisons easy, the script can also build a matching
-``is_recompute=False`` baseline (enabled by default) and report the index size
-delta after the update. Disable the baseline run with
-``--skip-compare-no-recompute`` if you only need the recompute flow.
-"""
-
-import argparse
-import json
-from collections.abc import Iterable
-from pathlib import Path
-from typing import Any
-
-from leann.api import LeannBuilder, LeannSearcher
-from leann.registry import register_project_directory
-
-from apps.chunking import create_text_chunks
-
-REPO_ROOT = Path(__file__).resolve().parents[1]
-
-DEFAULT_QUERY = "What's LEANN?"
-DEFAULT_INITIAL_FILES = [REPO_ROOT / "data" / "2501.14312v1 (1).pdf"]
-DEFAULT_UPDATE_FILES = [REPO_ROOT / "data" / "2506.08276v1.pdf"]
-
-
-def load_chunks_from_files(paths: list[Path]) -> list[str]:
-    from llama_index.core import SimpleDirectoryReader
-
-    documents = []
-    for path in paths:
-        p = path.expanduser().resolve()
-        if not p.exists():
-            raise FileNotFoundError(f"Input path not found: {p}")
-        if p.is_dir():
-            reader = SimpleDirectoryReader(str(p), recursive=False)
-            documents.extend(reader.load_data(show_progress=True))
-        else:
-            reader = SimpleDirectoryReader(input_files=[str(p)])
-            documents.extend(reader.load_data(show_progress=True))
-
-    if not documents:
-        return []
-
-    chunks = create_text_chunks(
-        documents,
-        chunk_size=512,
-        chunk_overlap=128,
-        use_ast_chunking=False,
-    )
-    return [c for c in chunks if isinstance(c, str) and c.strip()]
-
-
-def run_search(index_path: Path, query: str, top_k: int, *, recompute_embeddings: bool) -> list:
-    searcher = LeannSearcher(str(index_path))
-    try:
-        return searcher.search(
-            query=query,
-            top_k=top_k,
-            recompute_embeddings=recompute_embeddings,
-            batch_size=16,
-        )
-    finally:
-        searcher.cleanup()
-
-
-def print_results(title: str, results: Iterable) -> None:
-    print(f"\n=== {title} ===")
-    res_list = list(results)
-    print(f"results count: {len(res_list)}")
-    print("passages:")
-    if not res_list:
-        print("  (no passages returned)")
-    for res in res_list:
-        snippet = res.text.replace("\n", " ")[:120]
-        print(f"  - {res.id}: {snippet}... (score={res.score:.4f})")
-
-
-def build_initial_index(
-    index_path: Path,
-    paragraphs: list[str],
-    model_name: str,
-    embedding_mode: str,
-    is_recompute: bool,
-) -> None:
-    builder = LeannBuilder(
-        backend_name="hnsw",
-        embedding_model=model_name,
-        embedding_mode=embedding_mode,
-        is_compact=False,
-        is_recompute=is_recompute,
-    )
-    for idx, passage in enumerate(paragraphs):
-        builder.add_text(passage, metadata={"id": str(idx)})
-    builder.build_index(str(index_path))
-
-
-def update_index(
-    index_path: Path,
-    start_id: int,
-    paragraphs: list[str],
-    model_name: str,
-    embedding_mode: str,
-    is_recompute: bool,
-) -> None:
-    updater = LeannBuilder(
-        backend_name="hnsw",
-        embedding_model=model_name,
-        embedding_mode=embedding_mode,
-        is_compact=False,
-        is_recompute=is_recompute,
-    )
-    for offset, passage in enumerate(paragraphs, start=start_id):
-        updater.add_text(passage, metadata={"id": str(offset)})
-    updater.update_index(str(index_path))
-
-
-def ensure_index_dir(index_path: Path) -> None:
-    index_path.parent.mkdir(parents=True, exist_ok=True)
-
-
-def cleanup_index_files(index_path: Path) -> None:
-    """Remove leftover index artifacts for a clean rebuild."""
-
-    parent = index_path.parent
-    if not parent.exists():
-        return
-    stem = index_path.stem
-    for file in parent.glob(f"{stem}*"):
-        if file.is_file():
-            file.unlink()
-
-
-def index_file_size(index_path: Path) -> int:
-    """Return the size of the primary .index file for the given index path."""
-
-    index_file = index_path.parent / f"{index_path.stem}.index"
-    return index_file.stat().st_size if index_file.exists() else 0
-
-
-def load_metadata_snapshot(index_path: Path) -> dict[str, Any] | None:
-    meta_path = index_path.parent / f"{index_path.name}.meta.json"
-    if not meta_path.exists():
-        return None
-    try:
-        return json.loads(meta_path.read_text())
-    except json.JSONDecodeError:
-        return None
-
-
-def run_workflow(
-    *,
-    label: str,
-    index_path: Path,
-    initial_paragraphs: list[str],
-    update_paragraphs: list[str],
-    model_name: str,
-    embedding_mode: str,
-    is_recompute: bool,
-    query: str,
-    top_k: int,
-) -> dict[str, Any]:
-    prefix = f"[{label}] " if label else ""
-
-    ensure_index_dir(index_path)
-    cleanup_index_files(index_path)
-
-    print(f"{prefix}Building initial index...")
-    build_initial_index(
-        index_path,
-        initial_paragraphs,
-        model_name,
-        embedding_mode,
-        is_recompute=is_recompute,
-    )
-
-    initial_size = index_file_size(index_path)
-    before_results = run_search(
-        index_path,
-        query,
-        top_k,
-        recompute_embeddings=is_recompute,
-    )
-
-    print(f"\n{prefix}Updating index with additional passages...")
-    update_index(
-        index_path,
-        start_id=len(initial_paragraphs),
-        paragraphs=update_paragraphs,
-        model_name=model_name,
-        embedding_mode=embedding_mode,
-        is_recompute=is_recompute,
-    )
-
-    after_results = run_search(
-        index_path,
-        query,
-        top_k,
-        recompute_embeddings=is_recompute,
-    )
-    updated_size = index_file_size(index_path)
-
-    return {
-        "initial_size": initial_size,
-        "updated_size": updated_size,
-        "delta": updated_size - initial_size,
-        "before_results": before_results,
-        "after_results": after_results,
-        "metadata": load_metadata_snapshot(index_path),
-    }
-
-
-def main() -> None:
-    parser = argparse.ArgumentParser(description=__doc__)
-    parser.add_argument(
-        "--initial-files",
-        type=Path,
-        nargs="+",
-        default=DEFAULT_INITIAL_FILES,
-        help="Initial document files (PDF/TXT) used to build the base index",
-    )
-    parser.add_argument(
-        "--index-path",
-        type=Path,
-        default=Path(".leann/examples/leann-demo.leann"),
-        help="Destination index path (default: .leann/examples/leann-demo.leann)",
-    )
-    parser.add_argument(
-        "--initial-count",
-        type=int,
-        default=8,
-        help="Number of chunks to use from the initial documents (default: 8)",
-    )
-    parser.add_argument(
-        "--update-files",
-        type=Path,
-        nargs="*",
-        default=DEFAULT_UPDATE_FILES,
-        help="Additional documents to add during update (PDF/TXT)",
-    )
-    parser.add_argument(
-        "--update-count",
-        type=int,
-        default=4,
-        help="Number of chunks to append from update documents (default: 4)",
-    )
-    parser.add_argument(
-        "--update-text",
-        type=str,
-        default=(
-            "LEANN (Lightweight Embedding ANN) is an indexing toolkit focused on "
-            "recompute-aware HNSW graphs, allowing embeddings to be regenerated "
-            "on demand to keep disk usage minimal."
-        ),
-        help="Fallback text to append if --update-files is omitted",
-    )
-    parser.add_argument(
-        "--top-k",
-        type=int,
-        default=4,
-        help="Number of results to show for each search (default: 4)",
-    )
-    parser.add_argument(
-        "--query",
-        type=str,
-        default=DEFAULT_QUERY,
-        help="Query to run before/after the update",
-    )
-    parser.add_argument(
-        "--embedding-model",
-        type=str,
-        default="sentence-transformers/all-MiniLM-L6-v2",
-        help="Embedding model name",
-    )
-    parser.add_argument(
-        "--embedding-mode",
-        type=str,
-        default="sentence-transformers",
-        choices=["sentence-transformers", "openai", "mlx", "ollama"],
-        help="Embedding backend mode",
-    )
-    parser.add_argument(
-        "--compare-no-recompute",
-        dest="compare_no_recompute",
-        action="store_true",
-        help="Also run a baseline with is_recompute=False and report its index growth.",
-    )
-    parser.add_argument(
-        "--skip-compare-no-recompute",
-        dest="compare_no_recompute",
-        action="store_false",
-        help="Skip building the no-recompute baseline.",
-    )
-    parser.set_defaults(compare_no_recompute=True)
-    args = parser.parse_args()
-
-    ensure_index_dir(args.index_path)
-    register_project_directory(REPO_ROOT)
-
-    initial_chunks = load_chunks_from_files(list(args.initial_files))
-    if not initial_chunks:
-        raise ValueError("No text chunks extracted from the initial files.")
-
-    initial = initial_chunks[: args.initial_count]
-    if not initial:
-        raise ValueError("Initial chunk set is empty after applying --initial-count.")
-
-    if args.update_files:
-        update_chunks = load_chunks_from_files(list(args.update_files))
-        if not update_chunks:
-            raise ValueError("No text chunks extracted from the update files.")
-        to_add = update_chunks[: args.update_count]
-    else:
-        if not args.update_text:
-            raise ValueError("Provide --update-files or --update-text for the update step.")
-        to_add = [args.update_text]
-    if not to_add:
-        raise ValueError("Update chunk set is empty after applying --update-count.")
-
-    recompute_stats = run_workflow(
-        label="recompute",
-        index_path=args.index_path,
-        initial_paragraphs=initial,
-        update_paragraphs=to_add,
-        model_name=args.embedding_model,
-        embedding_mode=args.embedding_mode,
-        is_recompute=True,
-        query=args.query,
-        top_k=args.top_k,
-    )
-
-    print_results("initial search", recompute_stats["before_results"])
-    print_results("after update", recompute_stats["after_results"])
-    print(
-        f"\n[recompute] Index file size change: {recompute_stats['initial_size']} -> {recompute_stats['updated_size']} bytes"
-        f" (Δ {recompute_stats['delta']})"
-    )
-
-    if recompute_stats["metadata"]:
-        meta_view = {k: recompute_stats["metadata"].get(k) for k in ("is_compact", "is_pruned")}
-        print("[recompute] metadata snapshot:")
-        print(json.dumps(meta_view, indent=2))
-
-    if args.compare_no_recompute:
-        baseline_path = (
-            args.index_path.parent / f"{args.index_path.stem}-norecompute{args.index_path.suffix}"
-        )
-        baseline_stats = run_workflow(
-            label="no-recompute",
-            index_path=baseline_path,
-            initial_paragraphs=initial,
-            update_paragraphs=to_add,
-            model_name=args.embedding_model,
-            embedding_mode=args.embedding_mode,
-            is_recompute=False,
-            query=args.query,
-            top_k=args.top_k,
-        )
-
-        print(
-            f"\n[no-recompute] Index file size change: {baseline_stats['initial_size']} -> {baseline_stats['updated_size']} bytes"
-            f" (Δ {baseline_stats['delta']})"
-        )
-
-        after_texts = [res.text for res in recompute_stats["after_results"]]
-        baseline_after_texts = [res.text for res in baseline_stats["after_results"]]
-        if after_texts == baseline_after_texts:
-            print(
-                "[no-recompute] Search results match recompute baseline; see above for the shared output."
-            )
-        else:
-            print("[no-recompute] WARNING: search results differ from recompute baseline.")
-
-        if baseline_stats["metadata"]:
-            meta_view = {k: baseline_stats["metadata"].get(k) for k in ("is_compact", "is_pruned")}
-            print("[no-recompute] metadata snapshot:")
-            print(json.dumps(meta_view, indent=2))
-
-
-if __name__ == "__main__":
-    main()

examples/email_data/email.py

@@ -7,9 +7,9 @@ Contains simple parser for mbox files.
 
 import logging
 from pathlib import Path
-from typing import Any
+from typing import Any, Dict, List, Optional
-
 from fsspec import AbstractFileSystem
+
 from llama_index.core.readers.base import BaseReader
 from llama_index.core.schema import Document
 
@@ -27,7 +27,11 @@ class MboxReader(BaseReader):
     """
 
     DEFAULT_MESSAGE_FORMAT: str = (
-        "Date: {_date}\nFrom: {_from}\nTo: {_to}\nSubject: {_subject}\nContent: {_content}"
+        "Date: {_date}\n"
+        "From: {_from}\n"
+        "To: {_to}\n"
+        "Subject: {_subject}\n"
+        "Content: {_content}"
     )
 
     def __init__(
@@ -41,7 +45,9 @@ class MboxReader(BaseReader):
         try:
             from bs4 import BeautifulSoup  # noqa
         except ImportError:
-            raise ImportError("`beautifulsoup4` package not found: `pip install beautifulsoup4`")
+            raise ImportError(
+                "`beautifulsoup4` package not found: `pip install beautifulsoup4`"
+            )
 
         super().__init__(*args, **kwargs)
         self.max_count = max_count
@@ -50,9 +56,9 @@ class MboxReader(BaseReader):
     def load_data(
         self,
         file: Path,
-        extra_info: dict | None = None,
+        extra_info: Optional[Dict] = None,
-        fs: AbstractFileSystem | None = None,
+        fs: Optional[AbstractFileSystem] = None,
-    ) -> list[Document]:
+    ) -> List[Document]:
         """Parse file into string."""
         # Import required libraries
         import mailbox
@@ -68,7 +74,7 @@ class MboxReader(BaseReader):
             )
 
         i = 0
-        results: list[str] = []
+        results: List[str] = []
         # Load file using mailbox
         bytes_parser = BytesParser(policy=default).parse
         mbox = mailbox.mbox(file, factory=bytes_parser)  # type: ignore
@@ -118,7 +124,7 @@ class MboxReader(BaseReader):
 class EmlxMboxReader(MboxReader):
     """
     EmlxMboxReader - Modified MboxReader that handles directories of .emlx files.
-
+    
     Extends MboxReader to work with Apple Mail's .emlx format by:
     1. Reading .emlx files from a directory
     2. Converting them to mbox format in memory
@@ -128,13 +134,13 @@ class EmlxMboxReader(MboxReader):
     def load_data(
         self,
         directory: Path,
-        extra_info: dict | None = None,
+        extra_info: Optional[Dict] = None,
-        fs: AbstractFileSystem | None = None,
+        fs: Optional[AbstractFileSystem] = None,
-    ) -> list[Document]:
+    ) -> List[Document]:
         """Parse .emlx files from directory into strings using MboxReader logic."""
-        import os
         import tempfile
-
+        import os
+        
         if fs:
             logger.warning(
                 "fs was specified but EmlxMboxReader doesn't support loading "
@@ -144,37 +150,37 @@ class EmlxMboxReader(MboxReader):
         # Find all .emlx files in the directory
         emlx_files = list(directory.glob("*.emlx"))
         logger.info(f"Found {len(emlx_files)} .emlx files in {directory}")
-
+        
         if not emlx_files:
             logger.warning(f"No .emlx files found in {directory}")
             return []
 
         # Create a temporary mbox file
-        with tempfile.NamedTemporaryFile(mode="w", suffix=".mbox", delete=False) as temp_mbox:
+        with tempfile.NamedTemporaryFile(mode='w', suffix='.mbox', delete=False) as temp_mbox:
             temp_mbox_path = temp_mbox.name
-
+            
             # Convert .emlx files to mbox format
             for emlx_file in emlx_files:
                 try:
                     # Read the .emlx file
-                    with open(emlx_file, encoding="utf-8", errors="ignore") as f:
+                    with open(emlx_file, 'r', encoding='utf-8', errors='ignore') as f:
                         content = f.read()
-
+                    
                     # .emlx format: first line is length, rest is email content
-                    lines = content.split("\n", 1)
+                    lines = content.split('\n', 1)
                     if len(lines) >= 2:
                         email_content = lines[1]  # Skip the length line
-
+                        
                         # Write to mbox format (each message starts with "From " and ends with blank line)
                         temp_mbox.write(f"From {emlx_file.name} {email_content}\n\n")
-
+                    
                 except Exception as e:
                     logger.warning(f"Failed to process {emlx_file}: {e}")
                     continue
-
+            
             # Close the temporary file so MboxReader can read it
             temp_mbox.close()
-
+            
             try:
                 # Use the parent MboxReader's logic to parse the mbox file
                 return super().load_data(Path(temp_mbox_path), extra_info, fs)
@@ -182,5 +188,5 @@ class EmlxMboxReader(MboxReader):
                 # Clean up temporary file
                 try:
                     os.unlink(temp_mbox_path)
-                except OSError:
+                except:
-                    pass
+                    pass

examples/grep_search_example.py

@@ -1,35 +0,0 @@
-"""
-Grep Search Example
-
-Shows how to use grep-based text search instead of semantic search.
-Useful when you need exact text matches rather than meaning-based results.
-"""
-
-from leann import LeannSearcher
-
-# Load your index
-searcher = LeannSearcher("my-documents.leann")
-
-# Regular semantic search
-print("=== Semantic Search ===")
-results = searcher.search("machine learning algorithms", top_k=3)
-for result in results:
-    print(f"Score: {result.score:.3f}")
-    print(f"Text: {result.text[:80]}...")
-    print()
-
-# Grep-based search for exact text matches
-print("=== Grep Search ===")
-results = searcher.search("def train_model", top_k=3, use_grep=True)
-for result in results:
-    print(f"Score: {result.score}")
-    print(f"Text: {result.text[:80]}...")
-    print()
-
-# Find specific error messages
-error_results = searcher.search("FileNotFoundError", use_grep=True)
-print(f"Found {len(error_results)} files mentioning FileNotFoundError")
-
-# Search for function definitions
-func_results = searcher.search("class SearchResult", use_grep=True, top_k=5)
-print(f"Found {len(func_results)} class definitions")

examples/mail_reader_leann.py

@@ -0,0 +1,229 @@
+import os
+import asyncio
+import dotenv
+from pathlib import Path
+from typing import List, Any
+from leann.api import LeannBuilder, LeannSearcher, LeannChat
+from llama_index.core.node_parser import SentenceSplitter
+
+dotenv.load_dotenv()
+
+def create_leann_index_from_multiple_sources(messages_dirs: List[Path], index_path: str = "mail_index.leann", max_count: int = -1):
+    """
+    Create LEANN index from multiple mail data sources.
+    
+    Args:
+        messages_dirs: List of Path objects pointing to Messages directories
+        index_path: Path to save the LEANN index
+        max_count: Maximum number of emails to process per directory
+    """
+    print("Creating LEANN index from multiple mail data sources...")
+    
+    # Load documents using EmlxReader from LEANN_email_reader
+    from LEANN_email_reader import EmlxReader
+    reader = EmlxReader()
+    # from email_data.email import EmlxMboxReader
+    # from pathlib import Path
+    # reader = EmlxMboxReader()
+    
+    all_documents = []
+    total_processed = 0
+    
+    # Process each Messages directory
+    for i, messages_dir in enumerate(messages_dirs):
+        print(f"\nProcessing Messages directory {i+1}/{len(messages_dirs)}: {messages_dir}")
+        
+        try:
+            documents = reader.load_data(messages_dir)
+            if documents:
+                print(f"Loaded {len(documents)} email documents from {messages_dir}")
+                all_documents.extend(documents)
+                total_processed += len(documents)
+                
+                # Check if we've reached the max count
+                if max_count > 0 and total_processed >= max_count:
+                    print(f"Reached max count of {max_count} documents")
+                    break
+            else:
+                print(f"No documents loaded from {messages_dir}")
+        except Exception as e:
+            print(f"Error processing {messages_dir}: {e}")
+            continue
+    
+    if not all_documents:
+        print("No documents loaded from any source. Exiting.")
+        return None
+    
+    print(f"\nTotal loaded {len(all_documents)} email documents from {len(messages_dirs)} directories")
+    
+    # Create text splitter with 256 chunk size
+    text_splitter = SentenceSplitter(chunk_size=256, chunk_overlap=25)
+    
+    # Convert Documents to text strings and chunk them
+    all_texts = []
+    for doc in all_documents:
+        # Split the document into chunks
+        nodes = text_splitter.get_nodes_from_documents([doc])
+        for node in nodes:
+            all_texts.append(node.get_content())
+    
+    print(f"Created {len(all_texts)} text chunks from {len(all_documents)} documents")
+    
+    # Create LEANN index directory
+    INDEX_DIR = Path(index_path).parent
+    
+    if not INDEX_DIR.exists():
+        print(f"--- Index directory not found, building new index ---")
+        INDEX_DIR.mkdir(exist_ok=True)
+
+        print(f"--- Building new LEANN index ---")
+        
+        print(f"\n[PHASE 1] Building Leann index...")
+
+        # Use HNSW backend for better macOS compatibility
+        builder = LeannBuilder(
+            backend_name="hnsw",
+            embedding_model="facebook/contriever",
+            graph_degree=32, 
+            complexity=64,
+            is_compact=True,
+            is_recompute=True,
+            num_threads=1  # Force single-threaded mode
+        )
+
+        print(f"Adding {len(all_texts)} email chunks to index...")
+        for chunk_text in all_texts:
+            builder.add_text(chunk_text)
+            
+        builder.build_index(index_path)
+        print(f"\nLEANN index built at {index_path}!")
+    else:
+        print(f"--- Using existing index at {INDEX_DIR} ---")
+    
+    return index_path
+
+def create_leann_index(mail_path: str, index_path: str = "mail_index.leann", max_count: int = 1000):
+    """
+    Create LEANN index from mail data.
+    
+    Args:
+        mail_path: Path to the mail directory
+        index_path: Path to save the LEANN index
+        max_count: Maximum number of emails to process
+    """
+    print("Creating LEANN index from mail data...")
+    
+    # Load documents using EmlxReader from LEANN_email_reader
+    from LEANN_email_reader import EmlxReader
+    reader = EmlxReader()
+    # from email_data.email import EmlxMboxReader
+    # from pathlib import Path
+    # reader = EmlxMboxReader()
+    documents = reader.load_data(Path(mail_path))
+    
+    if not documents:
+        print("No documents loaded. Exiting.")
+        return None
+    
+    print(f"Loaded {len(documents)} email documents")
+    
+    # Create text splitter with 256 chunk size
+    text_splitter = SentenceSplitter(chunk_size=256, chunk_overlap=25)
+    
+    # Convert Documents to text strings and chunk them
+    all_texts = []
+    for doc in documents:
+        # Split the document into chunks
+        nodes = text_splitter.get_nodes_from_documents([doc])
+        for node in nodes:
+            all_texts.append(node.get_content())
+    
+    print(f"Created {len(all_texts)} text chunks from {len(documents)} documents")
+    
+    # Create LEANN index directory
+    INDEX_DIR = Path(index_path).parent
+    
+    if not INDEX_DIR.exists():
+        print(f"--- Index directory not found, building new index ---")
+        INDEX_DIR.mkdir(exist_ok=True)
+
+        print(f"--- Building new LEANN index ---")
+        
+        print(f"\n[PHASE 1] Building Leann index...")
+
+        # Use HNSW backend for better macOS compatibility
+        builder = LeannBuilder(
+            backend_name="hnsw",
+            embedding_model="facebook/contriever",
+            graph_degree=32, 
+            complexity=64,
+            is_compact=True,
+            is_recompute=True,
+            num_threads=1  # Force single-threaded mode
+        )
+
+        print(f"Adding {len(all_texts)} email chunks to index...")
+        for chunk_text in all_texts:
+            builder.add_text(chunk_text)
+            
+        builder.build_index(index_path)
+        print(f"\nLEANN index built at {index_path}!")
+    else:
+        print(f"--- Using existing index at {INDEX_DIR} ---")
+    
+    return index_path
+
+async def query_leann_index(index_path: str, query: str):
+    """
+    Query the LEANN index.
+    
+    Args:
+        index_path: Path to the LEANN index
+        query: The query string
+    """
+    print(f"\n[PHASE 2] Starting Leann chat session...")
+    chat = LeannChat(index_path=index_path)
+    
+    print(f"You: {query}")
+    chat_response = chat.ask(
+        query, 
+        top_k=5, 
+        recompute_beighbor_embeddings=True,
+        complexity=32,
+        beam_width=1
+    )
+    print(f"Leann: {chat_response}")
+
+async def main():
+    # Base path to the mail data directory
+    base_mail_path = "/Users/yichuan/Library/Mail/V10/0FCA0879-FD8C-4B7E-83BF-FDDA930791C5/[Gmail].mbox/All Mail.mbox/78BA5BE1-8819-4F9A-9613-EB63772F1DD0/Data"
+    
+    INDEX_DIR = Path("./mail_index_leann_raw_text_all")
+    INDEX_PATH = str(INDEX_DIR / "mail_documents.leann")
+    
+    # Find all Messages directories
+    from LEANN_email_reader import EmlxReader
+    messages_dirs = EmlxReader.find_all_messages_directories(base_mail_path)
+    
+    if not messages_dirs:
+        print("No Messages directories found. Exiting.")
+        return
+    
+    # Create or load the LEANN index from all sources
+    index_path = create_leann_index_from_multiple_sources(messages_dirs, INDEX_PATH)
+    
+    if index_path:
+        # Example queries
+        queries = [
+            "Hows Berkeley Graduate Student Instructor",
+            "how's the icloud related advertisement saying",
+            "Whats the number of class recommend to take per semester for incoming EECS students"
+
+        ]
+        
+        for query in queries:
+            print("\n" + "="*60)
+            await query_leann_index(index_path, query)
+
+if __name__ == "__main__":
+    asyncio.run(main()) 

examples/mail_reader_llamaindex.py

@@ -0,0 +1,86 @@
+import os
+from pathlib import Path
+from typing import List, Any
+from llama_index.core import VectorStoreIndex, StorageContext
+from llama_index.core.node_parser import SentenceSplitter
+
+# --- EMBEDDING MODEL ---
+from llama_index.embeddings.huggingface import HuggingFaceEmbedding
+import torch
+
+# --- END EMBEDDING MODEL ---
+
+# Import EmlxReader from the new module
+from LEANN_email_reader import EmlxReader
+
+def create_and_save_index(mail_path: str, save_dir: str = "mail_index_embedded", max_count: int = 1000):
+    print("Creating index from mail data with embedded metadata...")
+    documents = EmlxReader().load_data(mail_path, max_count=max_count)
+    if not documents:
+        print("No documents loaded. Exiting.")
+        return None
+    text_splitter = SentenceSplitter(chunk_size=256, chunk_overlap=25)
+    # Use facebook/contriever as the embedder
+    embed_model = HuggingFaceEmbedding(model_name="facebook/contriever")
+    # set on device
+    import torch
+    if torch.cuda.is_available():
+        embed_model._model.to("cuda")
+    # set mps
+    elif torch.backends.mps.is_available():
+        embed_model._model.to("mps")
+    else:
+        embed_model._model.to("cpu")
+    index = VectorStoreIndex.from_documents(
+        documents,
+        transformations=[text_splitter],
+        embed_model=embed_model
+    )
+    os.makedirs(save_dir, exist_ok=True)
+    index.storage_context.persist(persist_dir=save_dir)
+    print(f"Index saved to {save_dir}")
+    return index
+
+def load_index(save_dir: str = "mail_index_embedded"):
+    try:
+        storage_context = StorageContext.from_defaults(persist_dir=save_dir)
+        index = VectorStoreIndex.from_vector_store(
+            storage_context.vector_store,
+            storage_context=storage_context
+        )
+        print(f"Index loaded from {save_dir}")
+        return index
+    except Exception as e:
+        print(f"Error loading index: {e}")
+        return None
+
+def query_index(index, query: str):
+    if index is None:
+        print("No index available for querying.")
+        return
+    query_engine = index.as_query_engine()
+    response = query_engine.query(query)
+    print(f"Query: {query}")
+    print(f"Response: {response}")
+
+def main():
+    mail_path = "/Users/yichuan/Library/Mail/V10/0FCA0879-FD8C-4B7E-83BF-FDDA930791C5/[Gmail].mbox/All Mail.mbox/78BA5BE1-8819-4F9A-9613-EB63772F1DD0/Data/9/Messages"
+    save_dir = "mail_index_embedded"
+    if os.path.exists(save_dir) and os.path.exists(os.path.join(save_dir, "vector_store.json")):
+        print("Loading existing index...")
+        index = load_index(save_dir)
+    else:
+        print("Creating new index...")
+        index = create_and_save_index(mail_path, save_dir, max_count=10000)
+    if index:
+        queries = [
+            "Hows Berkeley Graduate Student Instructor",
+            "how's the icloud related advertisement saying"
+            "Whats the number of class recommend to take per semester for incoming EECS students"
+        ]
+        for query in queries:
+            print("\n" + "="*50)
+            query_index(index, query)
+
+if __name__ == "__main__":
+    main() 

examples/main_cli_example.py

@@ -0,0 +1,99 @@
+import faulthandler
+faulthandler.enable()
+
+import argparse
+from llama_index.core import SimpleDirectoryReader, Settings
+from llama_index.core.readers.base import BaseReader
+from llama_index.node_parser.docling import DoclingNodeParser
+from llama_index.readers.docling import DoclingReader
+from docling_core.transforms.chunker.hybrid_chunker import HybridChunker
+import asyncio
+import dotenv
+from leann.api import LeannBuilder, LeannSearcher, LeannChat
+import shutil
+from pathlib import Path
+
+dotenv.load_dotenv()
+
+reader = DoclingReader(export_type=DoclingReader.ExportType.JSON)
+file_extractor: dict[str, BaseReader] = {
+    ".docx": reader,
+    ".pptx": reader, 
+    ".pdf": reader,
+    ".xlsx": reader,
+    ".txt": reader,
+    ".md": reader,
+}
+node_parser = DoclingNodeParser(
+    chunker=HybridChunker(tokenizer="facebook/contriever", max_tokens=128)
+)
+print("Loading documents...")
+documents = SimpleDirectoryReader(
+    "examples/data", 
+    recursive=True, 
+    file_extractor=file_extractor,
+    encoding="utf-8",
+    required_exts=[".pdf", ".docx", ".pptx", ".xlsx", ".txt", ".md"]
+).load_data(show_progress=True)
+print("Documents loaded.")
+all_texts = []
+for doc in documents:
+    nodes = node_parser.get_nodes_from_documents([doc])
+    for node in nodes:
+        all_texts.append(node.get_content())
+
+INDEX_DIR = Path("./test_pdf_index_pangu_test")
+INDEX_PATH = str(INDEX_DIR / "pdf_documents.leann")
+
+if not INDEX_DIR.exists():
+    print(f"--- Index directory not found, building new index ---")
+    
+    print(f"\n[PHASE 1] Building Leann index...")
+
+    # Use HNSW backend for better macOS compatibility
+    builder = LeannBuilder(
+        backend_name="hnsw",
+        embedding_model="facebook/contriever",
+        graph_degree=32, 
+        complexity=64,
+        is_compact=True,
+        is_recompute=True,
+        num_threads=1  # Force single-threaded mode
+    )
+
+    print(f"Loaded {len(all_texts)} text chunks from documents.")
+    for chunk_text in all_texts:
+        builder.add_text(chunk_text)
+        
+    builder.build_index(INDEX_PATH)
+    print(f"\nLeann index built at {INDEX_PATH}!")
+else:
+    print(f"--- Using existing index at {INDEX_DIR} ---")
+
+async def main(args):
+    print(f"\n[PHASE 2] Starting Leann chat session...")
+    
+    llm_config = {
+        "type": args.llm,
+        "model": args.model,
+        "host": args.host
+    }
+
+    chat = LeannChat(index_path=INDEX_PATH, llm_config=llm_config)
+    
+    query = "Based on the paper, what are the main techniques LEANN explores to reduce the storage overhead and DLPM explore to achieve Fairness and Efiiciency trade-off?"
+    query = "What is the main idea of RL and give me 5 exapmle of classic RL algorithms?"
+    query = "什么是盘古大模型以及盘古开发过程中遇到了什么阴暗面，任务令一般在什么城市颁发"
+
+    print(f"You: {query}")
+    chat_response = chat.ask(query, top_k=20, recompute_beighbor_embeddings=True, complexity=32)
+    print(f"Leann: {chat_response}")
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description="Run Leann Chat with various LLM backends.")
+    parser.add_argument("--llm", type=str, default="hf", choices=["simulated", "ollama", "hf", "openai"], help="The LLM backend to use.")
+    parser.add_argument("--model", type=str, default='meta-llama/Llama-3.2-3B-Instruct', help="The model name to use (e.g., 'llama3:8b' for ollama, 'deepseek-ai/deepseek-llm-7b-chat' for hf, 'gpt-4o' for openai).")
+    parser.add_argument("--host", type=str, default="http://localhost:11434", help="The host for the Ollama API.")
+    args = parser.parse_args()
+
+    asyncio.run(main(args))

examples/mlx_demo.py

@@ -1,43 +0,0 @@
-import os
-
-from leann.api import LeannBuilder, LeannChat
-
-# Define the path for our new MLX-based index
-INDEX_PATH = "./mlx_diskann_index/leann"
-
-if os.path.exists(INDEX_PATH + ".meta.json"):
-    print(f"Index already exists at {INDEX_PATH}. Skipping build.")
-else:
-    print("Initializing LeannBuilder with MLX support...")
-    # 1. Configure LeannBuilder to use MLX
-    builder = LeannBuilder(
-        backend_name="hnsw",
-        embedding_model="mlx-community/Qwen3-Embedding-0.6B-4bit-DWQ",
-        embedding_mode="mlx",
-    )
-
-    # 2. Add documents
-    print("Adding documents...")
-    docs = [
-        "MLX is an array framework for machine learning on Apple silicon.",
-        "It was designed by Apple's machine learning research team.",
-        "The mlx-community organization provides pre-trained models in MLX format.",
-        "It supports operations on multi-dimensional arrays.",
-        "Leann can now use MLX for its embedding models.",
-    ]
-    for doc in docs:
-        builder.add_text(doc)
-
-    # 3. Build the index
-    print(f"Building the MLX-based index at: {INDEX_PATH}")
-    builder.build_index(INDEX_PATH)
-    print("\nSuccessfully built the index with MLX embeddings!")
-    print(f"Check the metadata file: {INDEX_PATH}.meta.json")
-
-
-chat = LeannChat(index_path=INDEX_PATH)
-# add query
-query = "MLX is an array framework for machine learning on Apple silicon."
-print(f"Query: {query}")
-response = chat.ask(query, top_k=3, recompute_beighbor_embeddings=True, complexity=3, beam_width=1)
-print(f"Response: {response}")

examples/multi_vector_aggregator.py

@@ -0,0 +1,319 @@
+#!/usr/bin/env python3
+"""
+Multi-Vector Aggregator for Fat Embeddings
+==========================================
+
+This module implements aggregation strategies for multi-vector embeddings,
+similar to ColPali's approach where multiple patch vectors represent a single document.
+
+Key features:
+- MaxSim aggregation (take maximum similarity across patches)
+- Voting-based aggregation (count patch matches)
+- Weighted aggregation (attention-score weighted)
+- Spatial clustering of matching patches
+- Document-level result consolidation
+"""
+
+import numpy as np
+from typing import List, Dict, Any, Tuple, Optional
+from dataclasses import dataclass
+from collections import defaultdict
+import json
+
+@dataclass
+class PatchResult:
+    """Represents a single patch search result."""
+    patch_id: int
+    image_name: str
+    image_path: str
+    coordinates: Tuple[int, int, int, int]  # (x1, y1, x2, y2)
+    score: float
+    attention_score: float
+    scale: float
+    metadata: Dict[str, Any]
+
+@dataclass
+class AggregatedResult:
+    """Represents an aggregated document-level result."""
+    image_name: str
+    image_path: str
+    doc_score: float
+    patch_count: int
+    best_patch: PatchResult
+    all_patches: List[PatchResult]
+    aggregation_method: str
+    spatial_clusters: Optional[List[List[PatchResult]]] = None
+
+class MultiVectorAggregator:
+    """
+    Aggregates multiple patch-level results into document-level results.
+    """
+    
+    def __init__(self, 
+                 aggregation_method: str = "maxsim",
+                 spatial_clustering: bool = True,
+                 cluster_distance_threshold: float = 100.0):
+        """
+        Initialize the aggregator.
+        
+        Args:
+            aggregation_method: "maxsim", "voting", "weighted", or "mean"
+            spatial_clustering: Whether to cluster spatially close patches
+            cluster_distance_threshold: Distance threshold for spatial clustering
+        """
+        self.aggregation_method = aggregation_method
+        self.spatial_clustering = spatial_clustering
+        self.cluster_distance_threshold = cluster_distance_threshold
+    
+    def aggregate_results(self, 
+                         search_results: List[Dict[str, Any]], 
+                         top_k: int = 10) -> List[AggregatedResult]:
+        """
+        Aggregate patch-level search results into document-level results.
+        
+        Args:
+            search_results: List of search results from LeannSearcher
+            top_k: Number of top documents to return
+            
+        Returns:
+            List of aggregated document results
+        """
+        # Group results by image
+        image_groups = defaultdict(list)
+        
+        for result in search_results:
+            metadata = result.metadata
+            if "image_name" in metadata and "patch_id" in metadata:
+                patch_result = PatchResult(
+                    patch_id=metadata["patch_id"],
+                    image_name=metadata["image_name"],
+                    image_path=metadata["image_path"],
+                    coordinates=tuple(metadata["coordinates"]),
+                    score=result.score,
+                    attention_score=metadata.get("attention_score", 0.0),
+                    scale=metadata.get("scale", 1.0),
+                    metadata=metadata
+                )
+                image_groups[metadata["image_name"]].append(patch_result)
+        
+        # Aggregate each image group
+        aggregated_results = []
+        for image_name, patches in image_groups.items():
+            if len(patches) == 0:
+                continue
+                
+            agg_result = self._aggregate_image_patches(image_name, patches)
+            aggregated_results.append(agg_result)
+        
+        # Sort by aggregated score and return top-k
+        aggregated_results.sort(key=lambda x: x.doc_score, reverse=True)
+        return aggregated_results[:top_k]
+    
+    def _aggregate_image_patches(self, image_name: str, patches: List[PatchResult]) -> AggregatedResult:
+        """Aggregate patches for a single image."""
+        
+        if self.aggregation_method == "maxsim":
+            doc_score = max(patch.score for patch in patches)
+            best_patch = max(patches, key=lambda p: p.score)
+            
+        elif self.aggregation_method == "voting":
+            # Count patches above threshold
+            threshold = np.percentile([p.score for p in patches], 75)
+            doc_score = sum(1 for patch in patches if patch.score >= threshold)
+            best_patch = max(patches, key=lambda p: p.score)
+            
+        elif self.aggregation_method == "weighted":
+            # Weight by attention scores
+            total_weighted_score = sum(p.score * p.attention_score for p in patches)
+            total_weights = sum(p.attention_score for p in patches)
+            doc_score = total_weighted_score / max(total_weights, 1e-8)
+            best_patch = max(patches, key=lambda p: p.score * p.attention_score)
+            
+        elif self.aggregation_method == "mean":
+            doc_score = np.mean([patch.score for patch in patches])
+            best_patch = max(patches, key=lambda p: p.score)
+            
+        else:
+            raise ValueError(f"Unknown aggregation method: {self.aggregation_method}")
+        
+        # Spatial clustering if enabled
+        spatial_clusters = None
+        if self.spatial_clustering:
+            spatial_clusters = self._cluster_patches_spatially(patches)
+        
+        return AggregatedResult(
+            image_name=image_name,
+            image_path=patches[0].image_path,
+            doc_score=float(doc_score),
+            patch_count=len(patches),
+            best_patch=best_patch,
+            all_patches=sorted(patches, key=lambda p: p.score, reverse=True),
+            aggregation_method=self.aggregation_method,
+            spatial_clusters=spatial_clusters
+        )
+    
+    def _cluster_patches_spatially(self, patches: List[PatchResult]) -> List[List[PatchResult]]:
+        """Cluster patches that are spatially close to each other."""
+        if len(patches) <= 1:
+            return [patches]
+        
+        clusters = []
+        remaining_patches = patches.copy()
+        
+        while remaining_patches:
+            # Start new cluster with highest scoring remaining patch
+            seed_patch = max(remaining_patches, key=lambda p: p.score)
+            current_cluster = [seed_patch]
+            remaining_patches.remove(seed_patch)
+            
+            # Add nearby patches to cluster
+            added_to_cluster = True
+            while added_to_cluster:
+                added_to_cluster = False
+                for patch in remaining_patches.copy():
+                    if self._is_patch_nearby(patch, current_cluster):
+                        current_cluster.append(patch)
+                        remaining_patches.remove(patch)
+                        added_to_cluster = True
+            
+            clusters.append(current_cluster)
+        
+        return sorted(clusters, key=lambda cluster: max(p.score for p in cluster), reverse=True)
+    
+    def _is_patch_nearby(self, patch: PatchResult, cluster: List[PatchResult]) -> bool:
+        """Check if a patch is spatially close to any patch in the cluster."""
+        patch_center = self._get_patch_center(patch.coordinates)
+        
+        for cluster_patch in cluster:
+            cluster_center = self._get_patch_center(cluster_patch.coordinates)
+            distance = np.sqrt((patch_center[0] - cluster_center[0])**2 + 
+                             (patch_center[1] - cluster_center[1])**2)
+            
+            if distance <= self.cluster_distance_threshold:
+                return True
+        
+        return False
+    
+    def _get_patch_center(self, coordinates: Tuple[int, int, int, int]) -> Tuple[float, float]:
+        """Get center point of a patch."""
+        x1, y1, x2, y2 = coordinates
+        return ((x1 + x2) / 2, (y1 + y2) / 2)
+    
+    def print_aggregated_results(self, results: List[AggregatedResult], max_patches_per_doc: int = 3):
+        """Pretty print aggregated results."""
+        print(f"\n🔍 Aggregated Results (method: {self.aggregation_method})")
+        print("=" * 80)
+        
+        for i, result in enumerate(results):
+            print(f"\n{i+1}. {result.image_name}")
+            print(f"   Doc Score: {result.doc_score:.4f} | Patches: {result.patch_count}")
+            print(f"   Path: {result.image_path}")
+            
+            # Show best patch
+            best = result.best_patch
+            print(f"   🌟 Best Patch: #{best.patch_id} at {best.coordinates} (score: {best.score:.4f})")
+            
+            # Show top patches
+            print(f"   📍 Top Patches:")
+            for j, patch in enumerate(result.all_patches[:max_patches_per_doc]):
+                print(f"      {j+1}. Patch #{patch.patch_id}: {patch.score:.4f} at {patch.coordinates}")
+            
+            # Show spatial clusters if available
+            if result.spatial_clusters and len(result.spatial_clusters) > 1:
+                print(f"   🗂️ Spatial Clusters: {len(result.spatial_clusters)}")
+                for j, cluster in enumerate(result.spatial_clusters[:2]):  # Show top 2 clusters
+                    cluster_score = max(p.score for p in cluster)
+                    print(f"      Cluster {j+1}: {len(cluster)} patches (best: {cluster_score:.4f})")
+
+def demo_aggregation():
+    """Demonstrate the multi-vector aggregation functionality."""
+    print("=== Multi-Vector Aggregation Demo ===")
+    
+    # Simulate some patch-level search results
+    # In real usage, these would come from LeannSearcher.search()
+    
+    class MockResult:
+        def __init__(self, score, metadata):
+            self.score = score
+            self.metadata = metadata
+    
+    # Simulate results for 2 images with multiple patches each
+    mock_results = [
+        # Image 1: cats_and_kitchen.jpg - 4 patches
+        MockResult(0.85, {
+            "image_name": "cats_and_kitchen.jpg",
+            "image_path": "/path/to/cats_and_kitchen.jpg",
+            "patch_id": 3,
+            "coordinates": [100, 50, 224, 174],  # Kitchen area
+            "attention_score": 0.92,
+            "scale": 1.0
+        }),
+        MockResult(0.78, {
+            "image_name": "cats_and_kitchen.jpg", 
+            "image_path": "/path/to/cats_and_kitchen.jpg",
+            "patch_id": 7,
+            "coordinates": [200, 300, 324, 424],  # Cat area
+            "attention_score": 0.88,
+            "scale": 1.0
+        }),
+        MockResult(0.72, {
+            "image_name": "cats_and_kitchen.jpg",
+            "image_path": "/path/to/cats_and_kitchen.jpg", 
+            "patch_id": 12,
+            "coordinates": [150, 100, 274, 224],  # Appliances
+            "attention_score": 0.75,
+            "scale": 1.0
+        }),
+        MockResult(0.65, {
+            "image_name": "cats_and_kitchen.jpg",
+            "image_path": "/path/to/cats_and_kitchen.jpg",
+            "patch_id": 15,
+            "coordinates": [50, 250, 174, 374],  # Furniture
+            "attention_score": 0.70,
+            "scale": 1.0
+        }),
+        
+        # Image 2: city_street.jpg - 3 patches  
+        MockResult(0.68, {
+            "image_name": "city_street.jpg",
+            "image_path": "/path/to/city_street.jpg",
+            "patch_id": 2,
+            "coordinates": [300, 100, 424, 224],  # Buildings
+            "attention_score": 0.80,
+            "scale": 1.0
+        }),
+        MockResult(0.62, {
+            "image_name": "city_street.jpg",
+            "image_path": "/path/to/city_street.jpg",
+            "patch_id": 8,
+            "coordinates": [100, 350, 224, 474],  # Street level
+            "attention_score": 0.75,
+            "scale": 1.0
+        }),
+        MockResult(0.55, {
+            "image_name": "city_street.jpg", 
+            "image_path": "/path/to/city_street.jpg",
+            "patch_id": 11,
+            "coordinates": [400, 200, 524, 324],  # Sky area
+            "attention_score": 0.60,
+            "scale": 1.0
+        }),
+    ]
+    
+    # Test different aggregation methods
+    methods = ["maxsim", "voting", "weighted", "mean"]
+    
+    for method in methods:
+        print(f"\n{'='*20} {method.upper()} AGGREGATION {'='*20}")
+        
+        aggregator = MultiVectorAggregator(
+            aggregation_method=method,
+            spatial_clustering=True,
+            cluster_distance_threshold=100.0
+        )
+        
+        aggregated = aggregator.aggregate_results(mock_results, top_k=5)
+        aggregator.print_aggregated_results(aggregated)
+
+if __name__ == "__main__":
+    demo_aggregation()

examples/resue_index.py

@@ -0,0 +1,18 @@
+import asyncio
+from leann.api import LeannChat
+from pathlib import Path
+
+INDEX_DIR = Path("./test_pdf_index_huawei")
+INDEX_PATH = str(INDEX_DIR / "pdf_documents.leann")
+
+async def main():
+    print(f"\n[PHASE 2] Starting Leann chat session...")
+    chat = LeannChat(index_path=INDEX_PATH)
+    query = "What is the main idea of RL and give me 5 exapmle of classic RL algorithms?"
+    query = "Based on the paper, what are the main techniques LEANN explores to reduce the storage overhead and DLPM explore to achieve Fairness and Efiiciency trade-off?"
+    # query = "什么是盘古大模型以及盘古开发过程中遇到了什么阴暗面，任务令一般在什么城市颁发"
+    response = chat.ask(query,top_k=20,recompute_beighbor_embeddings=True,complexity=32,beam_width=1)
+    print(f"\n[PHASE 2] Response: {response}")
+
+if __name__ == "__main__":
+    asyncio.run(main())

examples/run_evaluation.py

@@ -0,0 +1,157 @@
+#!/usr/bin/env python3
+"""
+This script runs a recall evaluation on a given LEANN index.
+It correctly compares results by fetching the text content for both the new search
+results and the golden standard results, making the comparison robust to ID changes.
+"""
+
+import json
+import argparse
+import time
+from pathlib import Path
+import sys
+import numpy as np
+from typing import List, Dict, Any
+import glob
+import pickle
+
+# Add project root to path to allow importing from leann
+project_root = Path(__file__).resolve().parent.parent
+sys.path.insert(0, str(project_root))
+
+from leann.api import LeannSearcher
+
+# --- Configuration ---
+NQ_QUERIES_FILE = Path("/opt/dlami/nvme/scaling_out/examples/nq_open.jsonl")
+
+# Ground truth files for different datasets
+GROUND_TRUTH_FILES = {
+    "rpj_wiki": "/opt/dlami/nvme/scaling_out/indices/rpj_wiki/facebook/contriever-msmarco/flat_results_nq_k3.json",
+    "dpr": "/opt/dlami/nvme/scaling_out/indices/dpr/facebook/contriever-msmarco/flat_results_nq_k3.json"
+}
+
+# Old passages for different datasets  
+OLD_PASSAGES_GLOBS = {
+    "rpj_wiki": "/opt/dlami/nvme/scaling_out/passages/rpj_wiki/8-shards/raw_passages-*-of-8.pkl.jsonl",
+    "dpr": "/opt/dlami/nvme/scaling_out/passages/dpr/1-shards/raw_passages-*-of-1.pkl.jsonl"
+}
+
+# --- Helper Class to Load Original Passages ---
+class OldPassageLoader:
+    """A simplified version of the old LazyPassages class to fetch golden results by ID."""
+    def __init__(self, passages_glob: str):
+        self.jsonl_paths = sorted(glob.glob(passages_glob))
+        self.offsets = {}
+        self.fps = [open(p, "r", encoding="utf-8") for p in self.jsonl_paths]
+        print("Building offset map for original passages...")
+        for i, shard_path_str in enumerate(self.jsonl_paths):
+            old_idx_path = Path(shard_path_str.replace(".jsonl", ".idx"))
+            if not old_idx_path.exists(): continue
+            with open(old_idx_path, 'rb') as f:
+                shard_offsets = pickle.load(f)
+                for pid, offset in shard_offsets.items():
+                    self.offsets[str(pid)] = (i, offset)
+        print("Offset map for original passages is ready.")
+
+    def get_passage_by_id(self, pid: str) -> Dict[str, Any]:
+        pid = str(pid)
+        if pid not in self.offsets:
+            raise ValueError(f"Passage ID {pid} not found in offsets")
+        file_idx, offset = self.offsets[pid]
+        fp = self.fps[file_idx]
+        fp.seek(offset)
+        return json.loads(fp.readline())
+
+    def __del__(self):
+        for fp in self.fps:
+            fp.close()
+
+def load_queries(file_path: Path) -> List[str]:
+    queries = []
+    with open(file_path, 'r', encoding='utf-8') as f:
+        for line in f:
+            data = json.loads(line)
+            queries.append(data['query'])
+    return queries
+
+def main():
+    parser = argparse.ArgumentParser(description="Run recall evaluation on a LEANN index.")
+    parser.add_argument("index_path", type=str, help="Path to the LEANN index to evaluate.")
+    parser.add_argument("--num-queries", type=int, default=10, help="Number of queries to evaluate.")
+    parser.add_argument("--top-k", type=int, default=3, help="The 'k' value for recall@k.")
+    parser.add_argument("--ef-search", type=int, default=120, help="The 'efSearch' parameter for HNSW.")
+    args = parser.parse_args()
+
+    print(f"--- Final, Correct Recall Evaluation (efSearch={args.ef_search}) ---")
+    
+    # Detect dataset type from index path
+    index_path_str = str(args.index_path)
+    if "rpj_wiki" in index_path_str:
+        dataset_type = "rpj_wiki"
+    elif "dpr" in index_path_str:
+        dataset_type = "dpr"
+    else:
+        print("WARNING: Unknown dataset type, defaulting to rpj_wiki")
+        dataset_type = "rpj_wiki"
+    
+    print(f"INFO: Detected dataset type: {dataset_type}")
+
+    try:
+        searcher = LeannSearcher(args.index_path)
+        queries = load_queries(NQ_QUERIES_FILE)
+        
+        golden_results_file = GROUND_TRUTH_FILES[dataset_type]
+        old_passages_glob = OLD_PASSAGES_GLOBS[dataset_type]
+        
+        print(f"INFO: Using ground truth file: {golden_results_file}")
+        print(f"INFO: Using old passages glob: {old_passages_glob}")
+        
+        with open(golden_results_file, 'r') as f:
+            golden_results_data = json.load(f)
+        
+        old_passage_loader = OldPassageLoader(old_passages_glob)
+
+        num_eval_queries = min(args.num_queries, len(queries))
+        queries = queries[:num_eval_queries]
+        
+        print(f"\nRunning evaluation on {num_eval_queries} queries...")
+        recall_scores = []
+        search_times = []
+
+        for i in range(num_eval_queries):
+            start_time = time.time()
+            new_results = searcher.search(queries[i], top_k=args.top_k, ef=args.ef_search)
+            search_times.append(time.time() - start_time)
+
+            # Correct Recall Calculation: Based on TEXT content
+            new_texts = {result.text for result in new_results}
+            golden_ids = golden_results_data["indices"][i][:args.top_k]
+            golden_texts = {old_passage_loader.get_passage_by_id(str(gid))['text'] for gid in golden_ids}
+
+            overlap = len(new_texts & golden_texts)
+            recall = overlap / len(golden_texts) if golden_texts else 0
+            recall_scores.append(recall)
+
+            print("\n--- EVALUATION RESULTS ---")
+            print(f"Query: {queries[i]}")
+            print(f"New Results: {new_texts}")
+            print(f"Golden Results: {golden_texts}")
+            print(f"Overlap: {overlap}")
+            print(f"Recall: {recall}")
+            print(f"Search Time: {search_times[-1]:.4f}s")
+            print(f"--------------------------------")
+
+        avg_recall = np.mean(recall_scores) if recall_scores else 0
+        avg_time = np.mean(search_times) if search_times else 0
+
+        print(f"\n🎉 --- Evaluation Complete ---")
+        print(f"Avg. Recall@{args.top_k} (efSearch={args.ef_search}): {avg_recall:.4f}")
+        print(f"Avg. Search Time: {avg_time:.4f}s")
+
+    except Exception as e:
+        print(f"\n❌ An error occurred during evaluation: {e}")
+        import traceback
+        traceback.print_exc()
+
+if __name__ == "__main__":
+    main()

examples/simple_demo.py

@@ -1,28 +1,21 @@
 """
 Simple demo showing basic leann usage
-Run: uv run python examples/basic_demo.py
+Run: uv run python examples/simple_demo.py
 """
 
 import argparse
-
+from leann import LeannBuilder, LeannSearcher, LeannChat
-from leann import LeannBuilder, LeannChat, LeannSearcher
 
 
 def main():
-    parser = argparse.ArgumentParser(
+    parser = argparse.ArgumentParser(description="Simple demo of Leann with selectable embedding models.")
-        description="Simple demo of Leann with selectable embedding models."
+    parser.add_argument("--embedding_model", type=str, default="sentence-transformers/all-mpnet-base-v2",
-    )
+                        help="The embedding model to use, e.g., 'sentence-transformers/all-mpnet-base-v2' or 'text-embedding-ada-002'.")
-    parser.add_argument(
-        "--embedding_model",
-        type=str,
-        default="sentence-transformers/all-mpnet-base-v2",
-        help="The embedding model to use, e.g., 'sentence-transformers/all-mpnet-base-v2' or 'text-embedding-ada-002'.",
-    )
     args = parser.parse_args()
 
     print(f"=== Leann Simple Demo with {args.embedding_model} ===")
     print()
-
+    
     # Sample knowledge base
     chunks = [
         "Machine learning is a subset of artificial intelligence that enables computers to learn without being explicitly programmed.",
@@ -34,7 +27,7 @@ def main():
         "Big data refers to extremely large datasets that require special tools and techniques to process.",
         "Cloud computing provides on-demand access to computing resources over the internet.",
     ]
-
+    
     print("1. Building index (no embeddings stored)...")
     builder = LeannBuilder(
         embedding_model=args.embedding_model,
@@ -44,45 +37,45 @@ def main():
         builder.add_text(chunk)
     builder.build_index("demo_knowledge.leann")
     print()
-
+    
     print("2. Searching with real-time embeddings...")
     searcher = LeannSearcher("demo_knowledge.leann")
-
+    
     queries = [
         "What is machine learning?",
-        "How does neural network work?",
+        "How does neural network work?", 
         "Tell me about data processing",
     ]
-
+    
     for query in queries:
         print(f"Query: {query}")
         results = searcher.search(query, top_k=2)
-
+        
         for i, result in enumerate(results, 1):
             print(f"  {i}. Score: {result.score:.3f}")
             print(f"     Text: {result.text[:100]}...")
         print()
-
+    
     print("3. Interactive chat demo:")
     print("   (Note: Requires OpenAI API key for real responses)")
-
+    
     chat = LeannChat("demo_knowledge.leann")
-
+    
     # Demo questions
     demo_questions: list[str] = [
         "What is the difference between machine learning and deep learning?",
         "How is data science related to big data?",
     ]
-
+    
     for question in demo_questions:
         print(f"   Q: {question}")
         response = chat.ask(question)
         print(f"   A: {response}")
         print()
-
+    
     print("Demo completed! Try running:")
-    print("   uv run python apps/document_rag.py")
+    print("   uv run python examples/document_search.py")
 
 
 if __name__ == "__main__":
-    main()
+    main()

examples/spoiler_free_book_rag.py

@@ -1,250 +0,0 @@
-#!/usr/bin/env python3
-"""
-Spoiler-Free Book RAG Example using LEANN Metadata Filtering
-
-This example demonstrates how to use LEANN's metadata filtering to create
-a spoiler-free book RAG system where users can search for information
-up to a specific chapter they've read.
-
-Usage:
-    python spoiler_free_book_rag.py
-"""
-
-import os
-import sys
-from typing import Any, Optional
-
-# Add LEANN to path (adjust path as needed)
-sys.path.insert(0, os.path.join(os.path.dirname(__file__), "../packages/leann-core/src"))
-
-from leann.api import LeannBuilder, LeannSearcher
-
-
-def chunk_book_with_metadata(book_title: str = "Sample Book") -> list[dict[str, Any]]:
-    """
-    Create sample book chunks with metadata for demonstration.
-
-    In a real implementation, this would parse actual book files (epub, txt, etc.)
-    and extract chapter boundaries, character mentions, etc.
-
-    Args:
-        book_title: Title of the book
-
-    Returns:
-        List of chunk dictionaries with text and metadata
-    """
-    # Sample book chunks with metadata
-    # In practice, you'd use proper text processing libraries
-
-    sample_chunks = [
-        {
-            "text": "Alice was beginning to get very tired of sitting by her sister on the bank, and of having nothing to do.",
-            "metadata": {
-                "book": book_title,
-                "chapter": 1,
-                "page": 1,
-                "characters": ["Alice", "Sister"],
-                "themes": ["boredom", "curiosity"],
-                "location": "riverbank",
-            },
-        },
-        {
-            "text": "So she was considering in her own mind (as well as she could, for the hot day made her feel very sleepy and stupid), whether the pleasure of making a daisy-chain would be worth the trouble of getting up and picking the daisies, when suddenly a White Rabbit with pink eyes ran close by her.",
-            "metadata": {
-                "book": book_title,
-                "chapter": 1,
-                "page": 2,
-                "characters": ["Alice", "White Rabbit"],
-                "themes": ["decision", "surprise", "magic"],
-                "location": "riverbank",
-            },
-        },
-        {
-            "text": "Alice found herself falling down a very deep well. Either the well was very deep, or she fell very slowly, for she had plenty of time as she fell to look about her and to wonder what was going to happen next.",
-            "metadata": {
-                "book": book_title,
-                "chapter": 2,
-                "page": 15,
-                "characters": ["Alice"],
-                "themes": ["falling", "wonder", "transformation"],
-                "location": "rabbit hole",
-            },
-        },
-        {
-            "text": "Alice meets the Cheshire Cat, who tells her that everyone in Wonderland is mad, including Alice herself.",
-            "metadata": {
-                "book": book_title,
-                "chapter": 6,
-                "page": 85,
-                "characters": ["Alice", "Cheshire Cat"],
-                "themes": ["madness", "philosophy", "identity"],
-                "location": "Duchess's house",
-            },
-        },
-        {
-            "text": "At the Queen's croquet ground, Alice witnesses the absurd trial that reveals the arbitrary nature of Wonderland's justice system.",
-            "metadata": {
-                "book": book_title,
-                "chapter": 8,
-                "page": 120,
-                "characters": ["Alice", "Queen of Hearts", "King of Hearts"],
-                "themes": ["justice", "absurdity", "authority"],
-                "location": "Queen's court",
-            },
-        },
-        {
-            "text": "Alice realizes that Wonderland was all a dream, even the Rabbit, as she wakes up on the riverbank next to her sister.",
-            "metadata": {
-                "book": book_title,
-                "chapter": 12,
-                "page": 180,
-                "characters": ["Alice", "Sister", "Rabbit"],
-                "themes": ["revelation", "reality", "growth"],
-                "location": "riverbank",
-            },
-        },
-    ]
-
-    return sample_chunks
-
-
-def build_spoiler_free_index(book_chunks: list[dict[str, Any]], index_name: str) -> str:
-    """
-    Build a LEANN index with book chunks that include spoiler metadata.
-
-    Args:
-        book_chunks: List of book chunks with metadata
-        index_name: Name for the index
-
-    Returns:
-        Path to the built index
-    """
-    print(f"📚 Building spoiler-free book index: {index_name}")
-
-    # Initialize LEANN builder
-    builder = LeannBuilder(
-        backend_name="hnsw", embedding_model="text-embedding-3-small", embedding_mode="openai"
-    )
-
-    # Add each chunk with its metadata
-    for chunk in book_chunks:
-        builder.add_text(text=chunk["text"], metadata=chunk["metadata"])
-
-    # Build the index
-    index_path = f"{index_name}_book_index"
-    builder.build_index(index_path)
-
-    print(f"✅ Index built successfully: {index_path}")
-    return index_path
-
-
-def spoiler_free_search(
-    index_path: str,
-    query: str,
-    max_chapter: int,
-    character_filter: Optional[list[str]] = None,
-) -> list[dict[str, Any]]:
-    """
-    Perform a spoiler-free search on the book index.
-
-    Args:
-        index_path: Path to the LEANN index
-        query: Search query
-        max_chapter: Maximum chapter number to include
-        character_filter: Optional list of characters to focus on
-
-    Returns:
-        List of search results safe for the reader
-    """
-    print(f"🔍 Searching: '{query}' (up to chapter {max_chapter})")
-
-    searcher = LeannSearcher(index_path)
-
-    metadata_filters = {"chapter": {"<=": max_chapter}}
-
-    if character_filter:
-        metadata_filters["characters"] = {"contains": character_filter[0]}
-
-    results = searcher.search(query=query, top_k=10, metadata_filters=metadata_filters)
-
-    return results
-
-
-def demo_spoiler_free_rag():
-    """
-    Demonstrate the spoiler-free book RAG system.
-    """
-    print("🎭 Spoiler-Free Book RAG Demo")
-    print("=" * 40)
-
-    # Step 1: Prepare book data
-    book_title = "Alice's Adventures in Wonderland"
-    book_chunks = chunk_book_with_metadata(book_title)
-
-    print(f"📖 Loaded {len(book_chunks)} chunks from '{book_title}'")
-
-    # Step 2: Build the index (in practice, this would be done once)
-    try:
-        index_path = build_spoiler_free_index(book_chunks, "alice_wonderland")
-    except Exception as e:
-        print(f"❌ Failed to build index (likely missing dependencies): {e}")
-        print(
-            "💡 This demo shows the filtering logic - actual indexing requires LEANN dependencies"
-        )
-        return
-
-    # Step 3: Demonstrate various spoiler-free searches
-    search_scenarios = [
-        {
-            "description": "Reader who has only read Chapter 1",
-            "query": "What can you tell me about the rabbit?",
-            "max_chapter": 1,
-        },
-        {
-            "description": "Reader who has read up to Chapter 5",
-            "query": "Tell me about Alice's adventures",
-            "max_chapter": 5,
-        },
-        {
-            "description": "Reader who has read most of the book",
-            "query": "What does the Cheshire Cat represent?",
-            "max_chapter": 10,
-        },
-        {
-            "description": "Reader who has read the whole book",
-            "query": "What can you tell me about the rabbit?",
-            "max_chapter": 12,
-        },
-    ]
-
-    for scenario in search_scenarios:
-        print(f"\n📚 Scenario: {scenario['description']}")
-        print(f"   Query: {scenario['query']}")
-
-        try:
-            results = spoiler_free_search(
-                index_path=index_path,
-                query=scenario["query"],
-                max_chapter=scenario["max_chapter"],
-            )
-
-            print(f"   📄 Found {len(results)} results:")
-            for i, result in enumerate(results[:3], 1):  # Show top 3
-                chapter = result.metadata.get("chapter", "?")
-                location = result.metadata.get("location", "?")
-                print(f"      {i}. Chapter {chapter} ({location}): {result.text[:80]}...")
-
-        except Exception as e:
-            print(f"   ❌ Search failed: {e}")
-
-
-if __name__ == "__main__":
-    print("📚 LEANN Spoiler-Free Book RAG Example")
-    print("=====================================")
-
-    try:
-        demo_spoiler_free_rag()
-    except ImportError as e:
-        print(f"❌ Cannot run demo due to missing dependencies: {e}")
-    except Exception as e:
-        print(f"❌ Error running demo: {e}")

llms.txt

@@ -1,28 +0,0 @@
-# llms.txt — LEANN MCP and Agent Integration
-product: LEANN
-homepage: https://github.com/yichuan-w/LEANN
-contact: https://github.com/yichuan-w/LEANN/issues
-
-# Installation
-install: uv tool install leann-core --with leann
-
-# MCP Server Entry Point
-mcp.server: leann_mcp
-mcp.protocol_version: 2024-11-05
-
-# Tools
-mcp.tools: leann_list, leann_search
-
-mcp.tool.leann_list.description: List available LEANN indexes
-mcp.tool.leann_list.input: {}
-
-mcp.tool.leann_search.description: Semantic search across a named LEANN index
-mcp.tool.leann_search.input.index_name: string, required
-mcp.tool.leann_search.input.query: string, required
-mcp.tool.leann_search.input.top_k: integer, optional, default=5, min=1, max=20
-mcp.tool.leann_search.input.complexity: integer, optional, default=32, min=16, max=128
-
-# Notes
-note: Build indexes with `leann build <name> --docs <files...>` before searching.
-example.add: claude mcp add --scope user leann-server -- leann_mcp
-example.verify: claude mcp list | cat

packages/__init__.py

@@ -0,0 +1 @@
+

packages/leann-backend-diskann/CMakeLists.txt

@@ -0,0 +1,8 @@
+# packages/leann-backend-diskann/CMakeLists.txt (最终简化版)
+
+cmake_minimum_required(VERSION 3.20)
+project(leann_backend_diskann_wrapper)
+
+# 告诉 CMake 直接进入 DiskANN 子模块并执行它自己的 CMakeLists.txt
+# DiskANN 会自己处理所有事情，包括编译 Python 绑定
+add_subdirectory(src/third_party/DiskANN)

packages/leann-backend-diskann/__init__.py

@@ -1 +1 @@
-# This file makes the directory a Python package
+# This file makes the directory a Python package 

packages/leann-backend-diskann/leann_backend_diskann/__init__.py

@@ -1,7 +1 @@
-from . import diskann_backend as diskann_backend
+from . import diskann_backend
-from . import graph_partition
-
-# Export main classes and functions
-from .graph_partition import GraphPartitioner, partition_graph
-
-__all__ = ["GraphPartitioner", "diskann_backend", "graph_partition", "partition_graph"]

packages/leann-backend-diskann/leann_backend_diskann/diskann_backend.py

@@ -1,77 +1,28 @@
-import contextlib
-import logging
-import os
-import struct
-import sys
-from pathlib import Path
-from typing import Any, Literal, Optional
-
 import numpy as np
-import psutil
+import os
-from leann.interface import (
+import json
-    LeannBackendBuilderInterface,
+import struct
-    LeannBackendFactoryInterface,
+from pathlib import Path
-    LeannBackendSearcherInterface,
+from typing import Dict, Any, List
-)
+import contextlib
-from leann.registry import register_backend
+import pickle
+
 from leann.searcher_base import BaseSearcher
-
+from leann.registry import register_backend
-logger = logging.getLogger(__name__)
+from leann.interface import (
-
+    LeannBackendFactoryInterface,
-
+    LeannBackendBuilderInterface,
-@contextlib.contextmanager
+    LeannBackendSearcherInterface
-def suppress_cpp_output_if_needed():
+)
-    """Suppress C++ stdout/stderr based on LEANN_LOG_LEVEL"""
-    # In CI we avoid fiddling with low-level file descriptors to prevent aborts
-    if os.getenv("CI") == "true":
-        yield
-        return
-
-    log_level = os.getenv("LEANN_LOG_LEVEL", "WARNING").upper()
-
-    # Only suppress if log level is WARNING or higher (ERROR, CRITICAL)
-    should_suppress = log_level in ["WARNING", "ERROR", "CRITICAL"]
-
-    if not should_suppress:
-        # Don't suppress, just yield
-        yield
-        return
-
-    # Save original file descriptors
-    stdout_fd = sys.stdout.fileno()
-    stderr_fd = sys.stderr.fileno()
-
-    # Save original stdout/stderr
-    stdout_dup = os.dup(stdout_fd)
-    stderr_dup = os.dup(stderr_fd)
-
-    try:
-        # Redirect to /dev/null
-        devnull = os.open(os.devnull, os.O_WRONLY)
-        os.dup2(devnull, stdout_fd)
-        os.dup2(devnull, stderr_fd)
-        os.close(devnull)
-
-        yield
-
-    finally:
-        # Restore original file descriptors
-        os.dup2(stdout_dup, stdout_fd)
-        os.dup2(stderr_dup, stderr_fd)
-        os.close(stdout_dup)
-        os.close(stderr_dup)
-
 
 def _get_diskann_metrics():
-    from . import _diskannpy as diskannpy  # type: ignore
+    from . import _diskannpy as diskannpy
-
     return {
         "mips": diskannpy.Metric.INNER_PRODUCT,
         "l2": diskannpy.Metric.L2,
         "cosine": diskannpy.Metric.COSINE,
     }
 
-
 @contextlib.contextmanager
 def chdir(path):
     original_dir = os.getcwd()
@@ -81,52 +32,13 @@ def chdir(path):
     finally:
         os.chdir(original_dir)
 
-
 def _write_vectors_to_bin(data: np.ndarray, file_path: Path):
     num_vectors, dim = data.shape
-    with open(file_path, "wb") as f:
+    with open(file_path, 'wb') as f:
-        f.write(struct.pack("I", num_vectors))
+        f.write(struct.pack('I', num_vectors))
-        f.write(struct.pack("I", dim))
+        f.write(struct.pack('I', dim))
         f.write(data.tobytes())
 
-
-def _calculate_smart_memory_config(data: np.ndarray) -> tuple[float, float]:
-    """
-    Calculate smart memory configuration for DiskANN based on data size and system specs.
-
-    Args:
-        data: The embedding data array
-
-    Returns:
-        tuple: (search_memory_maximum, build_memory_maximum) in GB
-    """
-    num_vectors, dim = data.shape
-
-    # Calculate embedding storage size
-    embedding_size_bytes = num_vectors * dim * 4  # float32 = 4 bytes
-    embedding_size_gb = embedding_size_bytes / (1024**3)
-
-    # search_memory_maximum: 1/10 of embedding size for optimal PQ compression
-    # This controls Product Quantization size - smaller means more compression
-    search_memory_gb = max(0.1, embedding_size_gb / 10)  # At least 100MB
-
-    # build_memory_maximum: Based on available system RAM for sharding control
-    # This controls how much memory DiskANN uses during index construction
-    available_memory_gb = psutil.virtual_memory().available / (1024**3)
-    total_memory_gb = psutil.virtual_memory().total / (1024**3)
-
-    # Use 50% of available memory, but at least 2GB and at most 75% of total
-    build_memory_gb = max(2.0, min(available_memory_gb * 0.5, total_memory_gb * 0.75))
-
-    logger.info(
-        f"Smart memory config - Data: {embedding_size_gb:.2f}GB, "
-        f"Search mem: {search_memory_gb:.2f}GB (PQ control), "
-        f"Build mem: {build_memory_gb:.2f}GB (sharding control)"
-    )
-
-    return search_memory_gb, build_memory_gb
-
-
 @register_backend("diskann")
 class DiskannBackend(LeannBackendFactoryInterface):
     @staticmethod
@@ -137,335 +49,85 @@ class DiskannBackend(LeannBackendFactoryInterface):
     def searcher(index_path: str, **kwargs) -> LeannBackendSearcherInterface:
         return DiskannSearcher(index_path, **kwargs)
 
-
 class DiskannBuilder(LeannBackendBuilderInterface):
     def __init__(self, **kwargs):
         self.build_params = kwargs
 
-    def _safe_cleanup_after_partition(self, index_dir: Path, index_prefix: str):
+    def build(self, data: np.ndarray, ids: List[str], index_path: str, **kwargs):
-        """
-        Safely cleanup files after partition.
-        In partition mode, C++ doesn't read _disk.index content,
-        so we can delete it if all derived files exist.
-        """
-        disk_index_file = index_dir / f"{index_prefix}_disk.index"
-        beam_search_file = index_dir / f"{index_prefix}_disk_beam_search.index"
-
-        # Required files that C++ partition mode needs
-        # Note: C++ generates these with _disk.index suffix
-        disk_suffix = "_disk.index"
-        required_files = [
-            f"{index_prefix}{disk_suffix}_medoids.bin",  # Critical: assert fails if missing
-            # Note: _centroids.bin is not created in single-shot build - C++ handles this automatically
-            f"{index_prefix}_pq_pivots.bin",  # PQ table
-            f"{index_prefix}_pq_compressed.bin",  # PQ compressed vectors
-        ]
-
-        # Check if all required files exist
-        missing_files = []
-        for filename in required_files:
-            file_path = index_dir / filename
-            if not file_path.exists():
-                missing_files.append(filename)
-
-        if missing_files:
-            logger.warning(
-                f"Cannot safely delete _disk.index - missing required files: {missing_files}"
-            )
-            logger.info("Keeping all original files for safety")
-            return
-
-        # Calculate space savings
-        space_saved = 0
-        files_to_delete = []
-
-        if disk_index_file.exists():
-            space_saved += disk_index_file.stat().st_size
-            files_to_delete.append(disk_index_file)
-
-        if beam_search_file.exists():
-            space_saved += beam_search_file.stat().st_size
-            files_to_delete.append(beam_search_file)
-
-        # Safe to delete!
-        for file_to_delete in files_to_delete:
-            try:
-                os.remove(file_to_delete)
-                logger.info(f"✅ Safely deleted: {file_to_delete.name}")
-            except Exception as e:
-                logger.warning(f"Failed to delete {file_to_delete.name}: {e}")
-
-        if space_saved > 0:
-            space_saved_mb = space_saved / (1024 * 1024)
-            logger.info(f"💾 Space saved: {space_saved_mb:.1f} MB")
-
-            # Show what files are kept
-            logger.info("📁 Kept essential files for partition mode:")
-            for filename in required_files:
-                file_path = index_dir / filename
-                if file_path.exists():
-                    size_mb = file_path.stat().st_size / (1024 * 1024)
-                    logger.info(f"  - {filename} ({size_mb:.1f} MB)")
-
-    def build(self, data: np.ndarray, ids: list[str], index_path: str, **kwargs):
         path = Path(index_path)
         index_dir = path.parent
         index_prefix = path.stem
         index_dir.mkdir(parents=True, exist_ok=True)
 
         if data.dtype != np.float32:
-            logger.warning(f"Converting data to float32, shape: {data.shape}")
             data = data.astype(np.float32)
 
         data_filename = f"{index_prefix}_data.bin"
         _write_vectors_to_bin(data, index_dir / data_filename)
 
+        label_map = {i: str_id for i, str_id in enumerate(ids)}
+        label_map_file = index_dir / "leann.labels.map"
+        with open(label_map_file, 'wb') as f:
+            pickle.dump(label_map, f)
+
         build_kwargs = {**self.build_params, **kwargs}
-
+        metric_enum = _get_diskann_metrics().get(build_kwargs.get("distance_metric", "mips").lower())
-        # Extract is_recompute from nested backend_kwargs if needed
-        is_recompute = build_kwargs.get("is_recompute", False)
-        if not is_recompute and "backend_kwargs" in build_kwargs:
-            is_recompute = build_kwargs["backend_kwargs"].get("is_recompute", False)
-
-        # Flatten all backend_kwargs parameters to top level for compatibility
-        if "backend_kwargs" in build_kwargs:
-            nested_params = build_kwargs.pop("backend_kwargs")
-            build_kwargs.update(nested_params)
-
-        metric_enum = _get_diskann_metrics().get(
-            build_kwargs.get("distance_metric", "mips").lower()
-        )
         if metric_enum is None:
-            raise ValueError(
+            raise ValueError(f"Unsupported distance_metric.")
-                f"Unsupported distance_metric '{build_kwargs.get('distance_metric', 'unknown')}'."
-            )
-
-        # Calculate smart memory configuration if not explicitly provided
-        if (
-            "search_memory_maximum" not in build_kwargs
-            or "build_memory_maximum" not in build_kwargs
-        ):
-            smart_search_mem, smart_build_mem = _calculate_smart_memory_config(data)
-        else:
-            smart_search_mem = build_kwargs.get("search_memory_maximum", 4.0)
-            smart_build_mem = build_kwargs.get("build_memory_maximum", 8.0)
 
         try:
-            from . import _diskannpy as diskannpy  # type: ignore
+            from . import _diskannpy as diskannpy
-
             with chdir(index_dir):
                 diskannpy.build_disk_float_index(
-                    metric_enum,
+                    metric_enum, data_filename, index_prefix,
-                    data_filename,
+                    build_kwargs.get("complexity", 64), build_kwargs.get("graph_degree", 32),
-                    index_prefix,
+                    build_kwargs.get("search_memory_maximum", 4.0), build_kwargs.get("build_memory_maximum", 8.0),
-                    build_kwargs.get("complexity", 64),
+                    build_kwargs.get("num_threads", 8), build_kwargs.get("pq_disk_bytes", 0), ""
-                    build_kwargs.get("graph_degree", 32),
-                    build_kwargs.get("search_memory_maximum", smart_search_mem),
-                    build_kwargs.get("build_memory_maximum", smart_build_mem),
-                    build_kwargs.get("num_threads", 8),
-                    build_kwargs.get("pq_disk_bytes", 0),
-                    "",
                 )
-
-            # Auto-partition if is_recompute is enabled
-            if build_kwargs.get("is_recompute", False):
-                logger.info("is_recompute=True, starting automatic graph partitioning...")
-                from .graph_partition import partition_graph
-
-                # Partition the index using absolute paths
-                # Convert to absolute paths to avoid issues with working directory changes
-                absolute_index_dir = Path(index_dir).resolve()
-                absolute_index_prefix_path = str(absolute_index_dir / index_prefix)
-                disk_graph_path, partition_bin_path = partition_graph(
-                    index_prefix_path=absolute_index_prefix_path,
-                    output_dir=str(absolute_index_dir),
-                    partition_prefix=index_prefix,
-                )
-
-                # Safe cleanup: In partition mode, C++ doesn't read _disk.index content
-                # but still needs the derived files (_medoids.bin, _centroids.bin, etc.)
-                self._safe_cleanup_after_partition(index_dir, index_prefix)
-
-                logger.info("✅ Graph partitioning completed successfully!")
-                logger.info(f"  - Disk graph: {disk_graph_path}")
-                logger.info(f"  - Partition file: {partition_bin_path}")
-
         finally:
             temp_data_file = index_dir / data_filename
             if temp_data_file.exists():
                 os.remove(temp_data_file)
-                logger.debug(f"Cleaned up temporary data file: {temp_data_file}")
-
 
 class DiskannSearcher(BaseSearcher):
     def __init__(self, index_path: str, **kwargs):
-        super().__init__(
+        super().__init__(index_path, backend_module_name="leann_backend_diskann.embedding_server", **kwargs)
-            index_path,
+        from . import _diskannpy as diskannpy
-            backend_module_name="leann_backend_diskann.diskann_embedding_server",
+
-            **kwargs,
+        distance_metric = kwargs.get("distance_metric", "mips").lower()
+        metric_enum = _get_diskann_metrics().get(distance_metric)
+        if metric_enum is None:
+            raise ValueError(f"Unsupported distance_metric '{distance_metric}'.")
+
+        self.num_threads = kwargs.get("num_threads", 8)
+        self.zmq_port = kwargs.get("zmq_port", 6666)
+
+        full_index_prefix = str(self.index_dir / self.index_path.stem)
+        self._index = diskannpy.StaticDiskFloatIndex(
+            metric_enum, full_index_prefix, self.num_threads, 
+            kwargs.get("num_nodes_to_cache", 0), 1, self.zmq_port, "", ""
         )
 
-        # Initialize DiskANN index with suppressed C++ output based on log level
+    def search(self, query: np.ndarray, top_k: int, **kwargs) -> Dict[str, Any]:
-        with suppress_cpp_output_if_needed():
+        recompute = kwargs.get("recompute_beighbor_embeddings", False)
-            from . import _diskannpy as diskannpy  # type: ignore
+        if recompute:
-
+            meta_file_path = self.index_dir / f"{self.index_path.name}.meta.json"
-            distance_metric = kwargs.get("distance_metric", "mips").lower()
+            if not meta_file_path.exists():
-            metric_enum = _get_diskann_metrics().get(distance_metric)
+                raise RuntimeError(f"FATAL: Recompute mode enabled but metadata file not found: {meta_file_path}")
-            if metric_enum is None:
+            zmq_port = kwargs.get("zmq_port", self.zmq_port)
-                raise ValueError(f"Unsupported distance_metric '{distance_metric}'.")
+            self._ensure_server_running(str(meta_file_path), port=zmq_port, **kwargs)
-
-            self.num_threads = kwargs.get("num_threads", 8)
-
-            # For DiskANN, we need to reinitialize the index when zmq_port changes
-            # Store the initialization parameters for later use
-            # Note: C++ load method expects the BASE path (without _disk.index suffix)
-            # C++ internally constructs: index_prefix + "_disk.index"
-            index_name = self.index_path.stem  # "simple_test.leann" -> "simple_test"
-            diskann_index_prefix = str(self.index_dir / index_name)  # /path/to/simple_test
-            full_index_prefix = diskann_index_prefix  # /path/to/simple_test (base path)
-
-            # Auto-detect partition files and set partition_prefix
-            partition_graph_file = self.index_dir / f"{index_name}_disk_graph.index"
-            partition_bin_file = self.index_dir / f"{index_name}_partition.bin"
-
-            partition_prefix = ""
-            if partition_graph_file.exists() and partition_bin_file.exists():
-                # C++ expects full path prefix, not just filename
-                partition_prefix = str(self.index_dir / index_name)  # /path/to/simple_test
-                logger.info(
-                    f"✅ Detected partition files, using partition_prefix='{partition_prefix}'"
-                )
-            else:
-                logger.debug("No partition files detected, using standard index files")
-
-            self._init_params = {
-                "metric_enum": metric_enum,
-                "full_index_prefix": full_index_prefix,
-                "num_threads": self.num_threads,
-                "num_nodes_to_cache": kwargs.get("num_nodes_to_cache", 0),
-                "cache_mechanism": 1,
-                "pq_prefix": "",
-                "partition_prefix": partition_prefix,
-            }
-
-            # Log partition configuration for debugging
-            if partition_prefix:
-                logger.info(
-                    f"✅ Detected partition files, using partition_prefix='{partition_prefix}'"
-                )
-            self._diskannpy = diskannpy
-            self._current_zmq_port = None
-            self._index = None
-            logger.debug("DiskANN searcher initialized (index will be loaded on first search)")
-
-    def _ensure_index_loaded(self, zmq_port: int):
-        """Ensure the index is loaded with the correct zmq_port."""
-        if self._index is None or self._current_zmq_port != zmq_port:
-            # Need to (re)load the index with the correct zmq_port
-            with suppress_cpp_output_if_needed():
-                if self._index is not None:
-                    logger.debug(f"Reloading DiskANN index with new zmq_port: {zmq_port}")
-                else:
-                    logger.debug(f"Loading DiskANN index with zmq_port: {zmq_port}")
-
-                self._index = self._diskannpy.StaticDiskFloatIndex(
-                    self._init_params["metric_enum"],
-                    self._init_params["full_index_prefix"],
-                    self._init_params["num_threads"],
-                    self._init_params["num_nodes_to_cache"],
-                    self._init_params["cache_mechanism"],
-                    zmq_port,
-                    self._init_params["pq_prefix"],
-                    self._init_params["partition_prefix"],
-                )
-                self._current_zmq_port = zmq_port
-
-    def search(
-        self,
-        query: np.ndarray,
-        top_k: int,
-        complexity: int = 64,
-        beam_width: int = 1,
-        prune_ratio: float = 0.0,
-        recompute_embeddings: bool = False,
-        pruning_strategy: Literal["global", "local", "proportional"] = "global",
-        zmq_port: Optional[int] = None,
-        batch_recompute: bool = False,
-        dedup_node_dis: bool = False,
-        **kwargs,
-    ) -> dict[str, Any]:
-        """
-        Search for nearest neighbors using DiskANN index.
-
-        Args:
-            query: Query vectors (B, D) where B is batch size, D is dimension
-            top_k: Number of nearest neighbors to return
-            complexity: Search complexity/candidate list size, higher = more accurate but slower
-            beam_width: Number of parallel IO requests per iteration
-            prune_ratio: Ratio of neighbors to prune via approximate distance (0.0-1.0)
-            recompute_embeddings: Whether to fetch fresh embeddings from server
-            pruning_strategy: PQ candidate selection strategy:
-                - "global": Use global pruning strategy (default)
-                - "local": Use local pruning strategy
-                - "proportional": Not supported in DiskANN, falls back to global
-            zmq_port: ZMQ port for embedding server communication. Must be provided if recompute_embeddings is True.
-            batch_recompute: Whether to batch neighbor recomputation (DiskANN-specific)
-            dedup_node_dis: Whether to cache and reuse distance computations (DiskANN-specific)
-            **kwargs: Additional DiskANN-specific parameters (for legacy compatibility)
-
-        Returns:
-            Dict with 'labels' (list of lists) and 'distances' (ndarray)
-        """
-        # Handle zmq_port compatibility: Ensure index is loaded with correct port
-        if recompute_embeddings:
-            if zmq_port is None:
-                raise ValueError("zmq_port must be provided if recompute_embeddings is True")
-            self._ensure_index_loaded(zmq_port)
-        else:
-            # If not recomputing, we still need an index, use a default port
-            if self._index is None:
-                self._ensure_index_loaded(6666)  # Default port when not recomputing
-
-        # DiskANN doesn't support "proportional" strategy
-        if pruning_strategy == "proportional":
-            raise NotImplementedError(
-                "DiskANN backend does not support 'proportional' pruning strategy. Use 'global' or 'local' instead."
-            )
 
         if query.dtype != np.float32:
             query = query.astype(np.float32)
 
-        # Map pruning_strategy to DiskANN's global_pruning parameter
+        labels, distances = self._index.batch_search(
-        if pruning_strategy == "local":
+            query, query.shape[0], top_k,
-            use_global_pruning = False
+            kwargs.get("complexity", 256), kwargs.get("beam_width", 4), self.num_threads,
-        else:  # "global"
+            kwargs.get("USE_DEFERRED_FETCH", False), kwargs.get("skip_search_reorder", False),
-            use_global_pruning = True
+            recompute, kwargs.get("dedup_node_dis", False), kwargs.get("prune_ratio", 0.0),
+            kwargs.get("batch_recompute", False), kwargs.get("global_pruning", False)
+        )
 
-        # Strategy:
+        string_labels = [[self.label_map.get(int_label, f"unknown_{int_label}") for int_label in batch_labels] for batch_labels in labels]
-        # - Traversal always uses PQ distances
-        # - If recompute_embeddings=True, do a single final rerank via deferred fetch
-        #   (fetch embeddings for the final candidate set only)
-        # - Do not recompute neighbor distances along the path
-        use_deferred_fetch = True if recompute_embeddings else False
-        recompute_neighors = False  # Expected typo. For backward compatibility.
 
-        with suppress_cpp_output_if_needed():
+        return {"labels": string_labels, "distances": distances}
-            labels, distances = self._index.batch_search(
-                query,
-                query.shape[0],
-                top_k,
-                complexity,
-                beam_width,
-                self.num_threads,
-                use_deferred_fetch,
-                kwargs.get("skip_search_reorder", False),
-                recompute_neighors,
-                dedup_node_dis,
-                prune_ratio,
-                batch_recompute,
-                use_global_pruning,
-            )
-
-        string_labels = [[str(int_label) for int_label in batch_labels] for batch_labels in labels]
-
-        return {"labels": string_labels, "distances": distances}

packages/leann-backend-diskann/leann_backend_diskann/diskann_embedding_server.py

@@ -1,492 +0,0 @@
-"""
-DiskANN-specific embedding server
-"""
-
-import argparse
-import json
-import logging
-import os
-import sys
-import threading
-import time
-from pathlib import Path
-from typing import Any, Optional
-
-import numpy as np
-import zmq
-
-# Set up logging based on environment variable
-LOG_LEVEL = os.getenv("LEANN_LOG_LEVEL", "WARNING").upper()
-logger = logging.getLogger(__name__)
-
-# Force set logger level (don't rely on basicConfig in subprocess)
-log_level = getattr(logging, LOG_LEVEL, logging.WARNING)
-logger.setLevel(log_level)
-
-# Ensure we have a handler if none exists
-if not logger.handlers:
-    handler = logging.StreamHandler()
-    formatter = logging.Formatter("%(asctime)s - %(levelname)s - %(message)s")
-    handler.setFormatter(formatter)
-    logger.addHandler(handler)
-    logger.propagate = False
-
-
-_RAW_PROVIDER_OPTIONS = os.getenv("LEANN_EMBEDDING_OPTIONS")
-try:
-    PROVIDER_OPTIONS: dict[str, Any] = (
-        json.loads(_RAW_PROVIDER_OPTIONS) if _RAW_PROVIDER_OPTIONS else {}
-    )
-except json.JSONDecodeError:
-    logger.warning("Failed to parse LEANN_EMBEDDING_OPTIONS; ignoring provider options")
-    PROVIDER_OPTIONS = {}
-
-
-def create_diskann_embedding_server(
-    passages_file: Optional[str] = None,
-    zmq_port: int = 5555,
-    model_name: str = "sentence-transformers/all-mpnet-base-v2",
-    embedding_mode: str = "sentence-transformers",
-    distance_metric: str = "l2",
-):
-    """
-    Create and start a ZMQ-based embedding server for DiskANN backend.
-    Uses ROUTER socket and protobuf communication as required by DiskANN C++ implementation.
-    """
-    logger.info(f"Starting DiskANN server on port {zmq_port} with model {model_name}")
-    logger.info(f"Using embedding mode: {embedding_mode}")
-
-    # Add leann-core to path for unified embedding computation
-    current_dir = Path(__file__).parent
-    leann_core_path = current_dir.parent.parent / "leann-core" / "src"
-    sys.path.insert(0, str(leann_core_path))
-
-    try:
-        from leann.api import PassageManager
-        from leann.embedding_compute import compute_embeddings
-
-        logger.info("Successfully imported unified embedding computation module")
-    except ImportError as e:
-        logger.error(f"Failed to import embedding computation module: {e}")
-        return
-    finally:
-        sys.path.pop(0)
-
-    # Check port availability
-    import socket
-
-    def check_port(port):
-        with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
-            return s.connect_ex(("localhost", port)) == 0
-
-    if check_port(zmq_port):
-        logger.error(f"Port {zmq_port} is already in use")
-        return
-
-    # Only support metadata file, fail fast for everything else
-    if not passages_file or not passages_file.endswith(".meta.json"):
-        raise ValueError("Only metadata files (.meta.json) are supported")
-
-    # Load metadata to get passage sources
-    with open(passages_file) as f:
-        meta = json.load(f)
-
-    logger.info(f"Loading PassageManager with metadata_file_path: {passages_file}")
-    passages = PassageManager(meta["passage_sources"], metadata_file_path=passages_file)
-    logger.info(f"Loaded PassageManager with {len(passages)} passages from metadata")
-
-    # Import protobuf after ensuring the path is correct
-    try:
-        from . import embedding_pb2
-    except ImportError as e:
-        logger.error(f"Failed to import protobuf module: {e}")
-        return
-
-    def zmq_server_thread():
-        """ZMQ server thread using REP socket for universal compatibility"""
-        context = zmq.Context()
-        socket = context.socket(
-            zmq.REP
-        )  # REP socket for both BaseSearcher and DiskANN C++ REQ clients
-        socket.bind(f"tcp://*:{zmq_port}")
-        logger.info(f"DiskANN ZMQ REP server listening on port {zmq_port}")
-
-        socket.setsockopt(zmq.RCVTIMEO, 1000)
-        socket.setsockopt(zmq.SNDTIMEO, 1000)
-        socket.setsockopt(zmq.LINGER, 0)
-
-        while True:
-            try:
-                # REP socket receives single-part messages
-                message = socket.recv()
-
-                # Check for empty messages - REP socket requires response to every request
-                if len(message) == 0:
-                    logger.debug("Received empty message, sending empty response")
-                    socket.send(b"")  # REP socket must respond to every request
-                    continue
-
-                logger.debug(f"Received ZMQ request of size {len(message)} bytes")
-                logger.debug(f"Message preview: {message[:50]}")  # Show first 50 bytes
-
-                e2e_start = time.time()
-
-                # Try protobuf first (for DiskANN C++ node_ids requests - primary use case)
-                texts = []
-                node_ids = []
-                is_text_request = False
-
-                try:
-                    req_proto = embedding_pb2.NodeEmbeddingRequest()
-                    req_proto.ParseFromString(message)
-                    node_ids = list(req_proto.node_ids)
-
-                    if not node_ids:
-                        raise RuntimeError(
-                            f"PROTOBUF: Received empty node_ids! Message size: {len(message)}"
-                        )
-
-                    logger.info(
-                        f"✅ PROTOBUF: Node ID request for {len(node_ids)} node embeddings: {node_ids[:10]}"
-                    )
-                except Exception as protobuf_error:
-                    logger.debug(f"Protobuf parsing failed: {protobuf_error}")
-                    # Fallback to msgpack (for BaseSearcher direct text requests)
-                    try:
-                        import msgpack
-
-                        request = msgpack.unpackb(message)
-                        # For BaseSearcher compatibility, request is a list of texts directly
-                        if isinstance(request, list) and all(
-                            isinstance(item, str) for item in request
-                        ):
-                            texts = request
-                            is_text_request = True
-                            logger.info(f"✅ MSGPACK: Direct text request for {len(texts)} texts")
-                        else:
-                            raise ValueError("Not a valid msgpack text request")
-                    except Exception as msgpack_error:
-                        raise RuntimeError(
-                            f"Both protobuf and msgpack parsing failed! Protobuf: {protobuf_error}, Msgpack: {msgpack_error}"
-                        )
-
-                # Look up texts by node IDs (only if not direct text request)
-                if not is_text_request:
-                    for nid in node_ids:
-                        try:
-                            passage_data = passages.get_passage(str(nid))
-                            txt = passage_data["text"]
-                            if not txt:
-                                raise RuntimeError(f"FATAL: Empty text for passage ID {nid}")
-                            texts.append(txt)
-                        except KeyError as e:
-                            logger.error(f"Passage ID {nid} not found: {e}")
-                            raise e
-                        except Exception as e:
-                            logger.error(f"Exception looking up passage ID {nid}: {e}")
-                            raise
-
-                    # Debug logging
-                    logger.debug(f"Processing {len(texts)} texts")
-                    logger.debug(f"Text lengths: {[len(t) for t in texts[:5]]}")  # Show first 5
-
-                # Process embeddings using unified computation
-                embeddings = compute_embeddings(
-                    texts,
-                    model_name,
-                    mode=embedding_mode,
-                    provider_options=PROVIDER_OPTIONS,
-                )
-                logger.info(
-                    f"Computed embeddings for {len(texts)} texts, shape: {embeddings.shape}"
-                )
-
-                # Prepare response based on request type
-                if is_text_request:
-                    # For BaseSearcher compatibility: return msgpack format
-                    import msgpack
-
-                    response_data = msgpack.packb(embeddings.tolist())
-                else:
-                    # For DiskANN C++ compatibility: return protobuf format
-                    resp_proto = embedding_pb2.NodeEmbeddingResponse()
-                    hidden_contiguous = np.ascontiguousarray(embeddings, dtype=np.float32)
-
-                    # Serialize embeddings data
-                    resp_proto.embeddings_data = hidden_contiguous.tobytes()
-                    resp_proto.dimensions.append(hidden_contiguous.shape[0])
-                    resp_proto.dimensions.append(hidden_contiguous.shape[1])
-
-                    response_data = resp_proto.SerializeToString()
-
-                # Send response back to the client
-                socket.send(response_data)
-
-                e2e_end = time.time()
-                logger.info(f"⏱️  ZMQ E2E time: {e2e_end - e2e_start:.6f}s")
-
-            except zmq.Again:
-                logger.debug("ZMQ socket timeout, continuing to listen")
-                continue
-            except Exception as e:
-                logger.error(f"Error in ZMQ server loop: {e}")
-                import traceback
-
-                traceback.print_exc()
-                raise
-
-    def zmq_server_thread_with_shutdown(shutdown_event):
-        """ZMQ server thread that respects shutdown signal.
-
-        This creates its own REP socket, binds to zmq_port, and periodically
-        checks shutdown_event using recv timeouts to exit cleanly.
-        """
-        logger.info("DiskANN ZMQ server thread started with shutdown support")
-
-        context = zmq.Context()
-        rep_socket = context.socket(zmq.REP)
-        rep_socket.bind(f"tcp://*:{zmq_port}")
-        logger.info(f"DiskANN ZMQ REP server listening on port {zmq_port}")
-
-        # Set receive timeout so we can check shutdown_event periodically
-        rep_socket.setsockopt(zmq.RCVTIMEO, 1000)  # 1 second timeout
-        rep_socket.setsockopt(zmq.SNDTIMEO, 1000)
-        rep_socket.setsockopt(zmq.LINGER, 0)
-
-        try:
-            while not shutdown_event.is_set():
-                try:
-                    e2e_start = time.time()
-                    # REP socket receives single-part messages
-                    message = rep_socket.recv()
-
-                    # Check for empty messages - REP socket requires response to every request
-                    if not message:
-                        logger.warning("Received empty message, sending empty response")
-                        rep_socket.send(b"")
-                        continue
-
-                    # Try protobuf first (same logic as original)
-                    texts = []
-                    is_text_request = False
-
-                    try:
-                        req_proto = embedding_pb2.NodeEmbeddingRequest()
-                        req_proto.ParseFromString(message)
-                        node_ids = list(req_proto.node_ids)
-
-                        # Look up texts by node IDs
-                        for nid in node_ids:
-                            try:
-                                passage_data = passages.get_passage(str(nid))
-                                txt = passage_data["text"]
-                                if not txt:
-                                    raise RuntimeError(f"FATAL: Empty text for passage ID {nid}")
-                                texts.append(txt)
-                            except KeyError:
-                                raise RuntimeError(f"FATAL: Passage with ID {nid} not found")
-
-                        logger.info(f"ZMQ received protobuf request for {len(node_ids)} node IDs")
-                    except Exception:
-                        # Fallback to msgpack for text requests
-                        try:
-                            import msgpack
-
-                            request = msgpack.unpackb(message)
-                            if isinstance(request, list) and all(
-                                isinstance(item, str) for item in request
-                            ):
-                                texts = request
-                                is_text_request = True
-                                logger.info(
-                                    f"ZMQ received msgpack text request for {len(texts)} texts"
-                                )
-                            else:
-                                raise ValueError("Not a valid msgpack text request")
-                        except Exception:
-                            logger.error("Both protobuf and msgpack parsing failed!")
-                            # Send error response
-                            resp_proto = embedding_pb2.NodeEmbeddingResponse()
-                            rep_socket.send(resp_proto.SerializeToString())
-                            continue
-
-                    # Process the request
-                    embeddings = compute_embeddings(
-                        texts,
-                        model_name,
-                        mode=embedding_mode,
-                        provider_options=PROVIDER_OPTIONS,
-                    )
-                    logger.info(f"Computed embeddings shape: {embeddings.shape}")
-
-                    # Validation
-                    if np.isnan(embeddings).any() or np.isinf(embeddings).any():
-                        logger.error("NaN or Inf detected in embeddings!")
-                        # Send error response
-                        if is_text_request:
-                            import msgpack
-
-                            response_data = msgpack.packb([])
-                        else:
-                            resp_proto = embedding_pb2.NodeEmbeddingResponse()
-                            response_data = resp_proto.SerializeToString()
-                        rep_socket.send(response_data)
-                        continue
-
-                    # Prepare response based on request type
-                    if is_text_request:
-                        # For direct text requests, return msgpack
-                        import msgpack
-
-                        response_data = msgpack.packb(embeddings.tolist())
-                    else:
-                        # For protobuf requests, return protobuf
-                        resp_proto = embedding_pb2.NodeEmbeddingResponse()
-                        hidden_contiguous = np.ascontiguousarray(embeddings, dtype=np.float32)
-
-                        resp_proto.embeddings_data = hidden_contiguous.tobytes()
-                        resp_proto.dimensions.append(hidden_contiguous.shape[0])
-                        resp_proto.dimensions.append(hidden_contiguous.shape[1])
-
-                        response_data = resp_proto.SerializeToString()
-
-                    # Send response back to the client
-                    rep_socket.send(response_data)
-
-                    e2e_end = time.time()
-                    logger.info(f"⏱️  ZMQ E2E time: {e2e_end - e2e_start:.6f}s")
-
-                except zmq.Again:
-                    # Timeout - check shutdown_event and continue
-                    continue
-                except Exception as e:
-                    if not shutdown_event.is_set():
-                        logger.error(f"Error in ZMQ server loop: {e}")
-                        try:
-                            # Send error response for REP socket
-                            resp_proto = embedding_pb2.NodeEmbeddingResponse()
-                            rep_socket.send(resp_proto.SerializeToString())
-                        except Exception:
-                            pass
-                    else:
-                        logger.info("Shutdown in progress, ignoring ZMQ error")
-                        break
-        finally:
-            try:
-                rep_socket.close(0)
-            except Exception:
-                pass
-            try:
-                context.term()
-            except Exception:
-                pass
-
-        logger.info("DiskANN ZMQ server thread exiting gracefully")
-
-    # Add shutdown coordination
-    shutdown_event = threading.Event()
-
-    def shutdown_zmq_server():
-        """Gracefully shutdown ZMQ server."""
-        logger.info("Initiating graceful shutdown...")
-        shutdown_event.set()
-
-        if zmq_thread.is_alive():
-            logger.info("Waiting for ZMQ thread to finish...")
-            zmq_thread.join(timeout=5)
-            if zmq_thread.is_alive():
-                logger.warning("ZMQ thread did not finish in time")
-
-        # Clean up ZMQ resources
-        try:
-            # Note: socket and context are cleaned up by thread exit
-            logger.info("ZMQ resources cleaned up")
-        except Exception as e:
-            logger.warning(f"Error cleaning ZMQ resources: {e}")
-
-        # Clean up other resources
-        try:
-            import gc
-
-            gc.collect()
-            logger.info("Additional resources cleaned up")
-        except Exception as e:
-            logger.warning(f"Error cleaning additional resources: {e}")
-
-        logger.info("Graceful shutdown completed")
-        sys.exit(0)
-
-    # Register signal handlers within this function scope
-    import signal
-
-    def signal_handler(sig, frame):
-        logger.info(f"Received signal {sig}, shutting down gracefully...")
-        shutdown_zmq_server()
-
-    signal.signal(signal.SIGTERM, signal_handler)
-    signal.signal(signal.SIGINT, signal_handler)
-
-    # Start ZMQ thread (NOT daemon!)
-    zmq_thread = threading.Thread(
-        target=lambda: zmq_server_thread_with_shutdown(shutdown_event),
-        daemon=False,  # Not daemon - we want to wait for it
-    )
-    zmq_thread.start()
-    logger.info(f"Started DiskANN ZMQ server thread on port {zmq_port}")
-
-    # Keep the main thread alive
-    try:
-        while not shutdown_event.is_set():
-            time.sleep(0.1)  # Check shutdown more frequently
-    except KeyboardInterrupt:
-        logger.info("DiskANN Server shutting down...")
-        shutdown_zmq_server()
-        return
-
-    # If we reach here, shutdown was triggered by signal
-    logger.info("Main loop exited, process should be shutting down")
-
-
-if __name__ == "__main__":
-    import sys
-
-    # Signal handlers are now registered within create_diskann_embedding_server
-
-    parser = argparse.ArgumentParser(description="DiskANN Embedding service")
-    parser.add_argument("--zmq-port", type=int, default=5555, help="ZMQ port to run on")
-    parser.add_argument(
-        "--passages-file",
-        type=str,
-        help="Metadata JSON file containing passage sources",
-    )
-    parser.add_argument(
-        "--model-name",
-        type=str,
-        default="sentence-transformers/all-mpnet-base-v2",
-        help="Embedding model name",
-    )
-    parser.add_argument(
-        "--embedding-mode",
-        type=str,
-        default="sentence-transformers",
-        choices=["sentence-transformers", "openai", "mlx", "ollama"],
-        help="Embedding backend mode",
-    )
-    parser.add_argument(
-        "--distance-metric",
-        type=str,
-        default="l2",
-        choices=["l2", "mips", "cosine"],
-        help="Distance metric for similarity computation",
-    )
-
-    args = parser.parse_args()
-
-    # Create and start the DiskANN embedding server
-    create_diskann_embedding_server(
-        passages_file=args.passages_file,
-        zmq_port=args.zmq_port,
-        model_name=args.model_name,
-        embedding_mode=args.embedding_mode,
-        distance_metric=args.distance_metric,
-    )

packages/leann-backend-diskann/leann_backend_diskann/embedding_pb2.py

@@ -1,28 +1,27 @@
+# -*- coding: utf-8 -*-
 # Generated by the protocol buffer compiler.  DO NOT EDIT!
 # source: embedding.proto
-# ruff: noqa
 """Generated protocol buffer code."""
-
+from google.protobuf.internal import builder as _builder
 from google.protobuf import descriptor as _descriptor
 from google.protobuf import descriptor_pool as _descriptor_pool
 from google.protobuf import symbol_database as _symbol_database
-from google.protobuf.internal import builder as _builder
-
 # @@protoc_insertion_point(imports)
 
 _sym_db = _symbol_database.Default()
 
 
-DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(
+
-    b'\n\x0f\x65mbedding.proto\x12\x0eprotoembedding"(\n\x14NodeEmbeddingRequest\x12\x10\n\x08node_ids\x18\x01 \x03(\r"Y\n\x15NodeEmbeddingResponse\x12\x17\n\x0f\x65mbeddings_data\x18\x01 \x01(\x0c\x12\x12\n\ndimensions\x18\x02 \x03(\x05\x12\x13\n\x0bmissing_ids\x18\x03 \x03(\rb\x06proto3'
+
-)
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x0f\x65mbedding.proto\x12\x0eprotoembedding\"(\n\x14NodeEmbeddingRequest\x12\x10\n\x08node_ids\x18\x01 \x03(\r\"Y\n\x15NodeEmbeddingResponse\x12\x17\n\x0f\x65mbeddings_data\x18\x01 \x01(\x0c\x12\x12\n\ndimensions\x18\x02 \x03(\x05\x12\x13\n\x0bmissing_ids\x18\x03 \x03(\rb\x06proto3')
 
 _builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, globals())
-_builder.BuildTopDescriptorsAndMessages(DESCRIPTOR, "embedding_pb2", globals())
+_builder.BuildTopDescriptorsAndMessages(DESCRIPTOR, 'embedding_pb2', globals())
-if not _descriptor._USE_C_DESCRIPTORS:
+if _descriptor._USE_C_DESCRIPTORS == False:
-    DESCRIPTOR._options = None
+
-    _NODEEMBEDDINGREQUEST._serialized_start = 35
+  DESCRIPTOR._options = None
-    _NODEEMBEDDINGREQUEST._serialized_end = 75
+  _NODEEMBEDDINGREQUEST._serialized_start=35
-    _NODEEMBEDDINGRESPONSE._serialized_start = 77
+  _NODEEMBEDDINGREQUEST._serialized_end=75
-    _NODEEMBEDDINGRESPONSE._serialized_end = 166
+  _NODEEMBEDDINGRESPONSE._serialized_start=77
+  _NODEEMBEDDINGRESPONSE._serialized_end=166
 # @@protoc_insertion_point(module_scope)

packages/leann-backend-diskann/leann_backend_diskann/embedding_server.py

@@ -0,0 +1,512 @@
+#!/usr/bin/env python3
+"""
+Embedding server for leann-backend-diskann - Fixed ZMQ REQ-REP pattern
+"""
+
+import pickle
+import argparse
+import threading
+import time
+import json
+from typing import Dict, Any, Optional, Union
+
+from transformers import AutoTokenizer, AutoModel
+import os
+from contextlib import contextmanager
+import zmq
+import numpy as np
+from pathlib import Path
+import pickle
+
+RED = "\033[91m"
+RESET = "\033[0m"
+
+# --- New Passage Loader from HNSW backend ---
+class SimplePassageLoader:
+    """
+    Simple passage loader that replaces config.py dependencies
+    """
+    def __init__(self, passages_data: Optional[Dict[str, Any]] = None):
+        self.passages_data = passages_data or {}
+    
+    def __getitem__(self, passage_id: Union[str, int]) -> Dict[str, str]:
+        """Get passage by ID"""
+        str_id = str(passage_id)
+        if str_id in self.passages_data:
+            return {"text": self.passages_data[str_id]}
+        else:
+            # Return empty text for missing passages
+            return {"text": ""}
+    
+    def __len__(self) -> int:
+        return len(self.passages_data)
+
+def load_passages_from_metadata(meta_file: str) -> SimplePassageLoader:
+    """
+    Load passages using metadata file with PassageManager for lazy loading
+    """
+    # Load metadata to get passage sources
+    with open(meta_file, 'r') as f:
+        meta = json.load(f)
+    
+    # Import PassageManager dynamically to avoid circular imports
+    import sys
+    from pathlib import Path
+    
+    # Find the leann package directory relative to this file
+    current_dir = Path(__file__).parent
+    leann_core_path = current_dir.parent.parent / "leann-core" / "src"
+    sys.path.insert(0, str(leann_core_path))
+    
+    try:
+        from leann.api import PassageManager
+        passage_manager = PassageManager(meta['passage_sources'])
+    finally:
+        sys.path.pop(0)
+    
+    # Load label map 
+    passages_dir = Path(meta_file).parent
+    label_map_file = passages_dir / "leann.labels.map"
+    
+    if label_map_file.exists():
+        import pickle
+        with open(label_map_file, 'rb') as f:
+            label_map = pickle.load(f)
+        print(f"Loaded label map with {len(label_map)} entries")
+    else:
+        raise FileNotFoundError(f"Label map file not found: {label_map_file}")
+    
+    print(f"Initialized lazy passage loading for {len(label_map)} passages")
+    
+    class LazyPassageLoader(SimplePassageLoader):
+        def __init__(self, passage_manager, label_map):
+            self.passage_manager = passage_manager
+            self.label_map = label_map
+            # Initialize parent with empty data
+            super().__init__({})
+        
+        def __getitem__(self, passage_id: Union[str, int]) -> Dict[str, str]:
+            """Get passage by ID with lazy loading"""
+            try:
+                int_id = int(passage_id)
+                if int_id in self.label_map:
+                    string_id = self.label_map[int_id]
+                    passage_data = self.passage_manager.get_passage(string_id)
+                    if passage_data and passage_data.get("text"):
+                        return {"text": passage_data["text"]}
+                    else:
+                        raise RuntimeError(f"FATAL: Empty text for ID {int_id} -> {string_id}")
+                else:
+                    raise RuntimeError(f"FATAL: ID {int_id} not found in label_map")
+            except Exception as e:
+                raise RuntimeError(f"FATAL: Exception getting passage {passage_id}: {e}")
+        
+        def __len__(self) -> int:
+            return len(self.label_map)
+    
+    return LazyPassageLoader(passage_manager, label_map)
+
+def load_passages_from_file(passages_file: str) -> SimplePassageLoader:
+    """
+    Load passages from a JSONL file with label map support
+    Expected format: {"id": "passage_id", "text": "passage_text", "metadata": {...}} (one per line)
+    """
+    
+    if not os.path.exists(passages_file):
+        raise FileNotFoundError(f"Passages file {passages_file} not found.")
+    
+    if not passages_file.endswith('.jsonl'):
+        raise ValueError(f"Expected .jsonl file format, got: {passages_file}")
+    
+    # Load label map (int -> string_id)
+    passages_dir = Path(passages_file).parent
+    label_map_file = passages_dir / "leann.labels.map"
+    
+    label_map = {}
+    if label_map_file.exists():
+        with open(label_map_file, 'rb') as f:
+            label_map = pickle.load(f)
+        print(f"Loaded label map with {len(label_map)} entries")
+    else:
+        raise FileNotFoundError(f"Label map file not found: {label_map_file}")
+    
+    # Load passages by string ID
+    string_id_passages = {}
+    with open(passages_file, 'r', encoding='utf-8') as f:
+        for line in f:
+            if line.strip():
+                passage = json.loads(line)
+                string_id_passages[passage['id']] = passage['text']
+    
+    # Create int ID -> text mapping using label map
+    passages_data = {}
+    for int_id, string_id in label_map.items():
+        if string_id in string_id_passages:
+            passages_data[str(int_id)] = string_id_passages[string_id]
+        else:
+            print(f"WARNING: String ID {string_id} from label map not found in passages")
+    
+    print(f"Loaded {len(passages_data)} passages from JSONL file {passages_file} using label map")
+    return SimplePassageLoader(passages_data)
+
+def create_embedding_server_thread(
+    zmq_port=5555,
+    model_name="sentence-transformers/all-mpnet-base-v2",
+    max_batch_size=128,
+    passages_file: Optional[str] = None,
+):
+    """
+    在当前线程中创建并运行 embedding server
+    这个函数设计为在单独的线程中调用
+    """
+    print(f"INFO: Initializing embedding server thread on port {zmq_port}")
+    
+    try:
+        # 检查端口是否已被占用
+        import socket
+        def check_port(port):
+            with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
+                return s.connect_ex(('localhost', port)) == 0
+
+        if check_port(zmq_port):
+            print(f"{RED}Port {zmq_port} is already in use{RESET}")
+            return
+
+        # 初始化模型
+        tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=True)
+        import torch
+
+        # 选择设备
+        mps_available = hasattr(torch.backends, 'mps') and torch.backends.mps.is_available()
+        cuda_available = torch.cuda.is_available()
+        
+        if cuda_available:
+            device = torch.device("cuda")
+            print("INFO: Using CUDA device")
+        elif mps_available:
+            device = torch.device("mps")
+            print("INFO: Using MPS device (Apple Silicon)")
+        else:
+            device = torch.device("cpu")
+            print("INFO: Using CPU device")
+        
+        # 加载模型
+        print(f"INFO: Loading model {model_name}")
+        model = AutoModel.from_pretrained(model_name).to(device).eval()
+
+        # 优化模型
+        if cuda_available or mps_available:
+            try:
+                model = model.half()
+                model = torch.compile(model)
+                print(f"INFO: Using FP16 precision with model: {model_name}")
+            except Exception as e:
+                print(f"WARNING: Model optimization failed: {e}")
+
+        # Load passages from file if provided
+        if passages_file and os.path.exists(passages_file):
+            # Check if it's a metadata file or a single passages file
+            if passages_file.endswith('.meta.json'):
+                passages = load_passages_from_metadata(passages_file)
+            else:
+                # Try to find metadata file in same directory
+                passages_dir = Path(passages_file).parent
+                meta_files = list(passages_dir.glob("*.meta.json"))
+                if meta_files:
+                    print(f"Found metadata file: {meta_files[0]}, using lazy loading")
+                    passages = load_passages_from_metadata(str(meta_files[0]))
+                else:
+                    # Fallback to original single file loading (will cause warnings)
+                    print("WARNING: No metadata file found, using single file loading (may cause missing passage warnings)")
+                    passages = load_passages_from_file(passages_file)
+        else:
+            print("WARNING: No passages file provided or file not found. Using an empty passage loader.")
+            passages = SimplePassageLoader()
+
+        print(f"INFO: Loaded {len(passages)} passages.")
+
+        class DeviceTimer:
+            """设备计时器"""
+            def __init__(self, name="", device=device):
+                self.name = name
+                self.device = device
+                self.start_time = 0
+                self.end_time = 0
+                
+                if cuda_available:
+                    self.start_event = torch.cuda.Event(enable_timing=True)
+                    self.end_event = torch.cuda.Event(enable_timing=True)
+                else:
+                    self.start_event = None
+                    self.end_event = None
+
+            @contextmanager
+            def timing(self):
+                self.start()
+                yield
+                self.end()
+
+            def start(self):
+                if cuda_available:
+                    torch.cuda.synchronize()
+                    self.start_event.record()
+                else:
+                    if self.device.type == "mps":
+                        torch.mps.synchronize()
+                    self.start_time = time.time()
+
+            def end(self):
+                if cuda_available:
+                    self.end_event.record()
+                    torch.cuda.synchronize()
+                else:
+                    if self.device.type == "mps":
+                        torch.mps.synchronize()
+                    self.end_time = time.time()
+
+            def elapsed_time(self):
+                if cuda_available:
+                    return self.start_event.elapsed_time(self.end_event) / 1000.0
+                else:
+                    return self.end_time - self.start_time
+
+            def print_elapsed(self):
+                print(f"Time taken for {self.name}: {self.elapsed_time():.6f} seconds")
+
+        def process_batch(texts_batch, ids_batch, missing_ids):
+            """处理文本批次"""
+            batch_size = len(texts_batch)
+            print(f"INFO: Processing batch of size {batch_size}")
+
+            tokenize_timer = DeviceTimer("tokenization (batch)", device)
+            to_device_timer = DeviceTimer("transfer to device (batch)", device)
+            embed_timer = DeviceTimer("embedding (batch)", device)
+            pool_timer = DeviceTimer("mean pooling (batch)", device)
+
+            with tokenize_timer.timing():
+                encoded_batch = tokenizer.batch_encode_plus(
+                    texts_batch,
+                    padding="max_length",
+                    truncation=True,
+                    max_length=256,
+                    return_tensors="pt",
+                    return_token_type_ids=False,
+                )
+            tokenize_timer.print_elapsed()
+
+            seq_length = encoded_batch["input_ids"].size(1)
+            print(f"Batch size: {batch_size}, Sequence length: {seq_length}")
+
+            with to_device_timer.timing():
+                enc = {k: v.to(device) for k, v in encoded_batch.items()}
+            to_device_timer.print_elapsed()
+
+            with torch.no_grad():
+                with embed_timer.timing():
+                    out = model(enc["input_ids"], enc["attention_mask"])
+                embed_timer.print_elapsed()
+
+                with pool_timer.timing():
+                    hidden_states = out.last_hidden_state if hasattr(out, "last_hidden_state") else out
+                    mask_expanded = enc["attention_mask"].unsqueeze(-1).expand(hidden_states.size()).float()
+                    sum_embeddings = torch.sum(hidden_states * mask_expanded, 1)
+                    sum_mask = torch.clamp(mask_expanded.sum(1), min=1e-9)
+                    batch_embeddings = sum_embeddings / sum_mask
+                pool_timer.print_elapsed()
+
+            return batch_embeddings.cpu().numpy()
+
+        # ZMQ server 主循环 - 修改为REP套接字
+        context = zmq.Context()
+        socket = context.socket(zmq.ROUTER)  # 改为REP套接字
+        socket.bind(f"tcp://127.0.0.1:{zmq_port}")
+        print(f"INFO: ZMQ ROUTER server listening on port {zmq_port}")
+
+        # 设置超时
+        socket.setsockopt(zmq.RCVTIMEO, 5000)  # 5秒接收超时
+        socket.setsockopt(zmq.SNDTIMEO, 300000)  # 300秒发送超时
+
+        from . import embedding_pb2
+
+        print(f"INFO: Embedding server ready to serve requests")
+
+        while True:
+            try:
+                parts = socket.recv_multipart()
+
+                # --- 恢复稳健的消息格式判断 ---
+                # 必须检查 parts 的长度，避免 IndexError
+                if len(parts) >= 3:
+                    identity = parts[0]
+                    # empty = parts[1]  # 中间的空帧我们通常不关心
+                    message = parts[2]
+                elif len(parts) == 2:
+                    # 也能处理没有空帧的情况
+                    identity = parts[0]
+                    message = parts[1]
+                else:
+                    # 如果收到格式错误的消息，打印警告并忽略它，而不是崩溃
+                    print(f"WARNING: Received unexpected message format with {len(parts)} parts. Ignoring.")
+                    continue
+                print(f"INFO: Received ZMQ request from client {identity.hex()[:8]}, size {len(message)} bytes")
+
+                e2e_start = time.time()
+                lookup_timer = DeviceTimer("text lookup", device)
+
+                # 解析请求
+                req_proto = embedding_pb2.NodeEmbeddingRequest()
+                req_proto.ParseFromString(message)
+                node_ids = req_proto.node_ids
+                print(f"INFO: Request for {len(node_ids)} node embeddings: {list(node_ids)}")
+
+                # 添加调试信息
+                if len(node_ids) > 0:
+                    print(f"DEBUG: Node ID range: {min(node_ids)} to {max(node_ids)}")
+                
+                # 查找文本
+                texts = []
+                missing_ids = []
+                with lookup_timer.timing():
+                    for nid in node_ids:
+                        txtinfo = passages[nid]
+                        txt = txtinfo["text"]
+                        if txt:
+                            texts.append(txt)
+                        else:
+                            # 如果文本为空，我们仍然需要一个占位符来进行批处理，
+                            # 但将其ID记录为缺失
+                            texts.append("") 
+                            missing_ids.append(nid)
+                lookup_timer.print_elapsed()
+
+                if missing_ids:
+                    print(f"WARNING: Missing passages for IDs: {missing_ids}")
+
+                # 处理批次
+                total_size = len(texts)
+                print(f"INFO: Total batch size: {total_size}, max_batch_size: {max_batch_size}")
+                
+                all_embeddings = []
+                
+                if total_size > max_batch_size:
+                    print(f"INFO: Splitting batch of size {total_size} into chunks of {max_batch_size}")
+                    for i in range(0, total_size, max_batch_size):
+                        end_idx = min(i + max_batch_size, total_size)
+                        print(f"INFO: Processing chunk {i//max_batch_size + 1}/{(total_size + max_batch_size - 1)//max_batch_size}: items {i} to {end_idx-1}")
+                        
+                        chunk_texts = texts[i:end_idx]
+                        chunk_ids = node_ids[i:end_idx]
+                        
+                        embeddings_chunk = process_batch(chunk_texts, chunk_ids, missing_ids)
+                        all_embeddings.append(embeddings_chunk)
+                        
+                        if cuda_available:
+                            torch.cuda.empty_cache()
+                        elif device.type == "mps":
+                            torch.mps.empty_cache()
+                            
+                    hidden = np.vstack(all_embeddings)
+                    print(f"INFO: Combined embeddings shape: {hidden.shape}")
+                else:
+                    hidden = process_batch(texts, node_ids, missing_ids)
+
+                # 序列化响应
+                ser_start = time.time()
+
+                resp_proto = embedding_pb2.NodeEmbeddingResponse()
+                hidden_contiguous = np.ascontiguousarray(hidden, dtype=np.float32)
+                resp_proto.embeddings_data = hidden_contiguous.tobytes()
+                resp_proto.dimensions.append(hidden_contiguous.shape[0])
+                resp_proto.dimensions.append(hidden_contiguous.shape[1])
+                resp_proto.missing_ids.extend(missing_ids)
+
+                response_data = resp_proto.SerializeToString()
+                
+                # REP 套接字发送单个响应
+                socket.send_multipart([identity, b'', response_data])
+
+                ser_end = time.time()
+
+                print(f"INFO: Serialize time: {ser_end - ser_start:.6f} seconds")
+
+                if device.type == "cuda":
+                    torch.cuda.synchronize()
+                elif device.type == "mps":
+                    torch.mps.synchronize()
+                e2e_end = time.time()
+                print(f"INFO: ZMQ E2E time: {e2e_end - e2e_start:.6f} seconds")
+
+            except zmq.Again:
+                print("INFO: ZMQ socket timeout, continuing to listen")
+                # REP套接字不需要重新创建，只需要继续监听
+                continue
+            except Exception as e:
+                print(f"ERROR: Error in ZMQ server: {e}")
+                try:
+                    # 发送空响应以维持REQ-REP状态
+                    empty_resp = embedding_pb2.NodeEmbeddingResponse()
+                    socket.send(empty_resp.SerializeToString())
+                except:
+                    # 如果发送失败，重新创建socket
+                    socket.close()
+                    socket = context.socket(zmq.REP)
+                    socket.bind(f"tcp://127.0.0.1:{zmq_port}")
+                    socket.setsockopt(zmq.RCVTIMEO, 5000)
+                    socket.setsockopt(zmq.SNDTIMEO, 300000)
+                    print("INFO: ZMQ socket recreated after error")
+
+    except Exception as e:
+        print(f"ERROR: Failed to start embedding server: {e}")
+        raise
+
+
+# 保持原有的 create_embedding_server 函数不变，只添加线程化版本
+def create_embedding_server(
+    domain="demo",
+    load_passages=True,
+    load_embeddings=False,
+    use_fp16=True,
+    use_int8=False,
+    use_cuda_graphs=False,
+    zmq_port=5555,
+    max_batch_size=128,
+    lazy_load_passages=False,
+    model_name="sentence-transformers/all-mpnet-base-v2",
+    passages_file: Optional[str] = None,
+):
+    """
+    原有的 create_embedding_server 函数保持不变
+    这个是阻塞版本，用于直接运行
+    """
+    create_embedding_server_thread(zmq_port, model_name, max_batch_size, passages_file)
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description="Embedding service")
+    parser.add_argument("--zmq-port", type=int, default=5555, help="ZMQ port to run on")
+    parser.add_argument("--domain", type=str, default="demo", help="Domain name")
+    parser.add_argument("--passages-file", type=str, help="JSON file containing passage ID to text mapping")
+    parser.add_argument("--load-passages", action="store_true", default=True)
+    parser.add_argument("--load-embeddings", action="store_true", default=False)
+    parser.add_argument("--use-fp16", action="store_true", default=False)
+    parser.add_argument("--use-int8", action="store_true", default=False)
+    parser.add_argument("--use-cuda-graphs", action="store_true", default=False)
+    parser.add_argument("--max-batch-size", type=int, default=128, help="Maximum batch size before splitting")
+    parser.add_argument("--lazy-load-passages", action="store_true", default=True)
+    parser.add_argument("--model-name", type=str, default="sentence-transformers/all-mpnet-base-v2", 
+                        help="Embedding model name")
+    args = parser.parse_args()
+
+    create_embedding_server(
+        domain=args.domain,
+        load_passages=args.load_passages,
+        load_embeddings=args.load_embeddings,
+        use_fp16=args.use_fp16,
+        use_int8=args.use_int8,
+        use_cuda_graphs=args.use_cuda_graphs,
+        zmq_port=args.zmq_port,
+        max_batch_size=args.max_batch_size,
+        lazy_load_passages=args.lazy_load_passages,
+        model_name=args.model_name,
+        passages_file=args.passages_file,
+    )

packages/leann-backend-diskann/leann_backend_diskann/graph_partition.py

@@ -1,299 +0,0 @@
-#!/usr/bin/env python3
-"""
-Graph Partition Module for LEANN DiskANN Backend
-
-This module provides Python bindings for the graph partition functionality
-of DiskANN, allowing users to partition disk-based indices for better
-performance.
-"""
-
-import os
-import shutil
-import subprocess
-import tempfile
-from pathlib import Path
-from typing import Optional
-
-
-class GraphPartitioner:
-    """
-    A Python interface for DiskANN's graph partition functionality.
-
-    This class provides methods to partition disk-based indices for improved
-    search performance and memory efficiency.
-    """
-
-    def __init__(self, build_type: str = "release"):
-        """
-        Initialize the GraphPartitioner.
-
-        Args:
-            build_type: Build type for the executables ("debug" or "release")
-        """
-        self.build_type = build_type
-        self._ensure_executables()
-
-    def _get_executable_path(self, name: str) -> str:
-        """Get the path to a graph partition executable."""
-        # Get the directory where this Python module is located
-        module_dir = Path(__file__).parent
-        # Navigate to the graph_partition directory
-        graph_partition_dir = module_dir.parent / "third_party" / "DiskANN" / "graph_partition"
-        executable_path = graph_partition_dir / "build" / self.build_type / "graph_partition" / name
-
-        if not executable_path.exists():
-            raise FileNotFoundError(f"Executable {name} not found at {executable_path}")
-
-        return str(executable_path)
-
-    def _ensure_executables(self):
-        """Ensure that the required executables are built."""
-        try:
-            self._get_executable_path("partitioner")
-            self._get_executable_path("index_relayout")
-        except FileNotFoundError:
-            # Try to build the executables automatically
-            print("Executables not found, attempting to build them...")
-            self._build_executables()
-
-    def _build_executables(self):
-        """Build the required executables."""
-        graph_partition_dir = (
-            Path(__file__).parent.parent / "third_party" / "DiskANN" / "graph_partition"
-        )
-        original_dir = os.getcwd()
-
-        try:
-            os.chdir(graph_partition_dir)
-
-            # Clean any existing build
-            if (graph_partition_dir / "build").exists():
-                shutil.rmtree(graph_partition_dir / "build")
-
-            # Run the build script
-            cmd = ["./build.sh", self.build_type, "split_graph", "/tmp/dummy"]
-            subprocess.run(cmd, capture_output=True, text=True, cwd=graph_partition_dir)
-
-            # Check if executables were created
-            partitioner_path = self._get_executable_path("partitioner")
-            relayout_path = self._get_executable_path("index_relayout")
-
-            print(f"✅ Built partitioner: {partitioner_path}")
-            print(f"✅ Built index_relayout: {relayout_path}")
-
-        except Exception as e:
-            raise RuntimeError(f"Failed to build executables: {e}")
-        finally:
-            os.chdir(original_dir)
-
-    def partition_graph(
-        self,
-        index_prefix_path: str,
-        output_dir: Optional[str] = None,
-        partition_prefix: Optional[str] = None,
-        **kwargs,
-    ) -> tuple[str, str]:
-        """
-        Partition a disk-based index for improved performance.
-
-        Args:
-            index_prefix_path: Path to the index prefix (e.g., "/path/to/index")
-            output_dir: Output directory for results (defaults to parent of index_prefix_path)
-            partition_prefix: Prefix for output files (defaults to basename of index_prefix_path)
-            **kwargs: Additional parameters for graph partitioning:
-                - gp_times: Number of LDG partition iterations (default: 10)
-                - lock_nums: Number of lock nodes (default: 10)
-                - cut: Cut adjacency list degree (default: 100)
-                - scale_factor: Scale factor (default: 1)
-                - data_type: Data type (default: "float")
-                - thread_nums: Number of threads (default: 10)
-
-        Returns:
-            Tuple of (disk_graph_index_path, partition_bin_path)
-
-        Raises:
-            RuntimeError: If the partitioning process fails
-        """
-        # Set default parameters
-        params = {
-            "gp_times": 10,
-            "lock_nums": 10,
-            "cut": 100,
-            "scale_factor": 1,
-            "data_type": "float",
-            "thread_nums": 10,
-            **kwargs,
-        }
-
-        # Determine output directory
-        if output_dir is None:
-            output_dir = str(Path(index_prefix_path).parent)
-
-        # Create output directory if it doesn't exist
-        Path(output_dir).mkdir(parents=True, exist_ok=True)
-
-        # Determine partition prefix
-        if partition_prefix is None:
-            partition_prefix = Path(index_prefix_path).name
-
-        # Get executable paths
-        partitioner_path = self._get_executable_path("partitioner")
-        relayout_path = self._get_executable_path("index_relayout")
-
-        # Create temporary directory for processing
-        with tempfile.TemporaryDirectory() as temp_dir:
-            # Change to the graph_partition directory for temporary files
-            graph_partition_dir = (
-                Path(__file__).parent.parent / "third_party" / "DiskANN" / "graph_partition"
-            )
-            original_dir = os.getcwd()
-
-            try:
-                os.chdir(graph_partition_dir)
-
-                # Create temporary data directory
-                temp_data_dir = Path(temp_dir) / "data"
-                temp_data_dir.mkdir(parents=True, exist_ok=True)
-
-                # Set up paths for temporary files
-                graph_path = temp_data_dir / "starling" / "_M_R_L_B" / "GRAPH"
-                graph_gp_path = (
-                    graph_path
-                    / f"GP_TIMES_{params['gp_times']}_LOCK_{params['lock_nums']}_GP_USE_FREQ0_CUT{params['cut']}_SCALE{params['scale_factor']}"
-                )
-                graph_gp_path.mkdir(parents=True, exist_ok=True)
-
-                # Find input index file
-                old_index_file = f"{index_prefix_path}_disk_beam_search.index"
-                if not os.path.exists(old_index_file):
-                    old_index_file = f"{index_prefix_path}_disk.index"
-
-                if not os.path.exists(old_index_file):
-                    raise RuntimeError(f"Index file not found: {old_index_file}")
-
-                # Run partitioner
-                gp_file_path = graph_gp_path / "_part.bin"
-                partitioner_cmd = [
-                    partitioner_path,
-                    "--index_file",
-                    old_index_file,
-                    "--data_type",
-                    params["data_type"],
-                    "--gp_file",
-                    str(gp_file_path),
-                    "-T",
-                    str(params["thread_nums"]),
-                    "--ldg_times",
-                    str(params["gp_times"]),
-                    "--scale",
-                    str(params["scale_factor"]),
-                    "--mode",
-                    "1",
-                ]
-
-                print(f"Running partitioner: {' '.join(partitioner_cmd)}")
-                result = subprocess.run(
-                    partitioner_cmd, capture_output=True, text=True, cwd=graph_partition_dir
-                )
-
-                if result.returncode != 0:
-                    raise RuntimeError(
-                        f"Partitioner failed with return code {result.returncode}.\n"
-                        f"stdout: {result.stdout}\n"
-                        f"stderr: {result.stderr}"
-                    )
-
-                # Run relayout
-                part_tmp_index = graph_gp_path / "_part_tmp.index"
-                relayout_cmd = [
-                    relayout_path,
-                    old_index_file,
-                    str(gp_file_path),
-                    params["data_type"],
-                    "1",
-                ]
-
-                print(f"Running relayout: {' '.join(relayout_cmd)}")
-                result = subprocess.run(
-                    relayout_cmd, capture_output=True, text=True, cwd=graph_partition_dir
-                )
-
-                if result.returncode != 0:
-                    raise RuntimeError(
-                        f"Relayout failed with return code {result.returncode}.\n"
-                        f"stdout: {result.stdout}\n"
-                        f"stderr: {result.stderr}"
-                    )
-
-                # Copy results to output directory
-                disk_graph_path = Path(output_dir) / f"{partition_prefix}_disk_graph.index"
-                partition_bin_path = Path(output_dir) / f"{partition_prefix}_partition.bin"
-
-                shutil.copy2(part_tmp_index, disk_graph_path)
-                shutil.copy2(gp_file_path, partition_bin_path)
-
-                print(f"Results copied to: {output_dir}")
-                return str(disk_graph_path), str(partition_bin_path)
-
-            finally:
-                os.chdir(original_dir)
-
-    def get_partition_info(self, partition_bin_path: str) -> dict:
-        """
-        Get information about a partition file.
-
-        Args:
-            partition_bin_path: Path to the partition binary file
-
-        Returns:
-            Dictionary containing partition information
-        """
-        if not os.path.exists(partition_bin_path):
-            raise FileNotFoundError(f"Partition file not found: {partition_bin_path}")
-
-        # For now, return basic file information
-        # In the future, this could parse the binary file for detailed info
-        stat = os.stat(partition_bin_path)
-        return {
-            "file_size": stat.st_size,
-            "file_path": partition_bin_path,
-            "modified_time": stat.st_mtime,
-        }
-
-
-def partition_graph(
-    index_prefix_path: str,
-    output_dir: Optional[str] = None,
-    partition_prefix: Optional[str] = None,
-    build_type: str = "release",
-    **kwargs,
-) -> tuple[str, str]:
-    """
-    Convenience function to partition a graph index.
-
-    Args:
-        index_prefix_path: Path to the index prefix
-        output_dir: Output directory (defaults to parent of index_prefix_path)
-        partition_prefix: Prefix for output files (defaults to basename of index_prefix_path)
-        build_type: Build type for executables ("debug" or "release")
-        **kwargs: Additional parameters for graph partitioning
-
-    Returns:
-        Tuple of (disk_graph_index_path, partition_bin_path)
-    """
-    partitioner = GraphPartitioner(build_type=build_type)
-    return partitioner.partition_graph(index_prefix_path, output_dir, partition_prefix, **kwargs)
-
-
-# Example usage:
-if __name__ == "__main__":
-    # Example: partition an index
-    try:
-        disk_graph_path, partition_bin_path = partition_graph(
-            "/path/to/your/index_prefix", gp_times=10, lock_nums=10, cut=100
-        )
-        print("Partitioning completed successfully!")
-        print(f"Disk graph index: {disk_graph_path}")
-        print(f"Partition binary: {partition_bin_path}")
-    except Exception as e:
-        print(f"Partitioning failed: {e}")

packages/leann-backend-diskann/pyproject.toml

@@ -1,21 +1,16 @@
 [build-system]
-requires = ["scikit-build-core>=0.10", "pybind11>=2.12.0", "numpy", "cmake>=3.30"]
+requires = ["scikit-build-core>=0.10", "pybind11>=2.12.0", "numpy"]
 build-backend = "scikit_build_core.build"
 
 [project]
 name = "leann-backend-diskann"
-version = "0.3.4"
+version = "0.1.0"
-dependencies = ["leann-core==0.3.4", "numpy", "protobuf>=3.19.0"]
+dependencies = ["leann-core==0.1.0", "numpy"]
 
 [tool.scikit-build]
-# Key: simplified CMake path
+# 关键：简化的 CMake 路径
 cmake.source-dir = "third_party/DiskANN"
-# Key: Python package in root directory, paths match exactly
+# 关键：Python 包在根目录，路径完全匹配
 wheel.packages = ["leann_backend_diskann"]
-# Use default redirect mode
+# 使用默认的 redirect 模式
-editable.mode = "redirect"
+editable.mode = "redirect"
-cmake.build-type = "Release"
-build.verbose = true
-build.tool-args = ["-j8"]
-# Let CMake find packages via Homebrew prefix
-cmake.define = {CMAKE_PREFIX_PATH = {env = "CMAKE_PREFIX_PATH"}, OpenMP_ROOT = {env = "OpenMP_ROOT"}}

packages/leann-backend-diskann/third_party/embedding.proto

@@ -2,12 +2,12 @@ syntax = "proto3";
 
 package protoembedding;
 
-message NodeEmbeddingRequest {
+message NodeEmbeddingRequest { 
-  repeated uint32 node_ids = 1;
+  repeated uint32 node_ids = 1; 
 }
 
 message NodeEmbeddingResponse {
   bytes embeddings_data = 1;        // All embedded binary datas
   repeated int32 dimensions = 2;    // Shape [batch_size, embedding_dim]
   repeated uint32 missing_ids = 3;  // Missing node ids
-}
+}

packages/leann-backend-hnsw/CMakeLists.txt

@@ -1,37 +1,25 @@
+# 最终简化版
 cmake_minimum_required(VERSION 3.24)
 project(leann_backend_hnsw_wrapper)
-set(CMAKE_C_COMPILER_WORKS 1)
-set(CMAKE_CXX_COMPILER_WORKS 1)
 
 # Set OpenMP path for macOS
 if(APPLE)
-    # Detect Homebrew installation path (Apple Silicon vs Intel)
+    set(OpenMP_C_FLAGS "-Xpreprocessor -fopenmp -I/opt/homebrew/opt/libomp/include")
-    if(EXISTS "/opt/homebrew/opt/libomp")
+    set(OpenMP_CXX_FLAGS "-Xpreprocessor -fopenmp -I/opt/homebrew/opt/libomp/include")
-        set(HOMEBREW_PREFIX "/opt/homebrew")
-    elseif(EXISTS "/usr/local/opt/libomp")
-        set(HOMEBREW_PREFIX "/usr/local")
-    else()
-        message(FATAL_ERROR "Could not find libomp installation. Please install with: brew install libomp")
-    endif()
-
-    set(OpenMP_C_FLAGS "-Xpreprocessor -fopenmp -I${HOMEBREW_PREFIX}/opt/libomp/include")
-    set(OpenMP_CXX_FLAGS "-Xpreprocessor -fopenmp -I${HOMEBREW_PREFIX}/opt/libomp/include")
     set(OpenMP_C_LIB_NAMES "omp")
     set(OpenMP_CXX_LIB_NAMES "omp")
-    set(OpenMP_omp_LIBRARY "${HOMEBREW_PREFIX}/opt/libomp/lib/libomp.dylib")
+    set(OpenMP_omp_LIBRARY "/opt/homebrew/opt/libomp/lib/libomp.dylib")
-
-    # Force use of system libc++ to avoid version mismatch
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -stdlib=libc++")
-    set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -stdlib=libc++")
-    set(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} -stdlib=libc++")
-
-    # Set minimum macOS version for better compatibility
-    set(CMAKE_OSX_DEPLOYMENT_TARGET "11.0" CACHE STRING "Minimum macOS version")
 endif()
 
-# Use system ZeroMQ instead of building from source
+# Build ZeroMQ from source
-find_package(PkgConfig REQUIRED)
+set(ZMQ_BUILD_TESTS OFF CACHE BOOL "" FORCE)
-pkg_check_modules(ZMQ REQUIRED libzmq)
+set(ENABLE_DRAFTS OFF CACHE BOOL "" FORCE)
+set(ENABLE_PRECOMPILED OFF CACHE BOOL "" FORCE)
+set(WITH_PERF_TOOL OFF CACHE BOOL "" FORCE)
+set(WITH_DOCS OFF CACHE BOOL "" FORCE)
+set(BUILD_SHARED OFF CACHE BOOL "" FORCE)
+set(BUILD_STATIC ON CACHE BOOL "" FORCE)
+add_subdirectory(third_party/libzmq)
 
 # Add cppzmq headers
 include_directories(third_party/cppzmq)
@@ -41,7 +29,6 @@ set(MSGPACK_USE_BOOST OFF CACHE BOOL "" FORCE)
 add_compile_definitions(MSGPACK_NO_BOOST)
 include_directories(third_party/msgpack-c/include)
 
-# Faiss configuration - streamlined build
 set(FAISS_ENABLE_PYTHON ON CACHE BOOL "" FORCE)
 set(FAISS_ENABLE_GPU OFF CACHE BOOL "" FORCE)
 set(FAISS_ENABLE_EXTRAS OFF CACHE BOOL "" FORCE)
@@ -49,43 +36,4 @@ set(BUILD_TESTING OFF CACHE BOOL "" FORCE)
 set(FAISS_ENABLE_C_API OFF CACHE BOOL "" FORCE)
 set(FAISS_OPT_LEVEL "generic" CACHE STRING "" FORCE)
 
-# Disable x86-specific SIMD optimizations (important for ARM64 compatibility)
+add_subdirectory(third_party/faiss)
-set(FAISS_ENABLE_AVX2 OFF CACHE BOOL "" FORCE)
-set(FAISS_ENABLE_AVX512 OFF CACHE BOOL "" FORCE)
-set(FAISS_ENABLE_SSE4_1 OFF CACHE BOOL "" FORCE)
-
-# ARM64-specific configuration
-if(CMAKE_SYSTEM_PROCESSOR MATCHES "aarch64|arm64")
-    message(STATUS "Configuring Faiss for ARM64 architecture")
-
-    if(CMAKE_SYSTEM_NAME STREQUAL "Linux")
-        # Use SVE optimization level for ARM64 Linux (as seen in Faiss conda build)
-        set(FAISS_OPT_LEVEL "sve" CACHE STRING "" FORCE)
-        message(STATUS "Setting FAISS_OPT_LEVEL to 'sve' for ARM64 Linux")
-    else()
-        # Use generic optimization for other ARM64 platforms (like macOS)
-        set(FAISS_OPT_LEVEL "generic" CACHE STRING "" FORCE)
-        message(STATUS "Setting FAISS_OPT_LEVEL to 'generic' for ARM64 ${CMAKE_SYSTEM_NAME}")
-    endif()
-
-    # ARM64 compatibility: Faiss submodule has been modified to fix x86 header inclusion
-    message(STATUS "Using ARM64-compatible Faiss submodule")
-endif()
-
-# Additional optimization options from INSTALL.md
-set(CMAKE_BUILD_TYPE "Release" CACHE STRING "" FORCE)
-set(BUILD_SHARED_LIBS OFF CACHE BOOL "" FORCE)  # Static library is faster to build
-
-# Avoid building demos and benchmarks
-set(BUILD_DEMOS OFF CACHE BOOL "" FORCE)
-set(BUILD_BENCHS OFF CACHE BOOL "" FORCE)
-
-# NEW: Tell Faiss to only build the generic version
-set(FAISS_BUILD_GENERIC ON CACHE BOOL "" FORCE)
-set(FAISS_BUILD_AVX2 OFF CACHE BOOL "" FORCE)
-set(FAISS_BUILD_AVX512 OFF CACHE BOOL "" FORCE)
-
-# IMPORTANT: Disable building AVX versions to speed up compilation
-set(FAISS_BUILD_AVX_VERSIONS OFF CACHE BOOL "" FORCE)
-
-add_subdirectory(third_party/faiss)

packages/leann-backend-hnsw/leann_backend_hnsw/__init__.py

@@ -1 +1 @@
-from . import hnsw_backend as hnsw_backend
+from . import hnsw_backend

packages/leann-backend-hnsw/leann_backend_hnsw/hnsw_backend.py

@@ -1,40 +1,29 @@
-import logging
-import os
-import shutil
-import time
-from pathlib import Path
-from typing import Any, Literal, Optional
-
 import numpy as np
-from leann.interface import (
+import os
-    LeannBackendBuilderInterface,
+import json
-    LeannBackendFactoryInterface,
+from pathlib import Path
-    LeannBackendSearcherInterface,
+from typing import Dict, Any, List
-)
+import pickle
-from leann.registry import register_backend
+import shutil
+
 from leann.searcher_base import BaseSearcher
+from .convert_to_csr import convert_hnsw_graph_to_csr
 
-from .convert_to_csr import convert_hnsw_graph_to_csr, prune_hnsw_embeddings_inplace
+from leann.registry import register_backend
-
+from leann.interface import (
-logger = logging.getLogger(__name__)
+    LeannBackendFactoryInterface,
-
+    LeannBackendBuilderInterface,
+    LeannBackendSearcherInterface
+)
 
 def get_metric_map():
-    from . import faiss  # type: ignore
+    from . import faiss
-
     return {
         "mips": faiss.METRIC_INNER_PRODUCT,
         "l2": faiss.METRIC_L2,
         "cosine": faiss.METRIC_INNER_PRODUCT,
     }
 
-
-def normalize_l2(data: np.ndarray) -> np.ndarray:
-    norms = np.linalg.norm(data, axis=1, keepdims=True)
-    norms[norms == 0] = 1  # Avoid division by zero
-    return data / norms
-
-
 @register_backend("hnsw")
 class HNSWBackend(LeannBackendFactoryInterface):
     @staticmethod
@@ -45,7 +34,6 @@ class HNSWBackend(LeannBackendFactoryInterface):
     def searcher(index_path: str, **kwargs) -> LeannBackendSearcherInterface:
         return HNSWSearcher(index_path, **kwargs)
 
-
 class HNSWBuilder(LeannBackendBuilderInterface):
     def __init__(self, **kwargs):
         self.build_params = kwargs.copy()
@@ -55,26 +43,22 @@ class HNSWBuilder(LeannBackendBuilderInterface):
         self.efConstruction = self.build_params.setdefault("efConstruction", 200)
         self.distance_metric = self.build_params.setdefault("distance_metric", "mips")
         self.dimensions = self.build_params.get("dimensions")
-        if not self.is_recompute and self.is_compact:
-            # Auto-correct: non-recompute requires non-compact storage for HNSW
-            logger.warning(
-                "is_recompute=False requires non-compact HNSW. Forcing is_compact=False."
-            )
-            self.is_compact = False
-            self.build_params["is_compact"] = False
-
-    def build(self, data: np.ndarray, ids: list[str], index_path: str, **kwargs):
-        from . import faiss  # type: ignore
 
+    def build(self, data: np.ndarray, ids: List[str], index_path: str, **kwargs):
+        from . import faiss
         path = Path(index_path)
         index_dir = path.parent
         index_prefix = path.stem
         index_dir.mkdir(parents=True, exist_ok=True)
 
         if data.dtype != np.float32:
-            logger.warning(f"Converting data to float32, shape: {data.shape}")
             data = data.astype(np.float32)
 
+        label_map = {i: str_id for i, str_id in enumerate(ids)}
+        label_map_file = index_dir / "leann.labels.map"
+        with open(label_map_file, 'wb') as f:
+            pickle.dump(label_map, f)
+
         metric_enum = get_metric_map().get(self.distance_metric.lower())
         if metric_enum is None:
             raise ValueError(f"Unsupported distance_metric '{self.distance_metric}'.")
@@ -84,7 +68,7 @@ class HNSWBuilder(LeannBackendBuilderInterface):
         index.hnsw.efConstruction = self.efConstruction
 
         if self.distance_metric.lower() == "cosine":
-            data = normalize_l2(data)
+            faiss.normalize_L2(data)
 
         index.add(data.shape[0], faiss.swig_ptr(data))
         index_file = index_dir / f"{index_prefix}.index"
@@ -92,53 +76,46 @@ class HNSWBuilder(LeannBackendBuilderInterface):
 
         if self.is_compact:
             self._convert_to_csr(index_file)
-        elif self.is_recompute:
-            prune_hnsw_embeddings_inplace(str(index_file))
 
     def _convert_to_csr(self, index_file: Path):
         """Convert built index to CSR format"""
         mode_str = "CSR-pruned" if self.is_recompute else "CSR-standard"
-        logger.info(f"INFO: Converting HNSW index to {mode_str} format...")
+        print(f"INFO: Converting HNSW index to {mode_str} format...")
 
         csr_temp_file = index_file.with_suffix(".csr.tmp")
 
         success = convert_hnsw_graph_to_csr(
-            str(index_file), str(csr_temp_file), prune_embeddings=self.is_recompute
+            str(index_file), 
+            str(csr_temp_file),
+            prune_embeddings=self.is_recompute
         )
 
         if success:
-            logger.info("✅ CSR conversion successful.")
+            print("✅ CSR conversion successful.")
-            # index_file_old = index_file.with_suffix(".old")
+            index_file_old = index_file.with_suffix(".old")
-            # shutil.move(str(index_file), str(index_file_old))
+            shutil.move(str(index_file), str(index_file_old))
             shutil.move(str(csr_temp_file), str(index_file))
-            logger.info(f"INFO: Replaced original index with {mode_str} version at '{index_file}'")
+            print(f"INFO: Replaced original index with {mode_str} version at '{index_file}'")
         else:
             # Clean up and fail fast
             if csr_temp_file.exists():
                 os.remove(csr_temp_file)
             raise RuntimeError("CSR conversion failed - cannot proceed with compact format")
 
-
 class HNSWSearcher(BaseSearcher):
     def __init__(self, index_path: str, **kwargs):
-        super().__init__(
+        super().__init__(index_path, backend_module_name="leann_backend_hnsw.hnsw_embedding_server", **kwargs)
-            index_path,
+        from . import faiss
-            backend_module_name="leann_backend_hnsw.hnsw_embedding_server",
-            **kwargs,
-        )
-        from . import faiss  # type: ignore
 
-        self.distance_metric = (
+        self.distance_metric = self.meta.get("distance_metric", "mips").lower()
-            self.meta.get("backend_kwargs", {}).get("distance_metric", "mips").lower()
-        )
         metric_enum = get_metric_map().get(self.distance_metric)
         if metric_enum is None:
             raise ValueError(f"Unsupported distance_metric '{self.distance_metric}'.")
 
-        backend_meta_kwargs = self.meta.get("backend_kwargs", {})
+        self.is_compact, self.is_pruned = (
-        self.is_compact = self.meta.get("is_compact", backend_meta_kwargs.get("is_compact", True))
+            self.meta.get('is_compact', True),
-        default_pruned = backend_meta_kwargs.get("is_recompute", self.is_compact)
+            self.meta.get('is_pruned', True)
-        self.is_pruned = bool(self.meta.get("is_pruned", default_pruned))
+        )
 
         index_file = self.index_dir / f"{self.index_path.stem}.index"
         if not index_file.exists():
@@ -146,110 +123,39 @@ class HNSWSearcher(BaseSearcher):
 
         hnsw_config = faiss.HNSWIndexConfig()
         hnsw_config.is_compact = self.is_compact
-        hnsw_config.is_recompute = (
+        hnsw_config.is_recompute = self.is_pruned or kwargs.get("is_recompute", False)
-            self.is_pruned
+
-        )  # In C++ code, it's called is_recompute, but it's only for loading IIUC.
+        if self.is_pruned and not hnsw_config.is_recompute:
+            raise RuntimeError("Index is pruned but recompute is disabled.")
 
         self._index = faiss.read_index(str(index_file), faiss.IO_FLAG_MMAP, hnsw_config)
 
-    def search(
+    def search(self, query: np.ndarray, top_k: int, **kwargs) -> Dict[str, Any]:
-        self,
+        from . import faiss
-        query: np.ndarray,
-        top_k: int,
-        zmq_port: Optional[int] = None,
-        complexity: int = 64,
-        beam_width: int = 1,
-        prune_ratio: float = 0.0,
-        recompute_embeddings: bool = True,
-        pruning_strategy: Literal["global", "local", "proportional"] = "global",
-        batch_size: int = 0,
-        **kwargs,
-    ) -> dict[str, Any]:
-        """
-        Search for nearest neighbors using HNSW index.
 
-        Args:
+        if self.is_pruned:
-            query: Query vectors (B, D) where B is batch size, D is dimension
+            meta_file_path = self.index_dir / f"{self.index_path.name}.meta.json"
-            top_k: Number of nearest neighbors to return
+            if not meta_file_path.exists():
-            complexity: Search complexity/efSearch, higher = more accurate but slower
+                raise RuntimeError(f"FATAL: Index is pruned but metadata file not found: {meta_file_path}")
-            beam_width: Number of parallel search paths/beam_size
+            zmq_port = kwargs.get("zmq_port", 5557)
-            prune_ratio: Ratio of neighbors to prune via PQ (0.0-1.0)
+            self._ensure_server_running(str(meta_file_path), port=zmq_port, **kwargs)
-            recompute_embeddings: Whether to fetch fresh embeddings from server
-            pruning_strategy: PQ candidate selection strategy:
-                - "global": Use global PQ queue size for selection (default)
-                - "local": Local pruning, sort and select best candidates
-                - "proportional": Base selection on new neighbor count ratio
-            zmq_port: ZMQ port for embedding server communication. Must be provided if recompute_embeddings is True.
-            batch_size: Neighbor processing batch size, 0=disabled (HNSW-specific)
-            **kwargs: Additional HNSW-specific parameters (for legacy compatibility)
-
-        Returns:
-            Dict with 'labels' (list of lists) and 'distances' (ndarray)
-        """
-        from . import faiss  # type: ignore
-
-        if not recompute_embeddings and self.is_pruned:
-            raise RuntimeError(
-                "Recompute is required for pruned/compact HNSW index. "
-                "Re-run search with --recompute, or rebuild with --no-recompute and --no-compact."
-            )
-        if recompute_embeddings:
-            if zmq_port is None:
-                raise ValueError("zmq_port must be provided if recompute_embeddings is True")
 
         if query.dtype != np.float32:
             query = query.astype(np.float32)
         if self.distance_metric == "cosine":
-            query = normalize_l2(query)
+            faiss.normalize_L2(query)
 
         params = faiss.SearchParametersHNSW()
-        if zmq_port is not None:
+        params.zmq_port = kwargs.get("zmq_port", 5557)
-            params.zmq_port = zmq_port  # C++ code won't use this if recompute_embeddings is False
+        params.efSearch = kwargs.get("complexity", 32)
-        params.efSearch = complexity
+        params.beam_size = kwargs.get("beam_width", 1)
-        params.beam_size = beam_width
 
-        # For OpenAI embeddings with cosine distance, disable relative distance check
+        batch_size = query.shape[0]
-        # This prevents early termination when all scores are in a narrow range
+        distances = np.empty((batch_size, top_k), dtype=np.float32)
-        embedding_model = self.meta.get("embedding_model", "").lower()
+        labels = np.empty((batch_size, top_k), dtype=np.int64)
-        if self.distance_metric == "cosine" and any(
-            openai_model in embedding_model for openai_model in ["text-embedding", "openai"]
-        ):
-            params.check_relative_distance = False
-        else:
-            params.check_relative_distance = True
 
-        # PQ pruning: direct mapping to HNSW's pq_pruning_ratio
+        self._index.search(query.shape[0], faiss.swig_ptr(query), top_k, faiss.swig_ptr(distances), faiss.swig_ptr(labels), params)
-        params.pq_pruning_ratio = prune_ratio
 
-        # Map pruning_strategy to HNSW parameters
+        string_labels = [[self.label_map.get(int_label, f"unknown_{int_label}") for int_label in batch_labels] for batch_labels in labels]
-        if pruning_strategy == "local":
-            params.local_prune = True
-            params.send_neigh_times_ratio = 0.0
-        elif pruning_strategy == "proportional":
-            params.local_prune = False
-            params.send_neigh_times_ratio = 1.0  # Any value > 1e-6 triggers proportional mode
-        else:  # "global"
-            params.local_prune = False
-            params.send_neigh_times_ratio = 0.0
 
-        # HNSW-specific batch processing parameter
+        return {"labels": string_labels, "distances": distances}
-        params.batch_size = batch_size
-
-        batch_size_query = query.shape[0]
-        distances = np.empty((batch_size_query, top_k), dtype=np.float32)
-        labels = np.empty((batch_size_query, top_k), dtype=np.int64)
-
-        search_time = time.time()
-        self._index.search(
-            query.shape[0],
-            faiss.swig_ptr(query),
-            top_k,
-            faiss.swig_ptr(distances),
-            faiss.swig_ptr(labels),
-            params,
-        )
-        search_time = time.time() - search_time
-        logger.info(f"  Search time in HNSWSearcher.search() backend: {search_time} seconds")
-        string_labels = [[str(int_label) for int_label in batch_labels] for batch_labels in labels]
-
-        return {"labels": string_labels, "distances": distances}

packages/leann-backend-hnsw/pyproject.toml

@@ -6,24 +6,12 @@ build-backend = "scikit_build_core.build"
 
 [project]
 name = "leann-backend-hnsw"
-version = "0.3.4"
+version = "0.1.0"
 description = "Custom-built HNSW (Faiss) backend for the Leann toolkit."
-dependencies = [
+dependencies = ["leann-core==0.1.0", "numpy"]
-    "leann-core==0.3.4",
-    "numpy",
-    "pyzmq>=23.0.0",
-    "msgpack>=1.0.0",
-]
 
 [tool.scikit-build]
 wheel.packages = ["leann_backend_hnsw"]
 editable.mode = "redirect"
-cmake.build-type = "Release"
+cmake.build-type = "Debug"
-build.verbose = true
+build.verbose = true
-build.tool-args = ["-j8"]
-
-# CMake definitions to optimize compilation and find Homebrew packages
-[tool.scikit-build.cmake.define]
-CMAKE_BUILD_PARALLEL_LEVEL = "8"
-CMAKE_PREFIX_PATH = {env = "CMAKE_PREFIX_PATH"}
-OpenMP_ROOT = {env = "OpenMP_ROOT"}