Merge branch 'main' into arch-eval

* feat(core): Add AST-aware code chunking with astchunk integration

This PR introduces intelligent code chunking that preserves semantic boundaries
(functions, classes, methods) for better code understanding in RAG applications.

Key Features:
- AST-aware chunking for Python, Java, C#, TypeScript files
- Graceful fallback to traditional chunking for unsupported languages
- New specialized code RAG application for repositories
- Enhanced CLI with --use-ast-chunking flag
- Comprehensive test suite with integration tests

Technical Implementation:
- New chunking_utils.py module with enhanced chunking logic
- Extended base RAG framework with AST chunking arguments
- Updated document RAG with --enable-code-chunking flag
- CLI integration with proper error handling and fallback

Benefits:
- Better semantic understanding of code structure
- Improved search quality for code-related queries
- Maintains backward compatibility with existing workflows
- Supports mixed content (code + documentation) seamlessly

Dependencies:
- Added astchunk and tree-sitter parsers to pyproject.toml
- All dependencies are optional - fallback works without them

Testing:
- Comprehensive test suite in test_astchunk_integration.py
- Integration tests with document RAG
- Error handling and edge case coverage

Documentation:
- Updated README.md with AST chunking highlights
- Added ASTCHUNK_INTEGRATION.md with complete guide
- Updated features.md with new capabilities

* Refactored chunk utils

* Remove useless import

* Update README.md

* Update apps/chunking/utils.py

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Update apps/code_rag.py

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Fix issue

* apply suggestion from @Copilot

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Fixes after pr review

* Fix tests not passing

* Fix linter error for documentation files

* Update .gitignore with unwanted files

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Andy Lee <andylizf@outlook.com>

.gitattributes

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

.github/workflows/build-reusable.yml

@@ -87,7 +87,7 @@ jobs:
     runs-on: ${{ matrix.os }}
 
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v5
         with:
           ref: ${{ inputs.ref }}
           submodules: recursive
@@ -98,21 +98,23 @@ jobs:
           python-version: ${{ matrix.python }}
 
       - name: Install uv
-        uses: astral-sh/setup-uv@v4
+        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 libopenblas-dev patchelf libabsl-dev libaio-dev libprotobuf-dev
+            pkg-config libabsl-dev libaio-dev libprotobuf-dev \
+            patchelf
 
           # Install Intel MKL for DiskANN
           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" >> $GITHUB_ENV
-          echo "LD_LIBRARY_PATH=/opt/intel/oneapi/mkl/latest/lib/intel64:$LD_LIBRARY_PATH" >> $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
 
       - name: Install system dependencies (macOS)
         if: runner.os == 'macOS'
@@ -304,3 +306,53 @@ jobs:
         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

@@ -14,6 +14,6 @@ jobs:
       - uses: actions/checkout@v4
       - uses: lycheeverse/lychee-action@v2
         with:
-          args: --no-progress --insecure README.md docs/ apps/ examples/ benchmarks/
+          args: --no-progress --insecure --user-agent 'curl/7.68.0' README.md docs/ apps/ examples/ benchmarks/
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.gitignore

@@ -18,6 +18,7 @@ demo/experiment_results/**/*.json
 *.eml
 *.emlx
 *.json
+!.vscode/*.json
 *.sh
 *.txt
 !CMakeLists.txt
@@ -92,3 +93,5 @@ packages/leann-backend-diskann/third_party/DiskANN/_deps/
 batchtest.py
 tests/__pytest_cache__/
 tests/__pycache__/
+
+benchmarks/data/

.vscode/extensions.json

@@ -0,0 +1,5 @@
+{
+    "recommendations": [
+        "charliermarsh.ruff",
+    ]
+}

.vscode/settings.json

@@ -0,0 +1,22 @@
+{
+  "python.defaultInterpreterPath": ".venv/bin/python",
+  "python.terminal.activateEnvironment": true,
+  "[python]": {
+    "editor.defaultFormatter": "charliermarsh.ruff",
+    "editor.formatOnSave": true,
+    "editor.codeActionsOnSave": {
+      "source.organizeImports": "explicit",
+      "source.fixAll": "explicit"
+    },
+    "editor.insertSpaces": true,
+    "editor.tabSize": 4
+  },
+  "ruff.enable": true,
+  "files.watcherExclude": {
+    "**/.venv/**": true,
+    "**/__pycache__/**": true,
+    "**/*.egg-info/**": true,
+    "**/build/**": true,
+    "**/dist/**": true
+  }
+}

README.md

@@ -5,7 +5,7 @@
 <p align="center">
   <img src="https://img.shields.io/badge/Python-3.9%20%7C%203.10%20%7C%203.11%20%7C%203.12%20%7C%203.13-blue.svg" alt="Python Versions">
   <img src="https://github.com/yichuan-w/LEANN/actions/workflows/build-and-publish.yml/badge.svg" alt="CI Status">
-  <img src="https://img.shields.io/badge/Platform-Ubuntu%20%7C%20macOS%20(ARM64%2FIntel)-lightgrey" alt="Platform">
+  <img src="https://img.shields.io/badge/Platform-Ubuntu%20%26%20Arch%20%26%20WSL%20%7C%20macOS%20(ARM64%2FIntel)-lightgrey" alt="Platform">
   <img src="https://img.shields.io/badge/License-MIT-green.svg" alt="MIT License">
   <img src="https://img.shields.io/badge/MCP-Native%20Integration-blue" alt="MCP Integration">
 </p>
@@ -31,7 +31,7 @@ LEANN achieves this through *graph-based selective recomputation* with *high-deg
   <img src="assets/effects.png" alt="LEANN vs Traditional Vector DB Storage Comparison" width="70%">
 </p>
 
-> **The numbers speak for themselves:** Index 60 million text chunks in just 6GB instead of 201GB. From emails to browser history, everything fits on your laptop. [See detailed benchmarks for different applications below ↓](#storage-comparison)
+> **The numbers speak for themselves:** Index 60 million text chunks in just 6GB instead of 201GB. From emails to browser history, everything fits on your laptop. [See detailed benchmarks for different applications below ↓](#-storage-comparison)
 
 
 🔒 **Privacy:** Your data never leaves your laptop. No OpenAI, no cloud, no "terms of service".
@@ -70,6 +70,8 @@ uv venv
 source .venv/bin/activate
 uv pip install leann
 ```
+<!--
+> Low-resource? See “Low-resource setups” in the [Configuration Guide](docs/configuration-guide.md#low-resource-setups). -->
 
 <details>
 <summary>
@@ -85,15 +87,60 @@ git submodule update --init --recursive
 ```
 
 **macOS:**
+
+Note: DiskANN requires MacOS 13.3 or later.
+
 ```bash
-brew install llvm libomp boost protobuf zeromq pkgconf
-CC=$(brew --prefix llvm)/bin/clang CXX=$(brew --prefix llvm)/bin/clang++ uv sync
+brew install libomp boost protobuf zeromq pkgconf
+uv sync --extra diskann
 ```
 
-**Linux:**
+**Linux (Ubuntu/Debian):**
+
+Note: On Ubuntu 20.04, you may need to build a newer Abseil and pin Protobuf (e.g., v3.20.x) for building DiskANN. See [Issue #30](https://github.com/yichuan-w/LEANN/issues/30) for a step-by-step note.
+
+You can manually install [Intel oneAPI MKL](https://www.intel.com/content/www/us/en/developer/tools/oneapi/onemkl.html) instead of `libmkl-full-dev` for DiskANN. You can also use `libopenblas-dev` for building HNSW only, by removing `--extra diskann` in the command below.
+
 ```bash
-sudo apt-get install libomp-dev libboost-all-dev protobuf-compiler libabsl-dev libmkl-full-dev libaio-dev libzmq3-dev
-uv sync
+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 \
+  libmkl-full-dev
+
+uv sync --extra diskann
+```
+
+**Linux (Arch Linux):**
+
+```bash
+sudo pacman -Syu && sudo pacman -S --needed base-devel cmake pkgconf git gcc \
+  boost boost-libs protobuf abseil-cpp libaio zeromq
+
+# For MKL in DiskANN
+sudo pacman -S --needed base-devel git
+git clone https://aur.archlinux.org/paru-bin.git
+cd paru-bin && makepkg -si
+paru -S intel-oneapi-mkl intel-oneapi-compiler
+source /opt/intel/oneapi/setvars.sh
+
+uv sync --extra diskann
+```
+
+**Linux (RHEL / CentOS Stream / Oracle / Rocky / AlmaLinux):**
+
+See [Issue #50](https://github.com/yichuan-w/LEANN/issues/50) for more details.
+
+```bash
+sudo dnf groupinstall -y "Development Tools"
+sudo dnf install -y libomp-devel boost-devel protobuf-compiler protobuf-devel \
+  abseil-cpp-devel libaio-devel zeromq-devel pkgconf-pkg-config
+
+# For MKL in DiskANN
+sudo dnf install -y intel-oneapi-mkl intel-oneapi-mkl-devel \
+  intel-oneapi-openmp || sudo dnf install -y intel-oneapi-compiler
+source /opt/intel/oneapi/setvars.sh
+
+uv sync --extra diskann
 ```
 
 </details>
@@ -129,6 +176,9 @@ response = chat.ask("How much storage does LEANN save?", top_k=1)
 
 LEANN supports RAG on various data sources including documents (`.pdf`, `.txt`, `.md`), Apple Mail, Google Search History, WeChat, and more.
 
+**AST-Aware Code Chunking** - LEANN also features intelligent code chunking that preserves semantic boundaries (functions, classes, methods) for Python, Java, C#, and TypeScript files, providing improved code understanding compared to traditional text-based approaches.
+📖 Read the [AST Chunking Guide →](docs/ast_chunking_guide.md) to learn more.
+
 ### Generation Model Setup
 
 LEANN supports multiple LLM providers for text generation (OpenAI API, HuggingFace, Ollama).
@@ -184,34 +234,34 @@ All RAG examples share these common parameters. **Interactive mode** is availabl
 
 ```bash
 # Core Parameters (General preprocessing for all examples)
---index-dir DIR          # Directory to store the index (default: current directory)
---query "YOUR QUESTION"  # Single query mode. Omit for interactive chat (type 'quit' to exit), and now you can play with your index interactively
---max-items N           # Limit data preprocessing (default: -1, process all data)
---force-rebuild         # Force rebuild index even if it exists
+--index-dir DIR              # Directory to store the index (default: current directory)
+--query "YOUR QUESTION"      # Single query mode. Omit for interactive chat (type 'quit' to exit), and now you can play with your index interactively
+--max-items N                # Limit data preprocessing (default: -1, process all data)
+--force-rebuild              # Force rebuild index even if it exists
 
 # Embedding Parameters
---embedding-model MODEL  # e.g., facebook/contriever, text-embedding-3-small, mlx-community/Qwen3-Embedding-0.6B-8bit or nomic-embed-text
---embedding-mode MODE    # sentence-transformers, openai, mlx, or ollama
+--embedding-model MODEL      # e.g., facebook/contriever, text-embedding-3-small, mlx-community/Qwen3-Embedding-0.6B-8bit or nomic-embed-text
+--embedding-mode MODE        # sentence-transformers, openai, mlx, or ollama
 
 # LLM Parameters (Text generation models)
---llm TYPE              # LLM backend: openai, ollama, or hf (default: openai)
---llm-model MODEL       # Model name (default: gpt-4o) e.g., gpt-4o-mini, llama3.2:1b, Qwen/Qwen2.5-1.5B-Instruct
---thinking-budget LEVEL # Thinking budget for reasoning models: low/medium/high (supported by o3, o3-mini, GPT-Oss:20b, and other reasoning models)
+--llm TYPE                   # LLM backend: openai, ollama, or hf (default: openai)
+--llm-model MODEL            # Model name (default: gpt-4o) e.g., gpt-4o-mini, llama3.2:1b, Qwen/Qwen2.5-1.5B-Instruct
+--thinking-budget LEVEL      # Thinking budget for reasoning models: low/medium/high (supported by o3, o3-mini, GPT-Oss:20b, and other reasoning models)
 
 # Search Parameters
---top-k N               # Number of results to retrieve (default: 20)
---search-complexity N   # Search complexity for graph traversal (default: 32)
+--top-k N                    # Number of results to retrieve (default: 20)
+--search-complexity N        # Search complexity for graph traversal (default: 32)
 
 # Chunking Parameters
---chunk-size N          # Size of text chunks (default varies by source: 256 for most, 192 for WeChat)
---chunk-overlap N       # Overlap between chunks (default varies: 25-128 depending on source)
+--chunk-size N               # Size of text chunks (default varies by source: 256 for most, 192 for WeChat)
+--chunk-overlap N            # Overlap between chunks (default varies: 25-128 depending on source)
 
 # Index Building Parameters
---backend-name NAME     # Backend to use: hnsw or diskann (default: hnsw)
---graph-degree N        # Graph degree for index construction (default: 32)
---build-complexity N    # Build complexity for index construction (default: 64)
---no-compact           # Disable compact index storage (compact storage IS enabled to save storage by default)
---no-recompute         # Disable embedding recomputation (recomputation IS enabled to save storage by default)
+--backend-name NAME          # Backend to use: hnsw or diskann (default: hnsw)
+--graph-degree N             # Graph degree for index construction (default: 32)
+--build-complexity N         # Build complexity for index construction (default: 64)
+--compact / --no-compact     # Use compact storage (default: true). Must be `no-compact` for `no-recompute` build.
+--recompute / --no-recompute # Enable/disable embedding recomputation (default: enabled). Should not do a `no-recompute` search in a `recompute` build.
 ```
 
 </details>
@@ -247,6 +297,12 @@ python -m apps.document_rag --data-dir "~/Documents/Papers" --chunk-size 1024
 
 # Filter only markdown and Python files with smaller chunks
 python -m apps.document_rag --data-dir "./docs" --chunk-size 256 --file-types .md .py
+
+# Enable AST-aware chunking for code files
+python -m apps.document_rag --enable-code-chunking --data-dir "./my_project"
+
+# Or use the specialized code RAG for better code understanding
+python -m apps.code_rag --repo-dir "./my_codebase" --query "How does authentication work?"
 ```
 
 </details>
@@ -424,21 +480,22 @@ Once the index is built, you can ask questions like:
 **The future of code assistance is here.** Transform your development workflow with LEANN's native MCP integration for Claude Code. Index your entire codebase and get intelligent code assistance directly in your IDE.
 
 **Key features:**
-- 🔍 **Semantic code search** across your entire project
+- 🔍 **Semantic code search** across your entire project, fully local index and lightweight
+- 🧠 **AST-aware chunking** preserves code structure (functions, classes)
 - 📚 **Context-aware assistance** for debugging and development
 - 🚀 **Zero-config setup** with automatic language detection
 
 ```bash
 # Install LEANN globally for MCP integration
-uv tool install leann-core
-
+uv tool install leann-core --with leann
+claude mcp add --scope user leann-server -- leann_mcp
 # Setup is automatic - just start using Claude Code!
 ```
 Try our fully agentic pipeline with auto query rewriting, semantic search planning, and more:
 
 ![LEANN MCP Integration](assets/mcp_leann.png)
 
-**Ready to supercharge your coding?** [Complete Setup Guide →](packages/leann-mcp/README.md)
+**🔥 Ready to supercharge your coding?** [Complete Setup Guide →](packages/leann-mcp/README.md)
 
 ## 🖥️ Command Line Interface
 
@@ -455,7 +512,8 @@ leann --help
 **To make it globally available:**
 ```bash
 # Install the LEANN CLI globally using uv tool
-uv tool install leann-core
+uv tool install leann-core --with leann
+
 
 # Now you can use leann from anywhere without activating venv
 leann --help
@@ -479,30 +537,36 @@ leann ask my-docs --interactive
 
 # List all your indexes
 leann list
+
+# Remove an index
+leann remove my-docs
 ```
 
 **Key CLI features:**
-- Auto-detects document formats (PDF, TXT, MD, DOCX)
-- Smart text chunking with overlap
+- Auto-detects document formats (PDF, TXT, MD, DOCX, PPTX + code files)
+- **🧠 AST-aware chunking** for Python, Java, C#, TypeScript files
+- Smart text chunking with overlap for all other content
 - Multiple LLM providers (Ollama, OpenAI, HuggingFace)
-- Organized index storage in `~/.leann/indexes/`
+- Organized index storage in `.leann/indexes/` (project-local)
 - Support for advanced search parameters
 
 <details>
 <summary><strong>📋 Click to expand: Complete CLI Reference</strong></summary>
 
+You can use `leann --help`, or `leann build --help`, `leann search --help`, `leann ask --help`, `leann list --help`, `leann remove --help` to get the complete CLI reference.
+
 **Build Command:**
 ```bash
-leann build INDEX_NAME --docs DIRECTORY [OPTIONS]
+leann build INDEX_NAME --docs DIRECTORY|FILE [DIRECTORY|FILE ...] [OPTIONS]
 
 Options:
   --backend {hnsw,diskann}     Backend to use (default: hnsw)
   --embedding-model MODEL      Embedding model (default: facebook/contriever)
-  --graph-degree N            Graph degree (default: 32)
-  --complexity N              Build complexity (default: 64)
-  --force                     Force rebuild existing index
-  --compact                   Use compact storage (default: true)
-  --recompute                 Enable recomputation (default: true)
+  --graph-degree N             Graph degree (default: 32)
+  --complexity N               Build complexity (default: 64)
+  --force                      Force rebuild existing index
+  --compact / --no-compact     Use compact storage (default: true). Must be `no-compact` for `no-recompute` build.
+  --recompute / --no-recompute Enable recomputation (default: true)
 ```
 
 **Search Command:**
@@ -510,9 +574,9 @@ Options:
 leann search INDEX_NAME QUERY [OPTIONS]
 
 Options:
-  --top-k N                   Number of results (default: 5)
-  --complexity N              Search complexity (default: 64)
-  --recompute-embeddings      Use recomputation for highest accuracy
+  --top-k N                     Number of results (default: 5)
+  --complexity N                Search complexity (default: 64)
+  --recompute / --no-recompute  Enable/disable embedding recomputation (default: enabled). Should not do a `no-recompute` search in a `recompute` build.
   --pruning-strategy {global,local,proportional}
 ```
 
@@ -527,6 +591,31 @@ Options:
   --top-k N                  Retrieval count (default: 20)
 ```
 
+**List Command:**
+```bash
+leann list
+
+# Lists all indexes across all projects with status indicators:
+# ✅ - Index is complete and ready to use
+# ❌ - Index is incomplete or corrupted
+# 📁 - CLI-created index (in .leann/indexes/)
+# 📄 - App-created index (*.leann.meta.json files)
+```
+
+**Remove Command:**
+```bash
+leann remove INDEX_NAME [OPTIONS]
+
+Options:
+  --force, -f    Force removal without confirmation
+
+# Smart removal: automatically finds and safely removes indexes
+# - Shows all matching indexes across projects
+# - Requires confirmation for cross-project removal
+# - Interactive selection when multiple matches found
+# - Supports both CLI and app-created indexes
+```
+
 </details>
 
 ## 🏗️ Architecture & How It Works
@@ -568,6 +657,7 @@ Options:
 ```bash
 uv pip install -e ".[dev]"  # Install dev dependencies
 python benchmarks/run_evaluation.py    # Will auto-download evaluation data and run benchmarks
+python benchmarks/run_evaluation.py benchmarks/data/indices/rpj_wiki/rpj_wiki --num-queries 2000    # After downloading data, you can run the benchmark with our biggest index
 ```
 
 The evaluation script downloads data automatically on first run. The last three results were tested with partial personal data, and you can reproduce them with your own data!

apps/base_rag_example.py

@@ -10,7 +10,7 @@ from typing import Any
 
 import dotenv
 from leann.api import LeannBuilder, LeannChat
-from llama_index.core.node_parser import SentenceSplitter
+from leann.registry import register_project_directory
 
 dotenv.load_dotenv()
 
@@ -108,6 +108,38 @@ class BaseRAGExample(ABC):
             help="Thinking budget for reasoning models (low/medium/high). Supported by GPT-Oss:20b and other reasoning models.",
         )
 
+        # 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(
@@ -214,6 +246,11 @@ class BaseRAGExample(ABC):
         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):
@@ -304,21 +341,3 @@ class BaseRAGExample(ABC):
             await self.run_single_query(args, index_path, args.query)
         else:
             await self.run_interactive_chat(args, index_path)
-
-
-def create_text_chunks(documents, chunk_size=256, chunk_overlap=25) -> list[str]:
-    """Helper function to create text chunks from documents."""
-    node_parser = SentenceSplitter(
-        chunk_size=chunk_size,
-        chunk_overlap=chunk_overlap,
-        separator=" ",
-        paragraph_separator="\n\n",
-    )
-
-    all_texts = []
-    for doc in documents:
-        nodes = node_parser.get_nodes_from_documents([doc])
-        if nodes:
-            all_texts.extend(node.get_content() for node in nodes)
-
-    return all_texts

apps/chunking/__init__.py

@@ -0,0 +1,22 @@
+"""
+Chunking utilities for LEANN RAG applications.
+Provides AST-aware and traditional text chunking functionality.
+"""
+
+from .utils import (
+    CODE_EXTENSIONS,
+    create_ast_chunks,
+    create_text_chunks,
+    create_traditional_chunks,
+    detect_code_files,
+    get_language_from_extension,
+)
+
+__all__ = [
+    "CODE_EXTENSIONS",
+    "create_ast_chunks",
+    "create_text_chunks",
+    "create_traditional_chunks",
+    "detect_code_files",
+    "get_language_from_extension",
+]

apps/chunking/utils.py

@@ -0,0 +1,320 @@
+"""
+Enhanced chunking utilities with AST-aware code chunking support.
+Provides unified interface for both traditional and AST-based text chunking.
+"""
+
+import logging
+from pathlib import Path
+from typing import Optional
+
+from llama_index.core.node_parser import SentenceSplitter
+
+logger = logging.getLogger(__name__)
+
+# Code file extensions supported by astchunk
+CODE_EXTENSIONS = {
+    ".py": "python",
+    ".java": "java",
+    ".cs": "csharp",
+    ".ts": "typescript",
+    ".tsx": "typescript",
+    ".js": "typescript",
+    ".jsx": "typescript",
+}
+
+# Default chunk parameters for different content types
+DEFAULT_CHUNK_PARAMS = {
+    "code": {
+        "max_chunk_size": 512,
+        "chunk_overlap": 64,
+    },
+    "text": {
+        "chunk_size": 256,
+        "chunk_overlap": 128,
+    },
+}
+
+
+def detect_code_files(documents, code_extensions=None) -> tuple[list, list]:
+    """
+    Separate documents into code files and regular text files.
+
+    Args:
+        documents: List of LlamaIndex Document objects
+        code_extensions: Dict mapping file extensions to languages (defaults to CODE_EXTENSIONS)
+
+    Returns:
+        Tuple of (code_documents, text_documents)
+    """
+    if code_extensions is None:
+        code_extensions = CODE_EXTENSIONS
+
+    code_docs = []
+    text_docs = []
+
+    for doc in documents:
+        # Get file path from metadata
+        file_path = doc.metadata.get("file_path", "")
+        if not file_path:
+            # Fallback to file_name
+            file_path = doc.metadata.get("file_name", "")
+
+        if file_path:
+            file_ext = Path(file_path).suffix.lower()
+            if file_ext in code_extensions:
+                # Add language info to metadata
+                doc.metadata["language"] = code_extensions[file_ext]
+                doc.metadata["is_code"] = True
+                code_docs.append(doc)
+            else:
+                doc.metadata["is_code"] = False
+                text_docs.append(doc)
+        else:
+            # If no file path, treat as text
+            doc.metadata["is_code"] = False
+            text_docs.append(doc)
+
+    logger.info(f"Detected {len(code_docs)} code files and {len(text_docs)} text files")
+    return code_docs, text_docs
+
+
+def get_language_from_extension(file_path: str) -> Optional[str]:
+    """Get the programming language from file extension."""
+    ext = Path(file_path).suffix.lower()
+    return CODE_EXTENSIONS.get(ext)
+
+
+def create_ast_chunks(
+    documents,
+    max_chunk_size: int = 512,
+    chunk_overlap: int = 64,
+    metadata_template: str = "default",
+) -> list[str]:
+    """
+    Create AST-aware chunks from code documents using astchunk.
+
+    Args:
+        documents: List of code documents
+        max_chunk_size: Maximum characters per chunk
+        chunk_overlap: Number of AST nodes to overlap between chunks
+        metadata_template: Template for chunk metadata
+
+    Returns:
+        List of text chunks with preserved code structure
+    """
+    try:
+        from astchunk import ASTChunkBuilder
+    except ImportError as e:
+        logger.error(f"astchunk not available: {e}")
+        logger.info("Falling back to traditional chunking for code files")
+        return create_traditional_chunks(documents, max_chunk_size, chunk_overlap)
+
+    all_chunks = []
+
+    for doc in documents:
+        # Get language from metadata (set by detect_code_files)
+        language = doc.metadata.get("language")
+        if not language:
+            logger.warning(
+                "No language detected for document, falling back to traditional chunking"
+            )
+            traditional_chunks = create_traditional_chunks([doc], max_chunk_size, chunk_overlap)
+            all_chunks.extend(traditional_chunks)
+            continue
+
+        try:
+            # Configure astchunk
+            configs = {
+                "max_chunk_size": max_chunk_size,
+                "language": language,
+                "metadata_template": metadata_template,
+                "chunk_overlap": chunk_overlap if chunk_overlap > 0 else 0,
+            }
+
+            # Add repository-level metadata if available
+            repo_metadata = {
+                "file_path": doc.metadata.get("file_path", ""),
+                "file_name": doc.metadata.get("file_name", ""),
+                "creation_date": doc.metadata.get("creation_date", ""),
+                "last_modified_date": doc.metadata.get("last_modified_date", ""),
+            }
+            configs["repo_level_metadata"] = repo_metadata
+
+            # Create chunk builder and process
+            chunk_builder = ASTChunkBuilder(**configs)
+            code_content = doc.get_content()
+
+            if not code_content or not code_content.strip():
+                logger.warning("Empty code content, skipping")
+                continue
+
+            chunks = chunk_builder.chunkify(code_content)
+
+            # Extract text content from chunks
+            for chunk in chunks:
+                if hasattr(chunk, "text"):
+                    chunk_text = chunk.text
+                elif isinstance(chunk, dict) and "text" in chunk:
+                    chunk_text = chunk["text"]
+                elif isinstance(chunk, str):
+                    chunk_text = chunk
+                else:
+                    # Try to convert to string
+                    chunk_text = str(chunk)
+
+                if chunk_text and chunk_text.strip():
+                    all_chunks.append(chunk_text.strip())
+
+            logger.info(
+                f"Created {len(chunks)} AST chunks from {language} file: {doc.metadata.get('file_name', 'unknown')}"
+            )
+
+        except Exception as e:
+            logger.warning(f"AST chunking failed for {language} file: {e}")
+            logger.info("Falling back to traditional chunking")
+            traditional_chunks = create_traditional_chunks([doc], max_chunk_size, chunk_overlap)
+            all_chunks.extend(traditional_chunks)
+
+    return all_chunks
+
+
+def create_traditional_chunks(
+    documents, chunk_size: int = 256, chunk_overlap: int = 128
+) -> list[str]:
+    """
+    Create traditional text chunks using LlamaIndex SentenceSplitter.
+
+    Args:
+        documents: List of documents to chunk
+        chunk_size: Size of each chunk in characters
+        chunk_overlap: Overlap between chunks
+
+    Returns:
+        List of text chunks
+    """
+    # Handle invalid chunk_size values
+    if chunk_size <= 0:
+        logger.warning(f"Invalid chunk_size={chunk_size}, using default value of 256")
+        chunk_size = 256
+
+    # Ensure chunk_overlap is not negative and not larger than chunk_size
+    if chunk_overlap < 0:
+        chunk_overlap = 0
+    if chunk_overlap >= chunk_size:
+        chunk_overlap = chunk_size // 2
+
+    node_parser = SentenceSplitter(
+        chunk_size=chunk_size,
+        chunk_overlap=chunk_overlap,
+        separator=" ",
+        paragraph_separator="\n\n",
+    )
+
+    all_texts = []
+    for doc in documents:
+        try:
+            nodes = node_parser.get_nodes_from_documents([doc])
+            if nodes:
+                chunk_texts = [node.get_content() for node in nodes]
+                all_texts.extend(chunk_texts)
+                logger.debug(f"Created {len(chunk_texts)} traditional chunks from document")
+        except Exception as e:
+            logger.error(f"Traditional chunking failed for document: {e}")
+            # As last resort, add the raw content
+            content = doc.get_content()
+            if content and content.strip():
+                all_texts.append(content.strip())
+
+    return all_texts
+
+
+def create_text_chunks(
+    documents,
+    chunk_size: int = 256,
+    chunk_overlap: int = 128,
+    use_ast_chunking: bool = False,
+    ast_chunk_size: int = 512,
+    ast_chunk_overlap: int = 64,
+    code_file_extensions: Optional[list[str]] = None,
+    ast_fallback_traditional: bool = True,
+) -> list[str]:
+    """
+    Create text chunks from documents with optional AST support for code files.
+
+    Args:
+        documents: List of LlamaIndex Document objects
+        chunk_size: Size for traditional text chunks
+        chunk_overlap: Overlap for traditional text chunks
+        use_ast_chunking: Whether to use AST chunking for code files
+        ast_chunk_size: Size for AST chunks
+        ast_chunk_overlap: Overlap for AST chunks
+        code_file_extensions: Custom list of code file extensions
+        ast_fallback_traditional: Fall back to traditional chunking on AST errors
+
+    Returns:
+        List of text chunks
+    """
+    if not documents:
+        logger.warning("No documents provided for chunking")
+        return []
+
+    # Create a local copy of supported extensions for this function call
+    local_code_extensions = CODE_EXTENSIONS.copy()
+
+    # Update supported extensions if provided
+    if code_file_extensions:
+        # Map extensions to languages (simplified mapping)
+        ext_mapping = {
+            ".py": "python",
+            ".java": "java",
+            ".cs": "c_sharp",
+            ".ts": "typescript",
+            ".tsx": "typescript",
+        }
+        for ext in code_file_extensions:
+            if ext.lower() not in local_code_extensions:
+                # Try to guess language from extension
+                if ext.lower() in ext_mapping:
+                    local_code_extensions[ext.lower()] = ext_mapping[ext.lower()]
+                else:
+                    logger.warning(f"Unsupported extension {ext}, will use traditional chunking")
+
+    all_chunks = []
+
+    if use_ast_chunking:
+        # Separate code and text documents using local extensions
+        code_docs, text_docs = detect_code_files(documents, local_code_extensions)
+
+        # Process code files with AST chunking
+        if code_docs:
+            logger.info(f"Processing {len(code_docs)} code files with AST chunking")
+            try:
+                ast_chunks = create_ast_chunks(
+                    code_docs, max_chunk_size=ast_chunk_size, chunk_overlap=ast_chunk_overlap
+                )
+                all_chunks.extend(ast_chunks)
+                logger.info(f"Created {len(ast_chunks)} AST chunks from code files")
+            except Exception as e:
+                logger.error(f"AST chunking failed: {e}")
+                if ast_fallback_traditional:
+                    logger.info("Falling back to traditional chunking for code files")
+                    traditional_code_chunks = create_traditional_chunks(
+                        code_docs, chunk_size, chunk_overlap
+                    )
+                    all_chunks.extend(traditional_code_chunks)
+                else:
+                    raise
+
+        # Process text files with traditional chunking
+        if text_docs:
+            logger.info(f"Processing {len(text_docs)} text files with traditional chunking")
+            text_chunks = create_traditional_chunks(text_docs, chunk_size, chunk_overlap)
+            all_chunks.extend(text_chunks)
+            logger.info(f"Created {len(text_chunks)} traditional chunks from text files")
+    else:
+        # Use traditional chunking for all files
+        logger.info(f"Processing {len(documents)} documents with traditional chunking")
+        all_chunks = create_traditional_chunks(documents, chunk_size, chunk_overlap)
+
+    logger.info(f"Total chunks created: {len(all_chunks)}")
+    return all_chunks

apps/code_rag.py

@@ -0,0 +1,211 @@
+"""
+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

@@ -9,7 +9,8 @@ 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, create_text_chunks
+from base_rag_example import BaseRAGExample
+from chunking import create_text_chunks
 from llama_index.core import SimpleDirectoryReader
 
 
@@ -44,6 +45,11 @@ class DocumentRAG(BaseRAGExample):
         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."""
@@ -76,9 +82,22 @@ class DocumentRAG(BaseRAGExample):
 
         print(f"Loaded {len(documents)} documents")
 
-        # Convert to text chunks
+        # 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
+            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
@@ -102,6 +121,10 @@ if __name__ == "__main__":
     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()

benchmarks/benchmark_no_recompute.py

@@ -0,0 +1,148 @@
+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/data/.gitattributes

@@ -1,82 +0,0 @@
-*.7z filter=lfs diff=lfs merge=lfs -text
-*.arrow filter=lfs diff=lfs merge=lfs -text
-*.bin filter=lfs diff=lfs merge=lfs -text
-*.bz2 filter=lfs diff=lfs merge=lfs -text
-*.ckpt filter=lfs diff=lfs merge=lfs -text
-*.ftz filter=lfs diff=lfs merge=lfs -text
-*.gz filter=lfs diff=lfs merge=lfs -text
-*.h5 filter=lfs diff=lfs merge=lfs -text
-*.joblib filter=lfs diff=lfs merge=lfs -text
-*.lfs.* filter=lfs diff=lfs merge=lfs -text
-*.lz4 filter=lfs diff=lfs merge=lfs -text
-*.mds filter=lfs diff=lfs merge=lfs -text
-*.mlmodel filter=lfs diff=lfs merge=lfs -text
-*.model filter=lfs diff=lfs merge=lfs -text
-*.msgpack filter=lfs diff=lfs merge=lfs -text
-*.npy filter=lfs diff=lfs merge=lfs -text
-*.npz filter=lfs diff=lfs merge=lfs -text
-*.onnx filter=lfs diff=lfs merge=lfs -text
-*.ot filter=lfs diff=lfs merge=lfs -text
-*.parquet filter=lfs diff=lfs merge=lfs -text
-*.pb filter=lfs diff=lfs merge=lfs -text
-*.pickle filter=lfs diff=lfs merge=lfs -text
-*.pkl filter=lfs diff=lfs merge=lfs -text
-*.pt filter=lfs diff=lfs merge=lfs -text
-*.pth filter=lfs diff=lfs merge=lfs -text
-*.rar filter=lfs diff=lfs merge=lfs -text
-*.safetensors filter=lfs diff=lfs merge=lfs -text
-saved_model/**/* filter=lfs diff=lfs merge=lfs -text
-*.tar.* filter=lfs diff=lfs merge=lfs -text
-*.tar filter=lfs diff=lfs merge=lfs -text
-*.tflite filter=lfs diff=lfs merge=lfs -text
-*.tgz filter=lfs diff=lfs merge=lfs -text
-*.wasm filter=lfs diff=lfs merge=lfs -text
-*.xz filter=lfs diff=lfs merge=lfs -text
-*.zip filter=lfs diff=lfs merge=lfs -text
-*.zst filter=lfs diff=lfs merge=lfs -text
-*tfevents* filter=lfs diff=lfs merge=lfs -text
-# Audio files - uncompressed
-*.pcm filter=lfs diff=lfs merge=lfs -text
-*.sam filter=lfs diff=lfs merge=lfs -text
-*.raw filter=lfs diff=lfs merge=lfs -text
-# Audio files - compressed
-*.aac filter=lfs diff=lfs merge=lfs -text
-*.flac filter=lfs diff=lfs merge=lfs -text
-*.mp3 filter=lfs diff=lfs merge=lfs -text
-*.ogg filter=lfs diff=lfs merge=lfs -text
-*.wav filter=lfs diff=lfs merge=lfs -text
-# Image files - uncompressed
-*.bmp filter=lfs diff=lfs merge=lfs -text
-*.gif filter=lfs diff=lfs merge=lfs -text
-*.png filter=lfs diff=lfs merge=lfs -text
-*.tiff filter=lfs diff=lfs merge=lfs -text
-# Image files - compressed
-*.jpg filter=lfs diff=lfs merge=lfs -text
-*.jpeg filter=lfs diff=lfs merge=lfs -text
-*.webp filter=lfs diff=lfs merge=lfs -text
-# Video files - compressed
-*.mp4 filter=lfs diff=lfs merge=lfs -text
-*.webm filter=lfs diff=lfs merge=lfs -text
-ground_truth/dpr/id_map.json filter=lfs diff=lfs merge=lfs -text
-indices/dpr/dpr_diskann.passages.idx filter=lfs diff=lfs merge=lfs -text
-indices/dpr/dpr_diskann.passages.jsonl filter=lfs diff=lfs merge=lfs -text
-indices/dpr/dpr_diskann_disk.index filter=lfs diff=lfs merge=lfs -text
-indices/dpr/leann.labels.map filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/leann.labels.map filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.index filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.0.idx filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.0.jsonl filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.1.idx filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.1.jsonl filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.2.idx filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.2.jsonl filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.3.idx filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.3.jsonl filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.4.idx filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.4.jsonl filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.5.idx filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.5.jsonl filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.6.idx filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.6.jsonl filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.7.idx filter=lfs diff=lfs merge=lfs -text
-indices/rpj_wiki/rpj_wiki.passages.7.jsonl filter=lfs diff=lfs merge=lfs -text

benchmarks/diskann_vs_hnsw_speed_comparison.py

@@ -10,6 +10,7 @@ This benchmark compares search performance between DiskANN and HNSW backends:
 """
 
 import gc
+import multiprocessing as mp
 import tempfile
 import time
 from pathlib import Path
@@ -17,6 +18,12 @@ 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."""
@@ -113,10 +120,10 @@ def benchmark_backend(
         ]
         score_validity_rate = len(valid_scores) / len(all_scores) if all_scores else 0
 
-        # Clean up
+        # Clean up (ensure embedding server shutdown and object GC)
         try:
-            if hasattr(searcher, "__del__"):
-                searcher.__del__()
+            if hasattr(searcher, "cleanup"):
+                searcher.cleanup()
             del searcher
             del builder
             gc.collect()
@@ -259,10 +266,21 @@ if __name__ == "__main__":
         print(f"\n❌ Benchmark failed: {e}")
         sys.exit(1)
     finally:
-        # Ensure clean exit
+        # 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
-        sys.exit(0)
+        # Use os._exit to bypass atexit handlers that may hang in rare cases
+        import os as _os
+
+        _os._exit(0)

benchmarks/simple_mac_tpt_test.py

@@ -183,6 +183,9 @@ class Benchmark:
         start_time = time.time()
         with torch.no_grad():
             self.model(input_ids=input_ids, attention_mask=attention_mask)
+        # mps sync
+        if torch.backends.mps.is_available():
+            torch.mps.synchronize()
         end_time = time.time()
 
         return end_time - start_time

docs/ast_chunking_guide.md

@@ -0,0 +1,128 @@
+# 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 "."
+```
+
+## 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/configuration-guide.md

@@ -97,29 +97,23 @@ ollama pull nomic-embed-text
 ```
 
 ### DiskANN
-**Best for**: Performance-critical applications and large datasets - **Production-ready with automatic graph partitioning**
+**Best for**: Large datasets, especially when you want `recompute=True`.
 
-**How it works:**
-- **Product Quantization (PQ) + Real-time Reranking**: Uses compressed PQ codes for fast graph traversal, then recomputes exact embeddings for final candidates
-- **Automatic Graph Partitioning**: When `is_recompute=True`, automatically partitions large indices and safely removes redundant files to save storage
-- **Superior Speed-Accuracy Trade-off**: Faster search than HNSW while maintaining high accuracy
+**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
 
-**Trade-offs compared to HNSW:**
-- ✅ **Faster search latency** (typically 2-8x speedup)
-- ✅ **Better scaling** for large datasets
-- ✅ **Smart storage management** with automatic partitioning
-- ✅ **Better graph locality** with `--ldg-times` parameter for SSD optimization
-- ⚠️ **Slightly larger index size** due to PQ tables and graph metadata
+**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
-
-# For large-scale deployments
---backend-name diskann --graph-degree 64 --build-complexity 128
 ```
 
-**Performance Benchmark**: Run `python benchmarks/diskann_vs_hnsw_speed_comparison.py` to compare DiskANN and HNSW on your system.
+**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
 
@@ -273,24 +267,114 @@ Every configuration choice involves trade-offs:
 
 The key is finding the right balance for your specific use case. Start small and simple, measure performance, then scale up only where needed.
 
-## Deep Dive: Critical Configuration Decisions
+## Low-resource setups
 
-### When to Disable Recomputation
+If you don’t have a local GPU or builds/searches are too slow, use one or more of the options below.
 
-LEANN's recomputation feature provides exact distance calculations but can be disabled for extreme QPS requirements:
+### 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
---no-recompute  # Disable selective recomputation
+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
 ```
 
-**Trade-offs**:
-- **With recomputation** (default): Exact distances, best quality, higher latency, minimal storage (only stores metadata, recomputes embeddings on-demand)
-- **Without recomputation**: Must store full embeddings, significantly higher memory and storage usage (10-100x more), but faster search
+### 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)
+
 
-**Disable when**:
-- You have abundant storage and memory
-- Need extremely low latency (< 100ms)
-- Running a read-heavy workload where storage cost is acceptable
 
 ## Further Reading
 

docs/features.md

@@ -3,6 +3,7 @@
 ## 🔥 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

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

@@ -441,9 +441,14 @@ class DiskannSearcher(BaseSearcher):
         else:  # "global"
             use_global_pruning = True
 
-        # Perform search with suppressed C++ output based on log level
-        use_deferred_fetch = kwargs.get("USE_DEFERRED_FETCH", True)
-        recompute_neighors = False
+        # Strategy:
+        # - 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():
             labels, distances = self._index.batch_search(
                 query,

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

@@ -83,9 +83,7 @@ def create_diskann_embedding_server(
 
     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.global_offset_map)} passages from metadata"
-    )
+    logger.info(f"Loaded PassageManager with {len(passages)} passages from metadata")
 
     # Import protobuf after ensuring the path is correct
     try:

packages/leann-backend-diskann/pyproject.toml

@@ -4,8 +4,8 @@ build-backend = "scikit_build_core.build"
 
 [project]
 name = "leann-backend-diskann"
-version = "0.2.9"
-dependencies = ["leann-core==0.2.9", "numpy", "protobuf>=3.19.0"]
+version = "0.3.2"
+dependencies = ["leann-core==0.3.2", "numpy", "protobuf>=3.19.0"]
 
 [tool.scikit-build]
 # Key: simplified CMake path

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

@@ -54,12 +54,13 @@ 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:
-            if self.is_compact:
-                # TODO: support this case @andy
-                raise ValueError(
-                    "is_recompute is False, but is_compact is True. This is not compatible now. change is compact to False and you can use the original HNSW index."
-                )
+        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
@@ -184,9 +185,11 @@ class HNSWSearcher(BaseSearcher):
         """
         from . import faiss  # type: ignore
 
-        if not recompute_embeddings:
-            if self.is_pruned:
-                raise RuntimeError("Recompute is required for pruned index.")
+        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")

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

@@ -90,9 +90,7 @@ def create_hnsw_embedding_server(
         embedding_dim: int = int(meta.get("dimensions", 0))
     except Exception:
         embedding_dim = 0
-    logger.info(
-        f"Loaded PassageManager with {len(passages.global_offset_map)} passages from metadata"
-    )
+    logger.info(f"Loaded PassageManager with {len(passages)} passages from metadata")
 
     # (legacy ZMQ thread removed; using shutdown-capable server only)
 

packages/leann-backend-hnsw/pyproject.toml

@@ -6,10 +6,10 @@ build-backend = "scikit_build_core.build"
 
 [project]
 name = "leann-backend-hnsw"
-version = "0.2.9"
+version = "0.3.2"
 description = "Custom-built HNSW (Faiss) backend for the Leann toolkit."
 dependencies = [
-    "leann-core==0.2.9",
+    "leann-core==0.3.2",
     "numpy",
     "pyzmq>=23.0.0",
     "msgpack>=1.0.0",

packages/leann-core/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "leann-core"
-version = "0.2.9"
+version = "0.3.2"
 description = "Core API and plugin system for LEANN"
 readme = "README.md"
 requires-python = ">=3.9"

packages/leann-core/src/leann/api.py

@@ -46,6 +46,7 @@ def compute_embeddings(
             - "sentence-transformers": Use sentence-transformers library (default)
             - "mlx": Use MLX backend for Apple Silicon
             - "openai": Use OpenAI embedding API
+            - "gemini": Use Google Gemini embedding API
         use_server: Whether to use embedding server (True for search, False for build)
 
     Returns:
@@ -118,9 +119,12 @@ class PassageManager:
     def __init__(
         self, passage_sources: list[dict[str, Any]], metadata_file_path: Optional[str] = None
     ):
-        self.offset_maps = {}
-        self.passage_files = {}
-        self.global_offset_map = {}  # Combined map for fast lookup
+        self.offset_maps: dict[str, dict[str, int]] = {}
+        self.passage_files: dict[str, str] = {}
+        # Avoid materializing a single gigantic global map to reduce memory
+        # footprint on very large corpora (e.g., 60M+ passages). Instead, keep
+        # per-shard maps and do a lightweight per-shard lookup on demand.
+        self._total_count: int = 0
 
         # Derive index base name for standard sibling fallbacks, e.g., <index_name>.passages.*
         index_name_base = None
@@ -141,12 +145,25 @@ class PassageManager:
                 default_name: Optional[str],
                 source_dict: dict[str, Any],
             ) -> list[Path]:
+                """
+                Build an ordered list of candidate paths. For relative paths specified in
+                metadata, prefer resolution relative to the metadata file directory first,
+                then fall back to CWD-based resolution, and finally to conventional
+                sibling defaults (e.g., <index_base>.passages.idx / .jsonl).
+                """
                 candidates: list[Path] = []
-                # 1) Primary as-is (absolute or relative)
+                # 1) Primary path
                 if primary:
                     p = Path(primary)
-                    candidates.append(p if p.is_absolute() else (Path.cwd() / p))
-                # 2) metadata-relative explicit relative key
+                    if p.is_absolute():
+                        candidates.append(p)
+                    else:
+                        # Prefer metadata-relative resolution for relative paths
+                        if metadata_file_path:
+                            candidates.append(Path(metadata_file_path).parent / p)
+                        # Also consider CWD-relative as a fallback for legacy layouts
+                        candidates.append(Path.cwd() / p)
+                # 2) metadata-relative explicit relative key (if present)
                 if metadata_file_path and source_dict.get(relative_key):
                     candidates.append(Path(metadata_file_path).parent / source_dict[relative_key])
                 # 3) metadata-relative standard sibling filename
@@ -176,23 +193,28 @@ class PassageManager:
                 raise FileNotFoundError(f"Passage index file not found: {index_file}")
 
             with open(index_file, "rb") as f:
-                offset_map = pickle.load(f)
+                offset_map: dict[str, int] = pickle.load(f)
                 self.offset_maps[passage_file] = offset_map
                 self.passage_files[passage_file] = passage_file
-
-                # Build global map for O(1) lookup
-                for passage_id, offset in offset_map.items():
-                    self.global_offset_map[passage_id] = (passage_file, offset)
+                self._total_count += len(offset_map)
 
     def get_passage(self, passage_id: str) -> dict[str, Any]:
-        if passage_id in self.global_offset_map:
-            passage_file, offset = self.global_offset_map[passage_id]
-            # Lazy file opening - only open when needed
-            with open(passage_file, encoding="utf-8") as f:
-                f.seek(offset)
-                return json.loads(f.readline())
+        # Fast path: check each shard map (there are typically few shards).
+        # This avoids building a massive combined dict while keeping lookups
+        # bounded by the number of shards.
+        for passage_file, offset_map in self.offset_maps.items():
+            try:
+                offset = offset_map[passage_id]
+                with open(passage_file, encoding="utf-8") as f:
+                    f.seek(offset)
+                    return json.loads(f.readline())
+            except KeyError:
+                continue
         raise KeyError(f"Passage ID not found: {passage_id}")
 
+    def __len__(self) -> int:
+        return self._total_count
+
 
 class LeannBuilder:
     def __init__(
@@ -204,6 +226,18 @@ class LeannBuilder:
         **backend_kwargs,
     ):
         self.backend_name = backend_name
+        # Normalize incompatible combinations early (for consistent metadata)
+        if backend_name == "hnsw":
+            is_recompute = backend_kwargs.get("is_recompute", True)
+            is_compact = backend_kwargs.get("is_compact", True)
+            if is_recompute is False and is_compact is True:
+                warnings.warn(
+                    "HNSW with is_recompute=False requires non-compact storage. Forcing is_compact=False.",
+                    UserWarning,
+                    stacklevel=2,
+                )
+                backend_kwargs["is_compact"] = False
+
         backend_factory: Optional[LeannBackendFactoryInterface] = BACKEND_REGISTRY.get(backend_name)
         if backend_factory is None:
             raise ValueError(f"Backend '{backend_name}' not found or not registered.")
@@ -294,6 +328,23 @@ class LeannBuilder:
     def build_index(self, index_path: str):
         if not self.chunks:
             raise ValueError("No chunks added.")
+
+        # Filter out invalid/empty text chunks early to keep passage and embedding counts aligned
+        valid_chunks: list[dict[str, Any]] = []
+        skipped = 0
+        for chunk in self.chunks:
+            text = chunk.get("text", "")
+            if isinstance(text, str) and text.strip():
+                valid_chunks.append(chunk)
+            else:
+                skipped += 1
+        if skipped > 0:
+            print(
+                f"Warning: Skipping {skipped} empty/invalid text chunk(s). Processing {len(valid_chunks)} valid chunks"
+            )
+            self.chunks = valid_chunks
+            if not self.chunks:
+                raise ValueError("All provided chunks are empty or invalid. Nothing to index.")
         if self.dimensions is None:
             self.dimensions = len(
                 compute_embeddings(
@@ -523,6 +574,7 @@ class LeannSearcher:
         self.embedding_model = self.meta_data["embedding_model"]
         # Support both old and new format
         self.embedding_mode = self.meta_data.get("embedding_mode", "sentence-transformers")
+        # Delegate portability handling to PassageManager
         self.passage_manager = PassageManager(
             self.meta_data.get("passage_sources", []), metadata_file_path=self.meta_path_str
         )
@@ -553,7 +605,9 @@ class LeannSearcher:
         logger.info(f"  Additional kwargs: {kwargs}")
 
         # Smart top_k detection and adjustment
-        total_docs = len(self.passage_manager.global_offset_map)
+        # Use PassageManager length (sum of shard sizes) to avoid
+        # depending on a massive combined map
+        total_docs = len(self.passage_manager)
         original_top_k = top_k
         if top_k > total_docs:
             top_k = total_docs
@@ -583,7 +637,7 @@ class LeannSearcher:
             zmq_port=zmq_port,
         )
         # logger.info(f"  Generated embedding shape: {query_embedding.shape}")
-        time.time() - start_time
+        # time.time() - start_time
         # logger.info(f"  Embedding time: {embedding_time} seconds")
 
         start_time = time.time()
@@ -649,8 +703,26 @@ class LeannSearcher:
         This method should be called after you're done using the searcher,
         especially in test environments or batch processing scenarios.
         """
-        if hasattr(self.backend_impl, "embedding_server_manager"):
-            self.backend_impl.embedding_server_manager.stop_server()
+        backend = getattr(self.backend_impl, "embedding_server_manager", None)
+        if backend is not None:
+            backend.stop_server()
+
+    # Enable automatic cleanup patterns
+    def __enter__(self):
+        return self
+
+    def __exit__(self, exc_type, exc, tb):
+        try:
+            self.cleanup()
+        except Exception:
+            pass
+
+    def __del__(self):
+        try:
+            self.cleanup()
+        except Exception:
+            # Avoid noisy errors during interpreter shutdown
+            pass
 
 
 class LeannChat:
@@ -730,3 +802,19 @@ class LeannChat:
         """
         if hasattr(self.searcher, "cleanup"):
             self.searcher.cleanup()
+
+    # Enable automatic cleanup patterns
+    def __enter__(self):
+        return self
+
+    def __exit__(self, exc_type, exc, tb):
+        try:
+            self.cleanup()
+        except Exception:
+            pass
+
+    def __del__(self):
+        try:
+            self.cleanup()
+        except Exception:
+            pass

packages/leann-core/src/leann/chat.py

@@ -422,7 +422,6 @@ class LLMInterface(ABC):
                 top_k=10,
                 complexity=64,
                 beam_width=8,
-                USE_DEFERRED_FETCH=True,
                 skip_search_reorder=True,
                 recompute_beighbor_embeddings=True,
                 dedup_node_dis=True,
@@ -434,7 +433,6 @@ class LLMInterface(ABC):
         Supported kwargs:
             - complexity (int): Search complexity parameter (default: 32)
             - beam_width (int): Beam width for search (default: 4)
-            - USE_DEFERRED_FETCH (bool): Enable deferred fetch mode (default: False)
             - skip_search_reorder (bool): Skip search reorder step (default: False)
             - recompute_beighbor_embeddings (bool): Enable ZMQ embedding server for neighbor recomputation (default: False)
             - dedup_node_dis (bool): Deduplicate nodes by distance (default: False)
@@ -682,6 +680,60 @@ class HFChat(LLMInterface):
         return response.strip()
 
 
+class GeminiChat(LLMInterface):
+    """LLM interface for Google Gemini models."""
+
+    def __init__(self, model: str = "gemini-2.5-flash", api_key: Optional[str] = None):
+        self.model = model
+        self.api_key = api_key or os.getenv("GEMINI_API_KEY")
+
+        if not self.api_key:
+            raise ValueError(
+                "Gemini API key is required. Set GEMINI_API_KEY environment variable or pass api_key parameter."
+            )
+
+        logger.info(f"Initializing Gemini Chat with model='{model}'")
+
+        try:
+            import google.genai as genai
+
+            self.client = genai.Client(api_key=self.api_key)
+        except ImportError:
+            raise ImportError(
+                "The 'google-genai' library is required for Gemini models. Please install it with 'uv pip install google-genai'."
+            )
+
+    def ask(self, prompt: str, **kwargs) -> str:
+        logger.info(f"Sending request to Gemini with model {self.model}")
+
+        try:
+            from google.genai.types import GenerateContentConfig
+
+            generation_config = GenerateContentConfig(
+                temperature=kwargs.get("temperature", 0.7),
+                max_output_tokens=kwargs.get("max_tokens", 1000),
+            )
+
+            # Handle top_p parameter
+            if "top_p" in kwargs:
+                generation_config.top_p = kwargs["top_p"]
+
+            response = self.client.models.generate_content(
+                model=self.model,
+                contents=prompt,
+                config=generation_config,
+            )
+            # Handle potential None response text
+            response_text = response.text
+            if response_text is None:
+                logger.warning("Gemini returned None response text")
+                return ""
+            return response_text.strip()
+        except Exception as e:
+            logger.error(f"Error communicating with Gemini: {e}")
+            return f"Error: Could not get a response from Gemini. Details: {e}"
+
+
 class OpenAIChat(LLMInterface):
     """LLM interface for OpenAI models."""
 
@@ -795,6 +847,8 @@ def get_llm(llm_config: Optional[dict[str, Any]] = None) -> LLMInterface:
         return HFChat(model_name=model or "deepseek-ai/deepseek-llm-7b-chat")
     elif llm_type == "openai":
         return OpenAIChat(model=model or "gpt-4o", api_key=llm_config.get("api_key"))
+    elif llm_type == "gemini":
+        return GeminiChat(model=model or "gemini-2.5-flash", api_key=llm_config.get("api_key"))
     elif llm_type == "simulated":
         return SimulatedChat()
     else:

packages/leann-core/src/leann/embedding_compute.py

@@ -57,6 +57,8 @@ def compute_embeddings(
         return compute_embeddings_mlx(texts, model_name)
     elif mode == "ollama":
         return compute_embeddings_ollama(texts, model_name, is_build=is_build)
+    elif mode == "gemini":
+        return compute_embeddings_gemini(texts, model_name, is_build=is_build)
     else:
         raise ValueError(f"Unsupported embedding mode: {mode}")
 
@@ -244,6 +246,16 @@ def compute_embeddings_openai(texts: list[str], model_name: str) -> np.ndarray:
     except ImportError as e:
         raise ImportError(f"OpenAI package not installed: {e}")
 
+    # Validate input list
+    if not texts:
+        raise ValueError("Cannot compute embeddings for empty text list")
+    # Extra validation: abort early if any item is empty/whitespace
+    invalid_count = sum(1 for t in texts if not isinstance(t, str) or not t.strip())
+    if invalid_count > 0:
+        raise ValueError(
+            f"Found {invalid_count} empty/invalid text(s) in input. Upstream should filter before calling OpenAI."
+        )
+
     api_key = os.getenv("OPENAI_API_KEY")
     if not api_key:
         raise RuntimeError("OPENAI_API_KEY environment variable not set")
@@ -263,8 +275,16 @@ def compute_embeddings_openai(texts: list[str], model_name: str) -> np.ndarray:
     print(f"len of texts: {len(texts)}")
 
     # OpenAI has limits on batch size and input length
-    max_batch_size = 1000  # Conservative batch size
+    max_batch_size = 800  # Conservative batch size because the token limit is 300K
     all_embeddings = []
+    # get the avg len of texts
+    avg_len = sum(len(text) for text in texts) / len(texts)
+    print(f"avg len of texts: {avg_len}")
+    # if avg len is less than 1000, use the max batch size
+    if avg_len > 300:
+        max_batch_size = 500
+
+    # if avg len is less than 1000, use the max batch size
 
     try:
         from tqdm import tqdm
@@ -650,3 +670,83 @@ def compute_embeddings_ollama(
     logger.info(f"Generated {len(embeddings)} embeddings, dimension: {embeddings.shape[1]}")
 
     return embeddings
+
+
+def compute_embeddings_gemini(
+    texts: list[str], model_name: str = "text-embedding-004", is_build: bool = False
+) -> np.ndarray:
+    """
+    Compute embeddings using Google Gemini API.
+
+    Args:
+        texts: List of texts to compute embeddings for
+        model_name: Gemini model name (default: "text-embedding-004")
+        is_build: Whether this is a build operation (shows progress bar)
+
+    Returns:
+        Embeddings array, shape: (len(texts), embedding_dim)
+    """
+    try:
+        import os
+
+        import google.genai as genai
+    except ImportError as e:
+        raise ImportError(f"Google GenAI package not installed: {e}")
+
+    api_key = os.getenv("GEMINI_API_KEY")
+    if not api_key:
+        raise RuntimeError("GEMINI_API_KEY environment variable not set")
+
+    # Cache Gemini client
+    cache_key = "gemini_client"
+    if cache_key in _model_cache:
+        client = _model_cache[cache_key]
+    else:
+        client = genai.Client(api_key=api_key)
+        _model_cache[cache_key] = client
+        logger.info("Gemini client cached")
+
+    logger.info(
+        f"Computing embeddings for {len(texts)} texts using Gemini API, model: '{model_name}'"
+    )
+
+    # Gemini supports batch embedding
+    max_batch_size = 100  # Conservative batch size for Gemini
+    all_embeddings = []
+
+    try:
+        from tqdm import tqdm
+
+        total_batches = (len(texts) + max_batch_size - 1) // max_batch_size
+        batch_range = range(0, len(texts), max_batch_size)
+        batch_iterator = tqdm(
+            batch_range, desc="Computing embeddings", unit="batch", total=total_batches
+        )
+    except ImportError:
+        # Fallback when tqdm is not available
+        batch_iterator = range(0, len(texts), max_batch_size)
+
+    for i in batch_iterator:
+        batch_texts = texts[i : i + max_batch_size]
+
+        try:
+            # Use the embed_content method from the new Google GenAI SDK
+            response = client.models.embed_content(
+                model=model_name,
+                contents=batch_texts,
+                config=genai.types.EmbedContentConfig(
+                    task_type="RETRIEVAL_DOCUMENT"  # For document embedding
+                ),
+            )
+
+            # Extract embeddings from response
+            for embedding_data in response.embeddings:
+                all_embeddings.append(embedding_data.values)
+        except Exception as e:
+            logger.error(f"Batch {i} failed: {e}")
+            raise
+
+    embeddings = np.array(all_embeddings, dtype=np.float32)
+    logger.info(f"Generated {len(embeddings)} embeddings, dimension: {embeddings.shape[1]}")
+
+    return embeddings

packages/leann-core/src/leann/embedding_server_manager.py

@@ -192,6 +192,7 @@ class EmbeddingServerManager:
             stderr_target = None  # Direct to console for visible logs
 
         # Start embedding server subprocess
+        logger.info(f"Starting server process with command: {' '.join(command)}")
         self.server_process = subprocess.Popen(
             command,
             cwd=project_root,
@@ -268,8 +269,12 @@ class EmbeddingServerManager:
             f"Terminating server process (PID: {self.server_process.pid}) for backend {self.backend_module_name}..."
         )
 
-        # Use simple termination - our improved server shutdown should handle this properly
-        self.server_process.terminate()
+        # Use simple termination first; if the server installed signal handlers,
+        # it will exit cleanly. Otherwise escalate to kill after a short wait.
+        try:
+            self.server_process.terminate()
+        except Exception:
+            pass
 
         try:
             self.server_process.wait(timeout=5)  # Give more time for graceful shutdown
@@ -278,7 +283,10 @@ class EmbeddingServerManager:
             logger.warning(
                 f"Server process {self.server_process.pid} did not terminate within 5 seconds, force killing..."
             )
-            self.server_process.kill()
+            try:
+                self.server_process.kill()
+            except Exception:
+                pass
             try:
                 self.server_process.wait(timeout=2)
                 logger.info(f"Server process {self.server_process.pid} killed successfully.")

packages/leann-core/src/leann/mcp.py

@@ -64,19 +64,6 @@ def handle_request(request):
                             "required": ["index_name", "query"],
                         },
                     },
-                    {
-                        "name": "leann_status",
-                        "description": "📊 Check the health and stats of your code indexes - like a medical checkup for your codebase knowledge!",
-                        "inputSchema": {
-                            "type": "object",
-                            "properties": {
-                                "index_name": {
-                                    "type": "string",
-                                    "description": "Optional: Name of specific index to check. If not provided, shows status of all indexes.",
-                                }
-                            },
-                        },
-                    },
                     {
                         "name": "leann_list",
                         "description": "📋 Show all your indexed codebases - your personal code library! Use this to see what's available for search.",
@@ -107,7 +94,7 @@ def handle_request(request):
                         },
                     }
 
-                # Build simplified command
+                # Build simplified command with non-interactive flag for MCP compatibility
                 cmd = [
                     "leann",
                     "search",
@@ -115,18 +102,10 @@ def handle_request(request):
                     args["query"],
                     f"--top-k={args.get('top_k', 5)}",
                     f"--complexity={args.get('complexity', 32)}",
+                    "--non-interactive",
                 ]
                 result = subprocess.run(cmd, capture_output=True, text=True)
 
-            elif tool_name == "leann_status":
-                if args.get("index_name"):
-                    # Check specific index status - for now, we'll use leann list and filter
-                    result = subprocess.run(["leann", "list"], capture_output=True, text=True)
-                    # We could enhance this to show more detailed status per index
-                else:
-                    # Show all indexes status
-                    result = subprocess.run(["leann", "list"], capture_output=True, text=True)
-
             elif tool_name == "leann_list":
                 result = subprocess.run(["leann", "list"], capture_output=True, text=True)
 

packages/leann-core/src/leann/registry.py

@@ -2,11 +2,17 @@
 
 import importlib
 import importlib.metadata
-from typing import TYPE_CHECKING
+import json
+import logging
+from pathlib import Path
+from typing import TYPE_CHECKING, Optional, Union
 
 if TYPE_CHECKING:
     from leann.interface import LeannBackendFactoryInterface
 
+# Set up logger for this module
+logger = logging.getLogger(__name__)
+
 BACKEND_REGISTRY: dict[str, "LeannBackendFactoryInterface"] = {}
 
 
@@ -14,7 +20,7 @@ def register_backend(name: str):
     """A decorator to register a new backend class."""
 
     def decorator(cls):
-        print(f"INFO: Registering backend '{name}'")
+        logger.debug(f"Registering backend '{name}'")
         BACKEND_REGISTRY[name] = cls
         return cls
 
@@ -39,3 +45,54 @@ def autodiscover_backends():
             # print(f"WARN: Could not import backend module '{backend_module_name}': {e}")
             pass
     # print("INFO: Backend auto-discovery finished.")
+
+
+def register_project_directory(project_dir: Optional[Union[str, Path]] = None):
+    """
+    Register a project directory in the global LEANN registry.
+
+    This allows `leann list` to discover indexes created by apps or other tools.
+
+    Args:
+        project_dir: Directory to register. If None, uses current working directory.
+    """
+    if project_dir is None:
+        project_dir = Path.cwd()
+    else:
+        project_dir = Path(project_dir)
+
+    # Only register directories that have some kind of LEANN content
+    # Either .leann/indexes/ (CLI format) or *.leann.meta.json files (apps format)
+    has_cli_indexes = (project_dir / ".leann" / "indexes").exists()
+    has_app_indexes = any(project_dir.rglob("*.leann.meta.json"))
+
+    if not (has_cli_indexes or has_app_indexes):
+        # Don't register if there are no LEANN indexes
+        return
+
+    global_registry = Path.home() / ".leann" / "projects.json"
+    global_registry.parent.mkdir(exist_ok=True)
+
+    project_str = str(project_dir.resolve())
+
+    # Load existing registry
+    projects = []
+    if global_registry.exists():
+        try:
+            with open(global_registry) as f:
+                projects = json.load(f)
+        except Exception:
+            logger.debug("Could not load existing project registry")
+            projects = []
+
+    # Add project if not already present
+    if project_str not in projects:
+        projects.append(project_str)
+
+        # Save updated registry
+        try:
+            with open(global_registry, "w") as f:
+                json.dump(projects, f, indent=2)
+            logger.debug(f"Registered project directory: {project_str}")
+        except Exception as e:
+            logger.warning(f"Could not save project registry: {e}")

packages/leann-mcp/README.md

@@ -13,10 +13,20 @@ This installs the `leann` CLI into an isolated tool environment and includes bot
 
 ## 🚀 Quick Setup
 
-Add the LEANN MCP server to Claude Code:
+Add the LEANN MCP server to Claude Code. Choose the scope based on how widely you want it available. Below is the command to install it globally; if you prefer a local install, skip this step:
 
 ```bash
-claude mcp add leann-server -- leann_mcp
+# Global (recommended): available in all projects for your user
+claude mcp add --scope user leann-server -- leann_mcp
+```
+
+- `leann-server`: the display name of the MCP server in Claude Code (you can change it).
+- `leann_mcp`: the Python entry point installed with LEANN that starts the MCP server.
+
+Verify it is registered globally:
+
+```bash
+claude mcp list | cat
 ```
 
 ## 🛠️ Available Tools
@@ -25,27 +35,36 @@ Once connected, you'll have access to these powerful semantic search tools in Cl
 
 - **`leann_list`** - List all available indexes across your projects
 - **`leann_search`** - Perform semantic searches across code and documents
-- **`leann_ask`** - Ask natural language questions and get AI-powered answers from your codebase
+
 
 ## 🎯 Quick Start Example
 
 ```bash
+# Add locally if you did not add it globally (current folder only; default if --scope is omitted)
+claude mcp add leann-server -- leann_mcp
+
 # Build an index for your project (change to your actual path)
-leann build my-project --docs ./
+# See the advanced examples below for more ways to configure indexing
+# Set the index name (replace 'my-project' with your own)
+leann build my-project --docs $(git ls-files)
 
 # Start Claude Code
 claude
 ```
 
-## 🚀 Advanced Usage Examples
+## 🚀 Advanced Usage Examples to build the index
 
 ### Index Entire Git Repository
 ```bash
-# Index all tracked files in your git repository, note right now we will skip submodules, but we can add it back easily if you want
+# Index all tracked files in your Git repository.
+# Note: submodules are currently skipped; we can add them back if needed.
 leann build my-repo --docs $(git ls-files) --embedding-mode sentence-transformers --embedding-model all-MiniLM-L6-v2 --backend hnsw
 
-# Index only specific file types from git
+# Index only tracked Python files from Git.
 leann build my-python-code --docs $(git ls-files "*.py") --embedding-mode sentence-transformers --embedding-model all-MiniLM-L6-v2 --backend hnsw
+
+# If you encounter empty requests caused by empty files (e.g., __init__.py), exclude zero-byte files. Thanks @ww2283 for pointing [that](https://github.com/yichuan-w/LEANN/issues/48) out
+leann build leann-prospec-lig --docs $(find ./src -name "*.py" -not -empty) --embedding-mode openai --embedding-model text-embedding-3-small
 ```
 
 ### Multiple Directories and Files
@@ -73,7 +92,7 @@ leann build docs-and-configs --docs $(git ls-files "*.md" "*.yml" "*.yaml" "*.js
 ```
 
 
-**Try this in Claude Code:**
+## **Try this in Claude Code:**
 ```
 Help me understand this codebase. List available indexes and search for authentication patterns.
 ```
@@ -82,6 +101,7 @@ Help me understand this codebase. List available indexes and search for authenti
   <img src="../../assets/claude_code_leann.png" alt="LEANN in Claude Code" width="80%">
 </p>
 
+If you see a prompt asking whether to proceed with LEANN, you can now use it in your chat!
 
 ## 🧠 How It Works
 
@@ -117,3 +137,11 @@ To remove LEANN
 ```
 uv pip uninstall leann leann-backend-hnsw leann-core
 ```
+
+To globally remove LEANN (for version update)
+```
+uv tool list | cat
+uv tool uninstall leann-core
+command -v leann || echo "leann gone"
+command -v leann_mcp || echo "leann_mcp gone"
+```

packages/leann/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "leann"
-version = "0.2.9"
+version = "0.3.2"
 description = "LEANN - The smallest vector index in the world. RAG Everything with LEANN!"
 readme = "README.md"
 requires-python = ">=3.9"

packages/wechat-exporter/__init__.py

@@ -0,0 +1 @@
+__all__ = []

packages/wechat-exporter/main.py

@@ -136,5 +136,9 @@ def export_sqlite(
     connection.commit()
 
 
-if __name__ == "__main__":
+def main():
     app()
+
+
+if __name__ == "__main__":
+    main()

pyproject.toml

@@ -10,11 +10,10 @@ requires-python = ">=3.9"
 dependencies = [
     "leann-core",
     "leann-backend-hnsw",
+    "typer>=0.12.3",
     "numpy>=1.26.0",
     "torch",
     "tqdm",
-    "flask",
-    "flask_compress",
     "datasets>=2.15.0",
     "evaluate",
     "colorama",
@@ -47,6 +46,13 @@ dependencies = [
     "pathspec>=0.12.1",
     "nbconvert>=7.16.6",
     "gitignore-parser>=0.1.12",
+    # AST-aware code chunking dependencies
+    "astchunk>=0.1.0",
+    "tree-sitter>=0.20.0",
+    "tree-sitter-python>=0.20.0",
+    "tree-sitter-java>=0.20.0",
+    "tree-sitter-c-sharp>=0.20.0",
+    "tree-sitter-typescript>=0.20.0",
 ]
 
 [project.optional-dependencies]
@@ -65,9 +71,7 @@ test = [
     "pytest>=7.0",
     "pytest-timeout>=2.0",
     "llama-index-core>=0.12.0",
-    "llama-index-readers-file>=0.4.0",
     "python-dotenv>=1.0.0",
-    "sentence-transformers>=2.2.0",
 ]
 
 diskann = [
@@ -84,6 +88,11 @@ documents = [
 
 [tool.setuptools]
 py-modules = []
+packages = ["wechat_exporter"]
+package-dir = { "wechat_exporter" = "packages/wechat-exporter" }
+
+[project.scripts]
+wechat-exporter = "wechat_exporter.main:main"
 
 
 [tool.uv.sources]
@@ -94,13 +103,8 @@ leann-backend-hnsw = { path = "packages/leann-backend-hnsw", editable = true }
 [tool.ruff]
 target-version = "py39"
 line-length = 100
-extend-exclude = [
-    "third_party",
-    "*.egg-info",
-    "__pycache__",
-    ".git",
-    ".venv",
-]
+extend-exclude = ["third_party"]
+
 
 [tool.ruff.lint]
 select = [
@@ -123,21 +127,12 @@ ignore = [
     "RUF012", # mutable class attributes should be annotated with typing.ClassVar
 ]
 
-[tool.ruff.lint.per-file-ignores]
-"test/**/*.py" = ["E402"]      # module level import not at top of file (common in tests)
-"examples/**/*.py" = ["E402"]  # module level import not at top of file (common in examples)
-
 [tool.ruff.format]
 quote-style = "double"
 indent-style = "space"
 skip-magic-trailing-comma = false
 line-ending = "auto"
 
-[dependency-groups]
-dev = [
-    "ruff>=0.12.4",
-]
-
 [tool.lychee]
 accept = ["200", "403", "429", "503"]
 timeout = 20

sky/leann-build.yaml

@@ -0,0 +1,76 @@
+name: leann-build
+
+resources:
+  # Choose a GPU for fast embeddings (examples: L4, A10G, A100). CPU also works but is slower.
+  accelerators: L4:1
+  # Optionally pin a cloud, otherwise SkyPilot will auto-select
+  # cloud: aws
+  disk_size: 100
+
+envs:
+  # Build parameters (override with: sky launch -c leann-gpu sky/leann-build.yaml -e key=value)
+  index_name: my-index
+  docs: ./data
+  backend: hnsw               # hnsw | diskann
+  complexity: 64
+  graph_degree: 32
+  num_threads: 8
+  # Embedding selection
+  embedding_mode: sentence-transformers   # sentence-transformers | openai | mlx | ollama
+  embedding_model: facebook/contriever
+  # Storage/latency knobs
+  recompute: true             # true => selective recomputation (recommended)
+  compact: true               # for HNSW only
+  # Optional pass-through
+  extra_args: ""
+  # Rebuild control
+  force: true
+
+# Sync local paths to the remote VM. Adjust as needed.
+file_mounts:
+  # Example: mount your local data directory used for building
+  ~/leann-data: ${docs}
+
+setup: |
+  set -e
+  # Install uv (package manager)
+  curl -LsSf https://astral.sh/uv/install.sh | sh
+  export PATH="$HOME/.local/bin:$PATH"
+
+  # Ensure modern libstdc++ for FAISS (GLIBCXX >= 3.4.30)
+  sudo apt-get update -y
+  sudo apt-get install -y libstdc++6 libgomp1
+  # Also upgrade conda's libstdc++ in base env (Skypilot images include conda)
+  if command -v conda >/dev/null 2>&1; then
+    conda install -y -n base -c conda-forge libstdcxx-ng
+  fi
+
+  # Install LEANN CLI and backends into the user environment
+  uv pip install --upgrade pip
+  uv pip install leann-core leann-backend-hnsw leann-backend-diskann
+
+run: |
+  export PATH="$HOME/.local/bin:$PATH"
+  # Derive flags from env
+  recompute_flag=""
+  if [ "${recompute}" = "false" ] || [ "${recompute}" = "0" ]; then
+    recompute_flag="--no-recompute"
+  fi
+  force_flag=""
+  if [ "${force}" = "true" ] || [ "${force}" = "1" ]; then
+    force_flag="--force"
+  fi
+
+  # Build command
+  python -m leann.cli build ${index_name} \
+    --docs ~/leann-data \
+    --backend ${backend} \
+    --complexity ${complexity} \
+    --graph-degree ${graph_degree} \
+    --num-threads ${num_threads} \
+    --embedding-mode ${embedding_mode} \
+    --embedding-model ${embedding_model} \
+    ${recompute_flag} ${force_flag} ${extra_args}
+
+  # Print where the index is stored for downstream rsync
+  echo "INDEX_OUT_DIR=~/.leann/indexes/${index_name}"

tests/test_astchunk_integration.py

@@ -0,0 +1,397 @@
+"""
+Test suite for astchunk integration with LEANN.
+Tests AST-aware chunking functionality, language detection, and fallback mechanisms.
+"""
+
+import os
+import subprocess
+import sys
+import tempfile
+from pathlib import Path
+from unittest.mock import patch
+
+import pytest
+
+# Add apps directory to path for imports
+sys.path.insert(0, str(Path(__file__).parent.parent / "apps"))
+
+from typing import Optional
+
+from chunking import (
+    create_ast_chunks,
+    create_text_chunks,
+    create_traditional_chunks,
+    detect_code_files,
+    get_language_from_extension,
+)
+
+
+class MockDocument:
+    """Mock LlamaIndex Document for testing."""
+
+    def __init__(self, content: str, file_path: str = "", metadata: Optional[dict] = None):
+        self.content = content
+        self.metadata = metadata or {}
+        if file_path:
+            self.metadata["file_path"] = file_path
+
+    def get_content(self) -> str:
+        return self.content
+
+
+class TestCodeFileDetection:
+    """Test code file detection and language mapping."""
+
+    def test_detect_code_files_python(self):
+        """Test detection of Python files."""
+        docs = [
+            MockDocument("print('hello')", "/path/to/file.py"),
+            MockDocument("This is text", "/path/to/file.txt"),
+        ]
+
+        code_docs, text_docs = detect_code_files(docs)
+
+        assert len(code_docs) == 1
+        assert len(text_docs) == 1
+        assert code_docs[0].metadata["language"] == "python"
+        assert code_docs[0].metadata["is_code"] is True
+        assert text_docs[0].metadata["is_code"] is False
+
+    def test_detect_code_files_multiple_languages(self):
+        """Test detection of multiple programming languages."""
+        docs = [
+            MockDocument("def func():", "/path/to/script.py"),
+            MockDocument("public class Test {}", "/path/to/Test.java"),
+            MockDocument("interface ITest {}", "/path/to/test.ts"),
+            MockDocument("using System;", "/path/to/Program.cs"),
+            MockDocument("Regular text content", "/path/to/document.txt"),
+        ]
+
+        code_docs, text_docs = detect_code_files(docs)
+
+        assert len(code_docs) == 4
+        assert len(text_docs) == 1
+
+        languages = [doc.metadata["language"] for doc in code_docs]
+        assert "python" in languages
+        assert "java" in languages
+        assert "typescript" in languages
+        assert "csharp" in languages
+
+    def test_detect_code_files_no_file_path(self):
+        """Test handling of documents without file paths."""
+        docs = [
+            MockDocument("some content"),
+            MockDocument("other content", metadata={"some_key": "value"}),
+        ]
+
+        code_docs, text_docs = detect_code_files(docs)
+
+        assert len(code_docs) == 0
+        assert len(text_docs) == 2
+        for doc in text_docs:
+            assert doc.metadata["is_code"] is False
+
+    def test_get_language_from_extension(self):
+        """Test language detection from file extensions."""
+        assert get_language_from_extension("test.py") == "python"
+        assert get_language_from_extension("Test.java") == "java"
+        assert get_language_from_extension("component.tsx") == "typescript"
+        assert get_language_from_extension("Program.cs") == "csharp"
+        assert get_language_from_extension("document.txt") is None
+        assert get_language_from_extension("") is None
+
+
+class TestChunkingFunctions:
+    """Test various chunking functionality."""
+
+    def test_create_traditional_chunks(self):
+        """Test traditional text chunking."""
+        docs = [
+            MockDocument(
+                "This is a test document. It has multiple sentences. We want to test chunking."
+            )
+        ]
+
+        chunks = create_traditional_chunks(docs, chunk_size=50, chunk_overlap=10)
+
+        assert len(chunks) > 0
+        assert all(isinstance(chunk, str) for chunk in chunks)
+        assert all(len(chunk.strip()) > 0 for chunk in chunks)
+
+    def test_create_traditional_chunks_empty_docs(self):
+        """Test traditional chunking with empty documents."""
+        chunks = create_traditional_chunks([], chunk_size=50, chunk_overlap=10)
+        assert chunks == []
+
+    @pytest.mark.skipif(
+        os.environ.get("CI") == "true",
+        reason="Skip astchunk tests in CI - dependency may not be available",
+    )
+    def test_create_ast_chunks_with_astchunk_available(self):
+        """Test AST chunking when astchunk is available."""
+        python_code = '''
+def hello_world():
+    """Print hello world message."""
+    print("Hello, World!")
+
+def add_numbers(a, b):
+    """Add two numbers and return the result."""
+    return a + b
+
+class Calculator:
+    """A simple calculator class."""
+
+    def __init__(self):
+        self.history = []
+
+    def add(self, a, b):
+        result = a + b
+        self.history.append(f"{a} + {b} = {result}")
+        return result
+'''
+
+        docs = [MockDocument(python_code, "/test/calculator.py", {"language": "python"})]
+
+        try:
+            chunks = create_ast_chunks(docs, max_chunk_size=200, chunk_overlap=50)
+
+            # Should have multiple chunks due to different functions/classes
+            assert len(chunks) > 0
+            assert all(isinstance(chunk, str) for chunk in chunks)
+            assert all(len(chunk.strip()) > 0 for chunk in chunks)
+
+            # Check that code structure is somewhat preserved
+            combined_content = " ".join(chunks)
+            assert "def hello_world" in combined_content
+            assert "class Calculator" in combined_content
+
+        except ImportError:
+            # astchunk not available, should fall back to traditional chunking
+            chunks = create_ast_chunks(docs, max_chunk_size=200, chunk_overlap=50)
+            assert len(chunks) > 0  # Should still get chunks from fallback
+
+    def test_create_ast_chunks_fallback_to_traditional(self):
+        """Test AST chunking falls back to traditional when astchunk is not available."""
+        docs = [MockDocument("def test(): pass", "/test/script.py", {"language": "python"})]
+
+        # Mock astchunk import to fail
+        with patch("chunking.create_ast_chunks"):
+            # First call (actual test) should import astchunk and potentially fail
+            # Let's call the actual function to test the import error handling
+            chunks = create_ast_chunks(docs)
+
+            # Should return some chunks (either from astchunk or fallback)
+            assert isinstance(chunks, list)
+
+    def test_create_text_chunks_traditional_mode(self):
+        """Test text chunking in traditional mode."""
+        docs = [
+            MockDocument("def test(): pass", "/test/script.py"),
+            MockDocument("This is regular text.", "/test/doc.txt"),
+        ]
+
+        chunks = create_text_chunks(docs, use_ast_chunking=False, chunk_size=50, chunk_overlap=10)
+
+        assert len(chunks) > 0
+        assert all(isinstance(chunk, str) for chunk in chunks)
+
+    def test_create_text_chunks_ast_mode(self):
+        """Test text chunking in AST mode."""
+        docs = [
+            MockDocument("def test(): pass", "/test/script.py"),
+            MockDocument("This is regular text.", "/test/doc.txt"),
+        ]
+
+        chunks = create_text_chunks(
+            docs,
+            use_ast_chunking=True,
+            ast_chunk_size=100,
+            ast_chunk_overlap=20,
+            chunk_size=50,
+            chunk_overlap=10,
+        )
+
+        assert len(chunks) > 0
+        assert all(isinstance(chunk, str) for chunk in chunks)
+
+    def test_create_text_chunks_custom_extensions(self):
+        """Test text chunking with custom code file extensions."""
+        docs = [
+            MockDocument("function test() {}", "/test/script.js"),  # Not in default extensions
+            MockDocument("Regular text", "/test/doc.txt"),
+        ]
+
+        # First without custom extensions - should treat .js as text
+        chunks_without = create_text_chunks(docs, use_ast_chunking=True, code_file_extensions=None)
+
+        # Then with custom extensions - should treat .js as code
+        chunks_with = create_text_chunks(
+            docs, use_ast_chunking=True, code_file_extensions=[".js", ".jsx"]
+        )
+
+        # Both should return chunks
+        assert len(chunks_without) > 0
+        assert len(chunks_with) > 0
+
+
+class TestIntegrationWithDocumentRAG:
+    """Integration tests with the document RAG system."""
+
+    @pytest.fixture
+    def temp_code_dir(self):
+        """Create a temporary directory with sample code files."""
+        with tempfile.TemporaryDirectory() as temp_dir:
+            temp_path = Path(temp_dir)
+
+            # Create sample Python file
+            python_file = temp_path / "example.py"
+            python_file.write_text('''
+def fibonacci(n):
+    """Calculate fibonacci number."""
+    if n <= 1:
+        return n
+    return fibonacci(n-1) + fibonacci(n-2)
+
+class MathUtils:
+    @staticmethod
+    def factorial(n):
+        if n <= 1:
+            return 1
+        return n * MathUtils.factorial(n-1)
+''')
+
+            # Create sample text file
+            text_file = temp_path / "readme.txt"
+            text_file.write_text("This is a sample text file for testing purposes.")
+
+            yield temp_path
+
+    @pytest.mark.skipif(
+        os.environ.get("CI") == "true",
+        reason="Skip integration tests in CI to avoid dependency issues",
+    )
+    def test_document_rag_with_ast_chunking(self, temp_code_dir):
+        """Test document RAG with AST chunking enabled."""
+        with tempfile.TemporaryDirectory() as index_dir:
+            cmd = [
+                sys.executable,
+                "apps/document_rag.py",
+                "--llm",
+                "simulated",
+                "--embedding-model",
+                "facebook/contriever",
+                "--embedding-mode",
+                "sentence-transformers",
+                "--index-dir",
+                index_dir,
+                "--data-dir",
+                str(temp_code_dir),
+                "--enable-code-chunking",
+                "--query",
+                "How does the fibonacci function work?",
+            ]
+
+            env = os.environ.copy()
+            env["HF_HUB_DISABLE_SYMLINKS"] = "1"
+            env["TOKENIZERS_PARALLELISM"] = "false"
+
+            try:
+                result = subprocess.run(
+                    cmd,
+                    capture_output=True,
+                    text=True,
+                    timeout=300,  # 5 minutes
+                    env=env,
+                )
+
+                # Should succeed even if astchunk is not available (fallback)
+                assert result.returncode == 0, f"Command failed: {result.stderr}"
+
+                output = result.stdout + result.stderr
+                assert "Index saved to" in output or "Using existing index" in output
+
+            except subprocess.TimeoutExpired:
+                pytest.skip("Test timed out - likely due to model download in CI")
+
+    @pytest.mark.skipif(
+        os.environ.get("CI") == "true",
+        reason="Skip integration tests in CI to avoid dependency issues",
+    )
+    def test_code_rag_application(self, temp_code_dir):
+        """Test the specialized code RAG application."""
+        with tempfile.TemporaryDirectory() as index_dir:
+            cmd = [
+                sys.executable,
+                "apps/code_rag.py",
+                "--llm",
+                "simulated",
+                "--embedding-model",
+                "facebook/contriever",
+                "--index-dir",
+                index_dir,
+                "--repo-dir",
+                str(temp_code_dir),
+                "--query",
+                "What classes are defined in this code?",
+            ]
+
+            env = os.environ.copy()
+            env["HF_HUB_DISABLE_SYMLINKS"] = "1"
+            env["TOKENIZERS_PARALLELISM"] = "false"
+
+            try:
+                result = subprocess.run(cmd, capture_output=True, text=True, timeout=300, env=env)
+
+                # Should succeed
+                assert result.returncode == 0, f"Command failed: {result.stderr}"
+
+                output = result.stdout + result.stderr
+                assert "Using AST-aware chunking" in output or "traditional chunking" in output
+
+            except subprocess.TimeoutExpired:
+                pytest.skip("Test timed out - likely due to model download in CI")
+
+
+class TestErrorHandling:
+    """Test error handling and edge cases."""
+
+    def test_text_chunking_empty_documents(self):
+        """Test text chunking with empty document list."""
+        chunks = create_text_chunks([])
+        assert chunks == []
+
+    def test_text_chunking_invalid_parameters(self):
+        """Test text chunking with invalid parameters."""
+        docs = [MockDocument("test content")]
+
+        # Should handle negative chunk sizes gracefully
+        chunks = create_text_chunks(
+            docs, chunk_size=0, chunk_overlap=0, ast_chunk_size=0, ast_chunk_overlap=0
+        )
+
+        # Should still return some result
+        assert isinstance(chunks, list)
+
+    def test_create_ast_chunks_no_language(self):
+        """Test AST chunking with documents missing language metadata."""
+        docs = [MockDocument("def test(): pass", "/test/script.py")]  # No language set
+
+        chunks = create_ast_chunks(docs)
+
+        # Should fall back to traditional chunking
+        assert isinstance(chunks, list)
+        assert len(chunks) >= 0  # May be empty if fallback also fails
+
+    def test_create_ast_chunks_empty_content(self):
+        """Test AST chunking with empty content."""
+        docs = [MockDocument("", "/test/script.py", {"language": "python"})]
+
+        chunks = create_ast_chunks(docs)
+
+        # Should handle empty content gracefully
+        assert isinstance(chunks, list)
+
+
+if __name__ == "__main__":
+    pytest.main([__file__, "-v"])

tests/test_document_rag.py

@@ -57,6 +57,51 @@ def test_document_rag_simulated(test_data_dir):
         assert "This is a simulated answer" in output
 
 
+@pytest.mark.skipif(
+    os.environ.get("CI") == "true",
+    reason="Skip AST chunking tests in CI to avoid dependency issues",
+)
+def test_document_rag_with_ast_chunking(test_data_dir):
+    """Test document_rag with AST-aware chunking enabled."""
+    with tempfile.TemporaryDirectory() as temp_dir:
+        # Use a subdirectory that doesn't exist yet to force index creation
+        index_dir = Path(temp_dir) / "test_ast_index"
+        cmd = [
+            sys.executable,
+            "apps/document_rag.py",
+            "--llm",
+            "simulated",
+            "--embedding-model",
+            "facebook/contriever",
+            "--embedding-mode",
+            "sentence-transformers",
+            "--index-dir",
+            str(index_dir),
+            "--data-dir",
+            str(test_data_dir),
+            "--enable-code-chunking",  # Enable AST chunking
+            "--query",
+            "What is Pride and Prejudice about?",
+        ]
+
+        env = os.environ.copy()
+        env["HF_HUB_DISABLE_SYMLINKS"] = "1"
+        env["TOKENIZERS_PARALLELISM"] = "false"
+
+        result = subprocess.run(cmd, capture_output=True, text=True, timeout=600, env=env)
+
+        # Check return code
+        assert result.returncode == 0, f"Command failed: {result.stderr}"
+
+        # Verify output
+        output = result.stdout + result.stderr
+        assert "Index saved to" in output or "Using existing index" in output
+        assert "This is a simulated answer" in output
+
+        # Should mention AST chunking if code files are present
+        # (might not be relevant for the test data, but command should succeed)
+
+
 @pytest.mark.skipif(not os.environ.get("OPENAI_API_KEY"), reason="OpenAI API key not available")
 @pytest.mark.skipif(
     os.environ.get("CI") == "true", reason="Skip OpenAI tests in CI to avoid API costs"