490329dc66
This commit provides a minimal, focused fix for CI hanging issues by addressing the root causes: **Key Changes:** 1. **ZMQ Resource Management:** - Remove `context.term()` calls that were causing hangs - Add `socket.setsockopt(zmq.LINGER, 0)` to prevent blocking on close - Keep socket operations simple with default timeouts (no artificial limits) 2. **Process Cleanup:** - Add timeout (1s) to final `process.wait()` in embedding server manager - Prevent infinite waiting that was causing CI hangs 3. **Resource Cleanup Methods:** - Add simple `cleanup()` methods to searchers and API classes - Focus on C++ object destruction for DiskANN backend - Avoid complex cleanup logic that could introduce new issues 4. **Basic Test Safety:** - Simple pytest-timeout configuration (300s) - Basic test session cleanup using psutil - Minimal conftest.py without complex logic **Philosophy:** This solution avoids the complex multi-layered fixes from the previous PR chain. Instead, it targets the specific root causes: - ZMQ context termination blocking - Process wait() without timeout - C++ resource leaks in backends 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
LEANN Tests
This directory contains automated tests for the LEANN project using pytest.
Test Files
test_readme_examples.py
Tests the examples shown in README.md:
- The basic example code that users see first (parametrized for both HNSW and DiskANN backends)
- Import statements work correctly
- Different backend options (HNSW, DiskANN)
- Different LLM configuration options (parametrized for both backends)
- All main README examples are tested with both HNSW and DiskANN backends using pytest parametrization
test_basic.py
Basic functionality tests that verify:
- All packages can be imported correctly
- C++ extensions (FAISS, DiskANN) load properly
- Basic index building and searching works for both HNSW and DiskANN backends
- Uses parametrized tests to test both backends
test_document_rag.py
Tests the document RAG example functionality:
- Tests with facebook/contriever embeddings
- Tests with OpenAI embeddings (if API key is available)
- Tests error handling with invalid parameters
- Verifies that normalized embeddings are detected and cosine distance is used
test_diskann_partition.py
Tests DiskANN graph partitioning functionality:
- Tests DiskANN index building without partitioning (baseline)
- Tests automatic graph partitioning with
is_recompute=True - Verifies that partition files are created and large files are cleaned up for storage saving
- Tests search functionality with partitioned indices
- Validates medoid and max_base_norm file generation and usage
- Includes performance comparison between DiskANN (with partition) and HNSW
- Note: These tests are skipped in CI due to hardware requirements and computation time
Running Tests
Install test dependencies:
# Using extras
uv pip install -e ".[test]"
Run all tests:
pytest tests/
# Or with coverage
pytest tests/ --cov=leann --cov-report=html
# Run in parallel (faster)
pytest tests/ -n auto
Run specific tests:
# Only basic tests
pytest tests/test_basic.py
# Only tests that don't require OpenAI
pytest tests/ -m "not openai"
# Skip slow tests
pytest tests/ -m "not slow"
# Run DiskANN partition tests (requires local machine, not CI)
pytest tests/test_diskann_partition.py
Run with specific backend:
# Test only HNSW backend
pytest tests/test_basic.py::test_backend_basic[hnsw]
pytest tests/test_readme_examples.py::test_readme_basic_example[hnsw]
# Test only DiskANN backend
pytest tests/test_basic.py::test_backend_basic[diskann]
pytest tests/test_readme_examples.py::test_readme_basic_example[diskann]
# All DiskANN tests (parametrized + specialized partition tests)
pytest tests/ -k diskann
CI/CD Integration
Tests are automatically run in GitHub Actions:
- After building wheel packages
- On multiple Python versions (3.9 - 3.13)
- On both Ubuntu and macOS
- Using pytest with appropriate markers and flags
pytest.ini Configuration
The pytest.ini file configures:
- Test discovery paths
- Default timeout (600 seconds)
- Environment variables (HF_HUB_DISABLE_SYMLINKS, TOKENIZERS_PARALLELISM)
- Custom markers for slow and OpenAI tests
- Verbose output with short tracebacks
Known Issues
- OpenAI tests are automatically skipped if no API key is provided