feat: Set up comprehensive Python testing infrastructure

[llbbl]

Add Python Testing Infrastructure

Summary

This PR sets up a comprehensive testing infrastructure for the Mamba Chat project, providing developers with a ready-to-use testing environment that follows Python best practices.

Changes Made

Package Management

• Added Poetry as the package manager for development dependencies
• Created pyproject.toml with full Poetry configuration
• Kept production dependencies in requirements.txt (as they have CUDA-specific requirements)

Testing Framework

• Configured pytest as the main testing framework
• Added pytest-cov for code coverage reporting
• Added pytest-mock for mocking utilities
• Set up 80% coverage threshold with HTML and XML reporting

Configuration

• Test Discovery: Configured to find tests matching test_*.py or *_test.py patterns
• Coverage Settings:
  • Source tracking for the project
  • Exclusions for test files, virtual environments, and boilerplate code
  • Multiple report formats (terminal, HTML, XML)
• Custom Markers:
  • @pytest.mark.unit - Fast, isolated unit tests
  • @pytest.mark.integration - Tests requiring external resources
  • @pytest.mark.slow - Long-running tests (skipped by default)

Directory Structure

tests/
├── __init__.py
├── conftest.py          # Shared fixtures and configuration
├── test_setup_validation.py  # Validation tests
├── unit/
│   └── __init__.py
└── integration/
    └── __init__.py

Fixtures

Created comprehensive fixtures in conftest.py:

• temp_dir - Temporary directory for test files
• temp_file - Helper to create temporary files
• mock_config - Mock configuration object
• mock_model - Mock ML model
• mock_tokenizer - Mock tokenizer
• mock_dataset - Mock dataset
• isolated_filesystem - Run tests in isolated directory
• capture_logs - Capture log output during tests

Additional Setup

• Updated .gitignore with testing artifacts and Claude-specific entries
• Created validation tests to ensure the infrastructure works correctly
• Added support for --runslow flag to run slow tests

How to Use

Install Dependencies

poetry install

Run Tests

# Run all tests
poetry run pytest

# Run with coverage report
poetry run pytest --cov

# Run without coverage
poetry run pytest --no-cov

# Run only unit tests
poetry run pytest -m unit

# Run integration tests
poetry run pytest -m integration

# Include slow tests
poetry run pytest --runslow

# Run specific test file
poetry run pytest tests/test_setup_validation.py

View Coverage Reports

• Terminal: Shown automatically when running tests
• HTML: Open htmlcov/index.html in your browser
• XML: Available at coverage.xml for CI integration

Notes

1. Production Dependencies: The main project dependencies (torch, transformers, etc.) remain in requirements.txt due to their CUDA-specific requirements. This testing infrastructure is focused on development dependencies only.

2. Coverage Threshold: Set to 80% by default. Tests will fail if coverage drops below this threshold.

3. Slow Tests: Tests marked with @pytest.mark.slow are skipped by default to keep the test suite fast. Use --runslow to include them.

4. Package Structure: Created a minimal mamba_chat package directory to satisfy Poetry's requirements. The actual project code structure can be organized as needed.

Next Steps

Developers can now immediately start writing tests by:

1. Creating test files in tests/unit/ or tests/integration/
2. Using the provided fixtures from conftest.py
3. Marking tests appropriately with @pytest.mark.unit, @pytest.mark.integration, or @pytest.mark.slow