- Cleaned up tools/__init__.py to remove imports for non-existent modules
- Restored server.py to proper state without platform tool endpoints
- Removed extra exception classes that aren't implemented yet
- Server now imports correctly (only missing fastmcp in dev environment)
The PyPI platform tools were implemented in separate git worktrees and need to be properly merged in the next step.
Add comprehensive PyPI publishing and account management functionality:
Features:
- upload_package_to_pypi: Upload distributions to PyPI/TestPyPI with safety checks
- check_pypi_credentials: Validate API tokens and credentials
- get_pypi_upload_history: View upload history for packages with statistics
- delete_pypi_release: Safe release deletion with dry-run and confirmation
- manage_pypi_maintainers: Add/remove/list package maintainers
- get_pypi_account_info: View account details, quotas, and limits
Implementation:
- Created pypi_query_mcp/tools/publishing.py with all 6 functions
- Added PyPIPublishingClient for authenticated API operations
- Comprehensive error handling with custom exceptions
- Full async/await patterns following existing codebase conventions
- Safety checks for destructive operations (deletion requires confirmation)
- Support for both production PyPI and TestPyPI
Integration:
- Added publishing-specific exceptions to core/exceptions.py
- Updated tools/__init__.py with publishing function imports
- Added 6 MCP server endpoints to server.py with proper error handling
- Created comprehensive tests in tests/test_publishing.py
Production-ready code with proper authentication, validation, and safety measures.
- Implemented PyPISearchClient with semantic search, filtering, and sorting
- Added 4 new search tools: search_packages, search_by_category, find_alternatives, get_trending_packages
- Created SearchFilter and SearchSort classes for flexible configuration
- Added SearchError exception for search-specific error handling
- Comprehensive test suite with 13 tests covering all search functionality
- Enhanced MCP server with 4 new search endpoints
- Support for filtering by Python version, license, category, downloads, maintenance status
- Multiple sorting options: relevance, popularity, quality, recency, name, downloads
- Semantic search using description similarity scoring
- Category-based package discovery with intelligent keyword matching
- Package alternatives finder using metadata analysis
- Trending packages analysis with download activity tracking
- Robust fallback mechanisms using curated package database
- All tests passing (13/13)
This implements feature #6 from the roadmap: "Advanced PyPI Search with filtering by Python version, license, maintenance status and sorting by popularity, recency, quality score with semantic search capabilities"
- Update all dependencies to latest versions (fastmcp, httpx, packaging, etc.)
- Downgrade click from yanked 8.2.2 to stable 8.1.7
- Fix code formatting and linting issues with ruff
- Most tests passing (2 test failures in dependency resolver need investigation)
- Implement parsing of development dependencies from extra markers
- Add comprehensive development keyword detection (20+ patterns)
- Enhance response format with dev dependency categorization
- Add development_optional_dependencies and provides_extra fields
- Test with pytest, setuptools, sphinx, wheel packages
- Add curated popular packages database with 100+ packages
- Implement GitHub API integration for real-time popularity metrics
- Create multi-tier fallback strategy (live API -> curated -> enhanced)
- Add period scaling and realistic download estimates
- Provide rich metadata with categories and descriptions
- Fix extra dependencies being filtered out by Python version checks
- Add proper handling for extra markers in dependency parsing
- Update parameter descriptions and documentation
- Add comprehensive examples and demo script
- Test with requests[socks], django[argon2,bcrypt], setuptools[test]
- Implement exponential backoff retry logic with jitter
- Add intelligent fallback mechanisms with realistic data estimates
- Enhance caching strategy with multi-tier validation (24hr + 7day TTL)
- Improve error handling and transparent user communication
- Add API health monitoring with consecutive failure tracking
- Fix version parameter being ignored - now properly fetches specified versions
- Enhance PyPIClient with version-specific URL construction
- Add version format validation with regex patterns
- Improve error handling for non-existent versions
- Test with Django 4.2.0, FastAPI 0.100.0, NumPy 1.20.0
- Add sort_versions_semantically function using packaging library
- Fix issue where pre-release versions appeared before stable (5.2rc1 vs 5.2.5)
- Handle edge cases: dev, post, invalid versions with graceful fallback
- Add comprehensive test suite covering all scenarios
- Maintain backward compatibility with existing functionality
- Fix blank line whitespace issues (W293) using ruff --unsafe-fixes
- Reformat code using ruff format for consistent styling
- Fix analyze_package_quality function to return list[Message] instead of string
- Add missing 'assessment' keyword to package analysis template
- Update tests to use real prompt functions instead of mocks for structure validation
- Fix import ordering in test files
- All 64 tests now pass with 47% code coverage
Signed-off-by: longhao <hal.long@outlook.com>
- Remove unused imports in stats_client.py and download_stats.py
- Fix import sorting in test files
- Remove unnecessary f-strings in server.py and demo script
- Clean up whitespace and formatting issues
- Ensure all files pass ruff and isort checks
Signed-off-by: longhao <hal.long@outlook.com>
- Add PyPIStatsClient for pypistats.org API integration
- Implement get_package_download_stats for recent download statistics
- Implement get_package_download_trends for time series analysis
- Implement get_top_packages_by_downloads for popularity rankings
- Add comprehensive MCP tools for download statistics
- Include download trends analysis with growth indicators
- Add repository information and metadata integration
- Provide comprehensive test coverage
- Add demo script and usage examples
- Update README with new features and examples
Signed-off-by: longhao <hal.long@outlook.com>
- Fix whitespace in docstrings and blank lines
- Remove unused variables in tests
- Rename unused loop variables to follow conventions
- All ruff checks now pass
Signed-off-by: Hal <hal.long@outlook.com>
- Add DependencyParser for parsing and categorizing package dependencies
- Add DependencyResolver for recursive dependency tree analysis
- Add PackageDownloader for downloading packages with dependencies
- Add resolve_dependencies MCP tool for comprehensive dependency analysis
- Add download_package MCP tool for package collection
- Support Python version filtering and extra dependencies
- Include comprehensive test coverage for new functionality
- Add demonstration script for new features
- Update README with new capabilities and usage examples
Signed-off-by: Hal <hal.long@outlook.com>
- Add ServerSettings class with pydantic-settings for type-safe configuration
- Support multiple PyPI mirror sources with priority-based fallback mechanism
- Implement RepositoryConfig and RepositoryManager for multi-repository support
- Add environment variable support for all configuration options
- Include private repository authentication configuration
- Add advanced dependency analysis settings (max depth, concurrency, security)
- Provide secure credential management with sensitive data masking
- Update documentation and configuration examples
- Add comprehensive test suite with 23 test cases covering all features
- Include demo script showcasing multi-mirror configuration capabilities
Configuration features:
- Primary, additional, and fallback index URLs
- Automatic duplicate URL removal with priority preservation
- Runtime configuration reloading
- Integration with repository manager for seamless multi-source queries
Signed-off-by: longhao <hal.long@outlook.com>
- Fix FastMCP server interface to use correct mcp.run() instead of app.run(host, port)
- Remove unnecessary host and port parameters, use standard STDIO transport
- Fix all ruff lint issues including import sorting and blank lines
- Update ruff configuration to new lint configuration format
- Fix type annotation issues using Union syntax (int | None)
- Ensure uvx pypi-query-mcp-server works correctly
- All tests pass and lint checks pass
