Skip to content

CHANGELOG.md

Latest commit

 

History

History
241 lines (192 loc) · 9.73 KB

CHANGELOG.md

File metadata and controls

241 lines (192 loc) · 9.73 KB

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.1.2] - 2025-10-07

Added

  • Release Infrastructure: Added tests/packaging/test_packaged_version.sh script to validate version consistency between pyproject.toml and installed package before deployment
  • CI/CD Integration: Updated .github/workflows/publish.yml to run packaging validation tests before PyPI publication
  • Developer Tooling: Added .claude/commands/release-prep.md slash command for streamlined release management workflow
  • Test Coverage: Added 26 new unit tests covering:
    • All exception class constructors and serialization (to_dict() method)
    • Data session error paths (validation methods, stats generation)
    • Version string format validation

Changed

  • Version Handling: Simplified _version.py by removing PackageNotFoundError fallback logic (package metadata always available in deployment environments)
  • Test Coverage: Improved overall coverage from 88.61% to 89.05%
    • _version.py: 71.43% → 100%
    • exceptions.py: 83.67% → 100%
    • data_session.py: 81.25% → 100%

Fixed

  • Version Validation: Packaging test now validates three-part semver format (MAJOR.MINOR.PATCH) with regex pattern matching

0.1.1 - 2025-10-06

Fixed

  • Version Reporting: Fixed _version.py to correctly report package version (0.1.1) instead of fallback "0.0.0" - now uses importlib.metadata.version("databeak") instead of __name__
  • Server Info: get_server_info and health_check MCP tools now correctly return actual package version

Changed

  • Version Management: Converted from dynamic versioning (via git tags) to static version in pyproject.toml for more reliable version control
  • Test Implementation: Updated version tests to read from pyproject.toml using tomllib parser instead of importlib.metadata for more direct source of truth

Added

  • Unit Tests: Added test_get_server_info_returns_actual_version unit test to validate version reporting
  • Integration Tests: Added 5 new integration tests in test_system_server_integration.py to verify version propagation through the full MCP client stack

0.1.0 - 2025-10-06

Fixed

  • Type Safety: Resolved 10 mypy type checking errors including missing type parameters, circular imports, and unused type ignore comments
  • Test Suite: Fixed 35 failing tests by correcting fixture names, adding dict-to-object conversion for FilterCondition, and handling None values in validators
  • Circular Import: Moved CellValue type alias to models/__init__.py to eliminate circular dependency between typed_dicts.py and data_models.py
  • Filter Operations: Added dict-to-FilterCondition conversion with == to = operator normalization for backward compatibility

Changed

  • Type Annotations: Added generic TypeVar T to apply_violation_limits for improved type safety
  • Pandera Integration: Fixed pandera.errors import path in validation_server.py
  • MyPy Configuration: Excluded examples/ directory from type checking to prevent false positives
  • Type Aliases: Migrated FilterValue to modern type statement syntax (PEP 695)
  • Pydantic Validators: Simplified validation logic in pydantic_validators.py, removed duplicate type hints, and streamlined field validator implementations

Removed

  • Dead Code: Removed unused session_service.py and statistics_service.py modules with dependency injection abstractions (0% actual usage)
  • Unused Models: Removed DataSchema and DataQualityRule Pydantic models (61 lines) that were only tested but never used in production code
  • Deprecated Enum: Removed OperationType enum that was marked for cleanup and had no active usage
  • Misleading Documentation: Removed docs/docker-deployment.md that referenced deleted Docker files
  • Example Code: Removed examples/dependency_injection_demo.py that referenced deleted modules
  • Try-Except Wrappers: Removed 58 unnecessary try-except wrapper blocks from MCP tool functions (TRY301 violations) - FastMCP handles exception transformation automatically
  • Transformation Service: Removed unused transformation_service.py module (676 lines, zero production usage)
  • Dead TypedDicts: Removed 22 unused TypedDict classes (343 lines, 88.7% reduction in typed_dicts.py)

0.0.4 - 2025-09-30

Added

  • Relaxed JSON Schema Validation: Custom jsonschema validator that accepts integers as strings or floats, handling LLM-generated parameters that don't strictly conform to integer types (PR #115)
  • Integration Testing Framework: Comprehensive integration testing with pytest config, isolated environment, and coverage reporting (PR #110)
  • ReadTheDocs Configuration: Official documentation site with MkDocs integration and automated deployment (PR #108)
  • Type Safety Enhancements: MCP-specific type definitions including NonNegativeIntString and PositiveIntString with regex validation (PR #111)

Changed

  • Header Type Architecture: Refactored HeaderConfig to use Abstract Base Class with Pydantic discriminated union pattern for cleaner polymorphic behavior (PR #113)
  • FastMCP Dependency: Updated to use released fastmcp version instead of development branch (PR #114)
  • Testing Configuration: Enhanced pytest configuration with integration test markers, coverage settings, and proper test isolation
  • Project Structure: Streamlined documentation and removed unmaintained agent files and Docker infrastructure (PR #116)

Removed

  • Docker Support: Removed Dockerfile and docker-compose.yml as Docker deployment is not officially supported
  • Agent Documentation: Removed specialized agent documentation files that are no longer actively maintained

0.0.3 - 2025-01-24

Fixed

  • Circular Import Resolution: Resolved critical circular import between session.py and session_service.py using lazy imports and TYPE_CHECKING blocks
  • Import Architecture: Converted all relative imports to absolute imports, eliminating 15 TID252 ruff violations
  • MyPy Configuration: Added mypy_path="src" to enable type checking from project root and pre-commit compatibility
  • Documentation Consistency: Updated all documentation to use standardized uv run mypy src/databeak/ command syntax

Changed

  • Agent Documentation: Streamlined test-coverage-analyzer agent description to focus on prescriptive guidance rather than current state information
  • Quality Pipeline: Pre-commit hooks now pass completely with proper MyPy integration

0.0.2 - 2025-01-19

Added

  • Production Quality Standards: Achieved zero ruff violations across all code quality categories
  • Enhanced Type Safety: 100% mypy compliance with comprehensive type annotations
  • Context-Based Logging: MCP-integrated logging for better traceability in tool functions
  • API Design Improvements: Keyword-only boolean parameters to eliminate boolean traps
  • Data Validation: Added proper validation constraints (e.g., negative value prevention)
  • Comprehensive Test Coverage: 1100+ unit tests with excellent coverage
  • Logging Guidelines: Clear documentation for Context vs standard logger usage

Changed

  • Architecture Simplification: Removed CorrelatedLogger, simplified to standard Python logging
  • Server Composition: Enhanced modular FastMCP server architecture
  • Exception Handling: Consistent exception message patterns with variable extraction
  • Parameter Naming: Resolved variable shadowing (format → export_format)
  • Test Structure: Improved pytest patterns with proper exception testing
  • Documentation: Updated all docs to reflect current architecture and quality standards

Fixed

  • Security: Eliminated silent exception handling (try-except-pass patterns)
  • Code Consistency: Fixed variable shadowing violations throughout codebase
  • Test Quality: Proper pytest.raises usage with specific match parameters
  • Error Messages: Extracted all exception messages to variables for maintainability
  • Import Organization: Cleaned up unused imports and dependencies

Removed

  • Code Violations: Eliminated all ruff violations (G004, FBT, EM, PT, A, ARG, S110, N818)
  • Boolean Traps: Removed confusing boolean positional parameters
  • Custom Logging Complexity: Simplified logging architecture
  • Duplicate Code: Consolidated AutoSaveConfig implementations
  • Outdated Documentation: Updated all references to current architecture

0.0.1 - 2025-09-18

Added

  • Initial release of DataBeak MCP Server
  • Core CSV operations: read, write, filter, transform
  • Data validation and quality checks
  • Statistical analysis and profiling capabilities
  • Outlier detection and handling
  • Support for multiple file formats (CSV, Excel, Parquet)
  • Async operations for high performance
  • Comprehensive error handling and logging
  • FastMCP integration for seamless AI assistant integration
  • Pandas-powered data manipulation
  • Full test coverage with pytest
  • Documentation with examples
  • Type hints and mypy compatibility