Add advanced code complexity and git history metrics #1
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Enhance cost estimation with additional metrics beyond COCOMO and SLOC: - Add Halstead complexity metrics (volume, difficulty, effort, bugs) - Add git history analysis (commits, contributors, age, releases, bus factor) - Add Maintainability Index calculation (Microsoft's formula) - Add maturity multiplier based on project age and scale - Add Halstead-based complexity multiplier - Update COCOMO estimator to incorporate all new multipliers - Add comprehensive unit tests for all new analyzers - Add end-to-end integration tests These metrics provide more accurate cost estimates for mature projects by accounting for repository history, code complexity, and maintainability.
Extend Halstead complexity metrics to support 12 programming languages beyond Python, using tree-sitter for universal AST parsing: - Add tree-sitter and tree-sitter-languages dependencies - Pin tree-sitter to 0.21.x for compatibility - Implement language detection via file extensions - Add operator/operand type mappings for each language - JavaScript, TypeScript, Java, C, C++, C#, Go, Rust, PHP, Ruby, Swift - Maintain graceful fallback to Python AST when tree-sitter unavailable - Add 7 new multi-language test cases - Language detection tests - JavaScript, TypeScript, Java, Go, Rust file analysis tests - Multi-language directory analysis tests - Fix test for invalid syntax (tree-sitter is resilient to errors) - Update CHANGELOG with multi-language feature details All 94 tests passing with improved language coverage.
The tree-sitter-languages package doesn't support Python 3.13 yet. Move tree-sitter dependencies to optional extras for graceful degradation: - Create [multilang] optional dependency group for tree-sitter - Restrict tree-sitter to Python <3.13 via environment markers - Add Python 3.13 to classifiers (now officially supported) - Update test workflow to test on Python 3.10-3.13 - Conditionally install multilang extras on Python <3.13 in CI - Update CHANGELOG to document optional dependency Without [multilang] installed: - Python-only Halstead analysis works via built-in AST - All other features work normally - Graceful fallback already implemented via TREE_SITTER_AVAILABLE flag With [multilang] installed (Python 3.10-3.12): - Full multi-language Halstead support for 12 languages - JavaScript, TypeScript, Java, C, C++, C#, Go, Rust, PHP, Ruby, Swift Installation: pip install ossval[multilang]
Fix CI failures on Python 3.13 where tree-sitter is not available: - Add pytest.skipif decorators to multi-language tests - Skip JavaScript, TypeScript, Java, Go, Rust tests when tree-sitter unavailable - Skip multi-language directory test when tree-sitter unavailable - Update invalid syntax test to handle both scenarios: - With tree-sitter: Returns metrics (resilient parsing) - Without tree-sitter: Returns None (AST fails on invalid syntax) - Import TREE_SITTER_AVAILABLE flag for test conditionals Test results: - With tree-sitter: 15 tests pass (9 Python + 6 multi-language) - Without tree-sitter: 9 tests pass, 6 tests skip (expected behavior) - Full suite: 94 tests pass on all Python versions This ensures CI passes on Python 3.13 without tree-sitter while maintaining full test coverage on Python 3.10-3.12 with [multilang] extras installed.
The workflow was manually installing pytest/ruff/black instead of using the [dev] extras group, which meant pytest-cov was missing. This caused pytest to fail since pyproject.toml has coverage options in addopts. - Install [dev] extras instead of individual packages - This ensures pytest-cov, mypy, and all other dev tools are installed - Aligns PR validation with the main test workflow
Consolidate CI workflows to avoid redundant test execution: Before: - test.yml: 8 test jobs (4 Python versions × 2 OS) on PR to main - pr-validation.yml: 1 test job (Python 3.13) on all PRs - Total: 9 test runs for PRs to main (duplicated) After: - test.yml: Comprehensive testing (multiple versions/OS) on PR to main - pr-validation.yml: Quick validation (linting, formatting, docs only) - Total: 8 test runs (no duplication) Changes: - Rename job from 'test-and-lint' to 'lint' - Remove pytest test execution from pr-validation.yml - Remove [dev] extras installation (only need ruff and black) - Keep black formatting and ruff linting checks - Keep documentation file verification Tests are still run comprehensively via test.yml which covers: - Python 3.10, 3.11, 3.12, 3.13 - Ubuntu and macOS - With and without tree-sitter (multilang extras)
oscarvalenzuelab
assigned
oscarvalenzuelab and
diegojorquera and unassigned
oscarvalenzuelab
oscarvalenzuelab
added
bug
diegojorquera
approved these changes
Dec 9, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
This PR enhances the cost estimation model with advanced metrics beyond COCOMO and SLOC:
Problem Solved
The previous model significantly underestimated costs for mature, complex projects (e.g., ReactJS estimated at only $16K). These new metrics account for:
Test Plan