What do these changes do

Adds AST-based analysis tooling to audit type hint coverage across the codebase. Provides actionable data on 720 functions/methods categorized by completeness, scope, internal references, test suite mapping, priority, and mypy error counts.

Deliverables:

• build_tools/analysis/type_hint_analyzer.py - Main analyzer using Python AST with mypy integration and command-line arguments
• build_tools/analysis/generate_csv.py - CSV report generator with command-line arguments
• build_tools/analysis/output/ - Output directory containing JSON and CSV data files (submodule stats with mypy errors, functions without/incomplete hints)
• TYPE_HINT_COMPLETE_REPORT.md - Executive summary
• TYPE_HINT_ANALYSIS.md - Detailed technical breakdown
• TYPE_HINT_QUICKSTART.md - Quick reference
• .markdownlintrc - Markdown linting configuration

Key Metrics:

• 82.22% functions have complete type hints (592/720)
• 7.78% incomplete (56/720)
• 10.00% no hints (72/720)
• 9 submodules at 100% coverage
• No high-priority gaps (all critical public APIs typed)
• Mypy errors range from 6-9 per submodule (most have 6)

Priority Classification:

• Maps functions to test suites (core/compact/extra/noauto)
• Counts internal references (991 to 0)
• Assigns priority based on scope + usage + test coverage
• Identifies 94 medium-priority items for improvement

Mypy Integration:

• Runs mypy on each submodule to count type-related errors
• Errors included in console output, JSON data, CSV files, and markdown reports
• Provides complementary view: type hint completeness vs. type consistency

Local Execution Support:

• Scripts auto-detect repository root (no hardcoded paths)
• Command-line arguments for flexible configuration:
  • type_hint_analyzer.py --output-dir DIR (default: ./output)
  • generate_csv.py --input FILE --output-dir DIR
• No dependency on /tmp directory
• Works seamlessly on Windows, macOS, and Linux

Submodules needing attention: transliterate (48%), classify (20%), el (20%), chat (25%), wangchanberta (44%), coref (40%)

What was wrong

No visibility into type hint coverage across the codebase. Unable to systematically identify gaps, prioritize type safety improvements, or track mypy errors per submodule. Previous analysis tools were not suitable for local development (hardcoded paths, /tmp dependency).

How this fixes it

Implements comprehensive analysis that:

1. Parses all Python files using ast module
2. Analyzes function signatures for parameter and return type hints
3. Categorizes by completeness (complete/incomplete/none)
4. Maps to project structure (submodule, scope, test suite)
5. Calculates internal reference frequency via text search
6. Assigns priority (high: public + >10 refs + core/compact tests; medium: public + 3-10 refs; low: private or rare)
7. Runs mypy on each submodule to count type-related errors
8. Generates reports in markdown, CSV, and JSON formats
9. All markdown reports validated with markdownlint for proper formatting
10. Auto-detects repository root for portable execution across different environments
11. Accepts command-line arguments for flexible output configuration
12. Outputs to build_tools/analysis/output/ directory

Analysis is repeatable via python3 build_tools/analysis/type_hint_analyzer.py from any location in the repository.

Your checklist for this pull request

• Passed code styles and structures
• Passed code linting checks and unit test
• All markdown files pass markdownlint validation
• Scripts work on local machines without hardcoded paths
• Command-line arguments implemented with help documentation

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.