Bug fixes (p-value mapping, None crashes), security (HTTPS for USPS API), CI improvements (pytest, Python 3.11-3.13 matrix), pylint 9.49 to 9.98, PyPI publishing, badges, LICENSE file, SEO-optimized docs

What's New

Comprehensive Test Suite

• 316 tests covering all modules: preprocessing (MAX & TAF), adapted algorithms, filters, common utilities
• All tests use synthetic data fixtures - no real claims data needed
• Previously skipped tests (BETOS 2020 crosswalk, low-value care) now run with synthetic data

Code Quality

• 0 flake8 warnings across entire codebase (fixed 163 pre-existing issues)
• pylint score: 9.94/10 - fixed unused imports, f-string logging, missing timeouts, unnecessary lambdas, param doc mismatches
• PEP 8 compliant throughout

Bug Fixes

• Fixed BETOS betpuf20.txt parsing - restructured reader to handle space-separated fields with quoted descriptions
• Fixed on_bad_lines parameter for pandas 2.0+ compatibility in BETOS code lookup files
• Fixed transform(max) to transform("max") eliminating 438 FutureWarnings

Performance Improvements

• Deduplicated pd.to_numeric() calls in flag_common_exclusions (MAX IP & OT)
• Merged consecutive map_partitions() calls in MAX PS eligibility processing
• Consolidated multiple .assign() calls into single operations

Documentation

• Comprehensive README with usage examples, expected data layout, algorithm references, and package structure
• SEO-optimized package description

CI/CD

• GitHub Actions updated to Python 3.13 with latest action versions

Dependencies

• Python >= 3.11 (tested on 3.13)
• Updated to pandas 2.0+, dask 202410+, pyarrow 15.0.0+

v2.0.0

Changes

• Python 3.12 support
• Pandas 2.0 compatibility updates
• OB/GYN module with ICD-10 delivery outcome codes
• Admitting diagnosis code column added to cleaning routines
• Improved repartitioning for cohort file exports using dask
• Beneficiary lookup speedup for cohort exports

v1.0.0

Initial stable release of medicaid-utils.

Features

• MAX file preprocessing (IP, OT, PS, CC)
• TAF file preprocessing (IP, OT, PS, RX, LT)
• Adapted algorithms: Elixhauser, PMCA, CDPS-MRx, NYU Billings ED, ED PQI, BETOS, Low Value Care
• Diagnosis and procedure code filtering
• Cohort extraction utilities
• Statistical analysis utilities