Refactor corpus I/O, parsers, and pipeline orchestration - #6

Open

apodobe wants to merge 11 commits into

mainfrom

cursor/refactor-corpus-pipeline-753f

Commits on Aug 11, 2026

Add atomic JSON I/O and fix lab row append dedup skip

cursoragent
and
apodobe
committed
7a07d66
refactor: pass patient_dob explicitly instead of global in local_structure_pdfs

cursoragent
and
apodobe
committed
8304ae3
Add shared time_util and cli_args utilities for corpus pipeline

cursoragent
and
apodobe
committed
8a5481f
Fix double parsing in merge_labs_corpus and add medsi/emias dedup test

cursoragent
and
apodobe
committed
f22afe0
Replace subprocess pipeline steps with direct function calls

cursoragent
and
apodobe
committed
f793670
Add unified vendor_registry module and consolidate detection logic

cursoragent
and
apodobe
committed
78b36b2
Group lab dedup by LOINC when available, fallback to canonical_key

cursoragent
and
apodobe
committed
366658b
Replace print() with logging in orchestration layer

cursoragent
and
apodobe
committed
d600563
Add ruff and mypy to dev deps and CI with minimal strictness

cursoragent
and
apodobe
committed
80d07e0
refactor: split local_structure_pdfs into medbots/parsers package

cursoragent
and
apodobe
committed
00a3db9
Fix ruff/mypy after parser split: pyproject ignores and meta typing

cursoragent
and
apodobe
committed
adc7034