Blood test PDF parser (pdfplumber)

[EduardPetraeus]

Problem

Blood test results from GetTested (and other labs) are PDF-only. No structured data.

Proposed Solution

1. PDF parser using pdfplumber (better than tabula for structured tables)
2. Extract: biomarker name, value, unit, reference range, date
3. Normalize biomarker names to standard vocabulary
4. Output: parquet files in blood_tests/raw/ with date partitioning
5. Silver merge: merge_blood_test_results.sql
6. Support multiple lab formats (GetTested, Sundhed.dk)

Acceptance Criteria

• Parse GetTested PDF to structured parquet
• Biomarker names normalized
• Reference ranges extracted
• Silver table queryable with historical trends

[EduardPetraeus]

Session 1 of C1 done: pdfplumber PDF parser (GetTested + sundhed.dk formats), lab_ingestion.py, merge SQL, MCP query_lab_results tool, 72 tests. Sessions 2-3 (genetics) remaining.