FDA Drug Label Ingestion & Parsing

Project Summary

This project ingests unstructured drug labeling data from the DailyMed public API, extracts meaningful sections, and structures it into clean JSON for downstream NLP/ML use.

Tech Stack

• Python
• requests, lxml, json, spacy
• Basic CLI orchestration

Workflow

1. ingest.py: Downloads HTML files using the DailyMed SPL web service.
2. clean.py: Parses HTML sections using LXML and extracts clinical sections like "INDICATIONS", "WARNINGS", etc.
3. nlp.py: analyzes sections and creates entity dicts.

Example Output

{
  "INDICATIONS AND USAGE": "This medication is used for...",
  "WARNINGS": "Do not use if you are allergic to...",
  ...
}