How yiannislevy/bloodpanel works
Project type and purpose
• This is a full-stack web application (“Blood Test OCR & AI Parser”) that lets users upload PDF blood tests, runs OCR and an LLM parser, then stores and displays structured results in a dashboard.
File structure
• Top-level dirs:
• backend/ (FastAPI service with routers, models, utils, DB setup)
• frontend/ (React SPA with components, API wrapper, utils)
• alembic/ & alembic.ini (DB migration config)
• Config and entry points: requirements.txt, package.json, backend/main.py, frontend/src/index.js
Extract README insights
• Key workflow: PDF upload → pdfplumber/pdf2image OCR → OpenAI GPT-4o parsing → Pydantic validation → PostgreSQL via SQLAlchemy → React dashboard
• Tech stack: React + react-dropzone, FastAPI, Pydantic, SQLAlchemy, Alembic, OpenAI GPT-4o, pdfplumber, pdf2image
• Blue for frontend/UI
• Green for backend internal modules
• Orange for external services (OpenAI)
• Grey for database
Detailed coverage
• DB migration path (Alembic) in the backend block.
• Pydantic as the glue between raw LLM output and ORM models.
• Routers (sessions, upload) are called out in the API layer.