PCOS Food Scanner (Python prototype)

Streamlit prototype: scan a barcode → get a PCOS-aware score (1–10) + serving advice + save/remove. Validates scoring logic before any mobile port.

Setup

cd ~/projects/pcos-scanner
python3 -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env   # fill in NVIDIA_API_KEY
streamlit run app.py

Layout

• app.py — Streamlit UI (Scan / Saved / Profile tabs)
• core/openfoodfacts.py — barcode → nutriments via Open Food Facts
• core/scoring.py — deterministic rule-based score
• core/personalize.py — NVIDIA OpenAI-compatible endpoint (https://integrate.api.nvidia.com/v1) adjusts score + writes explanation against user profile
• core/profile.py — load/save PCOS profile (4-type + symptom flags)
• core/db.py — Supabase CRUD for profile + saved foods + personalization cache
• migrations/ — SQL migrations auto-run on app startup

Deploy to Streamlit Cloud

Push this repo to GitHub. Go to https://share.streamlit.io and create a new app. Point the app at app.py. Add NVIDIA_API_KEY in Streamlit Secrets. Click Deploy.

Supabase setup

1. Create a Supabase project at https://supabase.com.
2. Copy the Project URL and anon key from Settings → API.
3. Copy the Postgres URI from Settings → Database → Connection String → URI mode (transaction pooler).
4. Add these to Streamlit Cloud Secrets:
   • SUPABASE_URL
   • SUPABASE_ANON_KEY
   • SUPABASE_DB_URL
   • SUPABASE_SERVICE_ROLE_KEY

On first deploy, Supabase will auto-run migrations on startup. To add a schema change later, drop a new file like migrations/002_add_column.sql and push — it runs automatically on the next redeploy.

See ~/.claude/plans/can-we-make-this-proud-quilt.md for full spec.