PillSee je inovativní AI-powered chatovací platforma zaměřená na poskytování přesných a relevantních informací o léčivých přípravcích pro české uživatele. Projekt je součástí iniciativy DigiMedic pro digitalizaci zdravotnictví.

✅ Repository Consolidation: COMPLETE (2025-12-06)
✅ Phase 1 Implementation: READY
📋 Sprint 1 BMAD: Designed, awaiting implementation

PillSee/
├── 📂 docs/                    # BMAD Architecture Documentation
│   ├── project-brief.md        # Vision, stakeholders, budget €24K
│   ├── prd.md                  # 15 user stories, API specs
│   ├── architecture.md         # 6-agent system specification
│   ├── security-strategy.md    # GDPR/MDR compliance
│   └── phase1/                 # Phase 1 documentation
│
├── 📂 pillsee-backend/         # Complete LangChain Backend
│   ├── app/                    # Application code (~2,881 LOC)
│   │   ├── agents/            # LangChain agents
│   │   ├── memory/            # GDPR-compliant memory
│   │   ├── tools/             # 6 LangChain tools
│   │   ├── retrievers/        # Supabase vectorstore
│   │   └── ...
│   ├── migrations/            # Database migrations (6 files)
│   ├── main.py                # FastAPI server
│   └── README_BACKEND.md      # Backend documentation
│
├── 📄 CONSOLIDATION_REPORT.md  # Complete consolidation report
└── 📄 README.md                # This file

• AI Chatbot: Inteligentní konverzační rozhraní pro dotazy o lécích
• Integrace SÚKL: Přímé napojení na databázi Státního ústavu pro kontrolu léčiv
• RAG Architecture: Retrieval-Augmented Generation pro přesné odpovědi
• GDPR Compliance: Anonymizace PII, 90-denní data retention
• Lékové Interakce: Kontrola drug-drug interactions
• Cenové Srovnání: Affiliate links na e-shopy

• Framework: FastAPI + LangChain
• LLM: OpenAI GPT-4o-mini
• Database: PostgreSQL 17 + pgvector (Supabase)
• Embeddings: text-embedding-3-small (1536D)
• Memory: GDPR-compliant conversation history
• Docker: Complete containerization

• Orchestration: LangGraph multi-agent system
• Agents: 6-agent architecture (Supervisor, Triage, RAG Expert, Safety Monitor, Interaction Checker, Dosage Advisor)
• RAG Strategies: 6 strategies (hybrid, HyDE, multi-query, parent-doc, compression, ensemble)

• Framework: Next.js 14
• UI Library: shadcn/ui + Tailwind CSS
• AI SDK: Vercel AI SDK
• Type Safety: TypeScript

• Hosting Backend: Cloud Run
• Hosting Frontend: Vercel
• Database: Supabase (managed PostgreSQL)
• CDN: Cloudflare
• Monitoring: Sentry + Langsmith

6 kompletních migrací v pillsee-backend/migrations/:

1. Core Drug Tables - drug_info s SÚKL daty
2. SPC/PIL Vectors - Dokumenty s 1536D embeddings
3. Pricing Data - E-commerce ceny + affiliate odkazy
4. Drug Interactions - Interakce + WHO ATC klasifikace
5. RAG Vector Search - pgvector IVFFlat indexy + GDPR sessions
6. LangChain Tables - Phase 1 specifické tabulky

cd /Users/petrsovadina/Desktop/Develope/PillSee

cd pillsee-backend
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

cp .env.example .env
# Edituj .env:
# - OPENAI_API_KEY=...
# - SUPABASE_URL=...
# - SUPABASE_KEY=...

# Spusť migrace v pořadí 001-006
psql "postgresql://..." -f migrations/001_core_drug_tables.sql
# ... (viz pillsee-backend/README_BACKEND.md)

# Development
uvicorn main:app --reload

# Docker
docker-compose up

• Backend README: pillsee-backend/README_BACKEND.md
• Consolidation Report: CONSOLIDATION_REPORT.md
• Architecture Spec: docs/architecture.md
• Security Strategy: docs/security-strategy.md

• Phase 1 Summary: docs/phase1/PillSee_PHASE1_SUMMARY.md
• Deployment Guide: docs/phase1/PillSee_DEPLOYMENT_GUIDE.md
• Implementation Checklist: docs/phase1/PillSee_IMPLEMENTATION_CHECKLIST.md
• Quick Reference: docs/phase1/PillSee_QUICK_REFERENCE.md

• Project Brief: docs/project-brief.md - Vision, timeline, budget
• PRD: docs/prd.md - 15 user stories, API specifications
• Architecture: docs/architecture.md - Complete 6-agent system
• Security: docs/security-strategy.md - GDPR/MDR compliance

• LangChain backend s RAG
• 6 LangChain tools
• GDPR-compliant memory
• Supabase integration
• Docker setup
• Database migrations

• 6-agent LangGraph orchestration
• Triage Agent - query classification
• RAG Expert - 6 RAG strategies
• Safety Monitor - MDR validation
• Interaction Checker - drug-drug interactions
• Dosage Advisor - informational dosage

• Implement 6 RAG strategies
• GDPR/MDR utilities
• PII anonymization
• Comprehensive testing

• Next.js 14 frontend
• Vercel AI SDK streaming
• shadcn/ui components
• Real-time chat interface

• ✅ PII anonymization (email, phone, ID numbers)
• ✅ 90-day data retention
• ✅ Right to erasure
• ✅ Hashed IP addresses
• ✅ Audit logging

• ✅ No diagnostic claims
• ✅ Informational only disclaimers
• ✅ Safety warnings
• ✅ Compliance audit trails

Pro přispěvání do projektu:

1. Přečti docs/architecture.md pro pochopení systému
2. Zkontroluj CONSOLIDATION_REPORT.md pro aktuální stav
3. Následuj coding standards v pillsee-backend/

Pro otázky viz:

• Backend docs: pillsee-backend/README_BACKEND.md
• Architecture: docs/architecture.md
• Consolidation report: CONSOLIDATION_REPORT.md

Project Status: ✅ Phase 1 Complete, Ready for Sprint 1
Last Updated: 2025-12-06
Version: 1.0.0-phase1
License: Proprietary