Legal Document Summarization & GDPR Recital Tracker

Streamlit app that summarizes legal documents, surfaces risks/clauses, and tracks GDPR recitals with email + Google Sheets notifications.

Features

• Upload PDF/DOCX, get chunked LLM summaries (Groq) and risk/clauses detection.
• Visualizations: risk matrix, heatmap, distributions, and exportable PDF report.
• Q&A over the uploaded document via Groq.
• GDPR recital tracker with email + Sheets push.
• RAG CLI helper in rag_pipeline.py (HF Hub LLM + FAISS embeddings).

Quickstart

1. Clone and set up Python 3.10+.
2. Create a virtualenv and install:

   pip install -r requirements.txt

3. Copy .env.example to .env and fill in keys/secrets.
4. Run the Streamlit app:

   streamlit run app.py

5. Optional: RAG CLI

   python rag_pipeline.py --pdf path/to/file.pdf

Required Environment

• GROQ_API_KEY – Groq LLM key.
• HUGGINGFACEHUB_API_TOKEN – for HF Hub LLM (RAG path).
• Email: SENDER_EMAIL, EMAIL_PASS, RECEIVER_EMAIL, FEEDBACK_EMAIL (and EMAIL_ADDRESS/EMAIL_PASSWORD for tracker).
• Google Sheets: GOOGLE_CREDENTIALS_PATH, GOOGLE_SHEET_ID.

Project Structure

• app.py – Streamlit navigation.
• legal_document_analysis.py – document ingestion, summarization, risk visuals, PDF export, feedback.
• Update_tracker.py – GDPR recital tracker + email/Sheets.
• rag_pipeline.py – CLI RAG pipeline.
• requirements.txt – runtime deps; see requirements-dev.txt for tooling.

Development

• Lint/format: ruff and black (see requirements-dev.txt).
• Tests: pytest in tests/ (offline; no API calls).
• CI: basic GitHub Actions workflow at .github/workflows/ci.yml.

Notes

• No secrets are committed; keep credentials.json and .env out of git (see .gitignore).
• The PDF export now uses built-in fonts and cleans up temporary files.
• Network calls include timeouts and user-agent headers.

License

MIT License – see LICENSE.