feat: add Streamlit UI for consent form compliance checking

[GovindhKishore]

Adds a Streamlit frontend that exposes the full compliance pipeline to end users.

What this adds

• PDF and TXT consent form upload
• LLM backend toggle - Gemini 2.5 Flash (local testing) and OpenAI GPT-4 (production)
• Detected study type and check count shown before LLM call
• Results split into Tier 1 universal checks and Tier 2 conditional checks with ✅ / ❌ verdict, reason, and GA4GH citation per check

Files changed

• src/ui/__init__.py - new
• src/ui/app.py - new
• src/compliance/checker.py - extract_text updated to accept BytesIO (Streamlit uploads) alongside file paths; run_compliance_check now accepts pre-extracted text directly and llm made optional to support Gemini backend

To run
streamlit run src/ui/app.py