Akshanshsensei/PDF-Constrained-Conversational-Agent
revert: remove FastAPI wrapper to use native Gradio UI and ?view=api docs
feat: wrap Gradio in FastAPI to expose Swagger UI at /docs
chore: remove dev-only utility scripts and binary artifacts
ci: trigger HF Space rebuild with requirements.txt touch
ci: force HF Space rebuild
style: remove all emojis from UI strings and tool status messages
fix: re-add type='messages' to gr.Chatbot for Gradio 5.33.0 tuples format error
fix: remove type='messages' from gr.Chatbot for Gradio 5.33.0 compatibility
feat: Add Chain-of-Thought thinking accordion to LLM streaming output
fix: get_page_count tool should return true page count from metadata, not non-empty page length
fix: revert gradio 6.0 theme parameter back to gr.Blocks for 5.33.0 compatibility
feat: Add Agentic Tool-Use layer for precise factual calculations
Fix: upgrade Gradio SDK to 5.33.0, disable API schema endpoint, sync upload script
Clean up README by removing unnecessary metadata
chore: track pdf with git lfs
chore: add Hugging Face Spaces configuration README
final push
chore: bind gradio to 0.0.0.0 for huggingface spaces
feat(v1): index persistence, genai SDK migration, page citations, recursive memory, multilingual support
feat: V1 upgrade - switch PDF parser from pypdf to PyMuPDF
fix: 4 pre-submission critical fixes from code review
feat: v0 baseline - hybrid RAG pipeline with anti-hallucination
chore: add venv run script
feat: add Gradio UI and main application entry point
feat(agent): add constrained LLM generator with streaming
feat(agent): add Redis-backed conversational memory with sliding window
feat(core): add CrossEncoderReranker for high-precision chunk selection
feat(core): add HybridRetriever with RRF and hallucination threshold
feat(core): add BM25 lexical retriever
feat(core): add FAISSEmbedder with Gemini and local fallback
feat(core): add text chunker with sliding window
feat(core): add PDF parser with header/footer filtering
feat: initial project setup (config, logger, dependencies)
