How minhthinhh/chatbot_work_ai works
Project type and purpose
• This is a single‐page web “chatbot” application built in Python with Streamlit.
• It ingests a corpus of PDF documents, processes them into a vector store, and answers user chat queries by retrieving relevant chunks and calling the OpenAI API.
File structure
• Top‐level files:
• chat_app.py – Streamlit entry point (UI + controller)
• Dockerfile & docker-compose.yml – containerization and orchestration
• requirements.txt – Python dependencies
• data/ – raw PDF documents
• utils/ – core pipeline modules:
– doc_handler.py (PDF parsing & chunking)
– build_graph.py (optional graph or chunk indexing)
– retriever_pipeline.py (vectorization, retrieval logic)
README for additional insights
• Three run modes: direct Streamlit, docker-compose, or raw Docker CLI.
• Expects an OPENAI_API_KEY.
• Exposes UI on port 8501.
Diagramming recommendations
• Frontend/UI (blue)
• Application logic (green)
• Data storage (yellow)
• External service (orange)