Skip to main content

Explanation

  • PDF-to-Text Extraction
  • Embedding + Vector Database (ChromaDB + Sentence Transformers)
  • RAG-based Q&A powered by HuggingFace Transformers
  • Feedback system with editable responses and star rating
  • SQLite backend for storing evaluations
  • Streamlit frontend

Usage Flow

  1. Select a predefined question or enter a custom one.
  2. The chatbot fetches relevant chunks from uploaded documents.
  3. The LLM generates a response using those chunks.
  4. You can view, edit, rate, and submit feedback.
  5. All feedback is stored in feedback.db.