How sagnikdebnath07/origamis-pro works
Project type and purpose
• This is a lightweight full-stack Python application built with Streamlit that implements a Retrieval-Augmented Generation (RAG) chatbot.
• The frontend (app.py) provides an interactive chat UI; the backend (chain.py) handles content loading, vector indexing (FAISS), and orchestration of an OpenAI language model via LangChain.
File structure
• Top level:
• .devcontainer/ – defines the development container settings.
• README.md – project overview and instructions.
• requirements.txt – Python dependencies.
• app.py – Streamlit frontend.
• chain.py – core retrieval and generation logic.
• No separate “frontend” or “backend” folders, but functionality is cleanly separated into two modules.
• Configuration files (.devcontainer.json, requirements.txt) sit alongside code.
README for additional insights
• Confirms a two-layer architecture: “Frontend(app.py)” vs. “Backend(chain.py).”
• Backend uses: WebBaseLoader → RecursiveCharacterTextSplitter → FAISS vector store → LangChain RetrievalChain → ChatOpenAI model.
• OpenAI API key is provided via Streamlit secrets.
• Conversation state is maintained via Streamlit’s session state.