75-langchain-chatbot

75 - Generative AI Chatbot (LangChain)

What It Does

Builds an AI-powered question-answering chatbot using the LangChain framework. The bot uses a Retrieval-Augmented Generation (RAG) pattern: it loads a knowledge base of text documents, splits them into chunks, stores them in a vector database (FAISS), and retrieves the most relevant context before generating an answer. Includes a ChatGPT-style frontend.

Project Structure

75-langchain-chatbot/
  main.py            # FastAPI entry point with chat endpoint
  bot.py             # LangChain chain construction (retriever + LLM)
  knowledge.py       # Knowledge base loader and text splitter
  config.py          # LLM model config and API key placeholder
  data/
    knowledge_base.txt  # Mock knowledge documents
  requirements.txt   # Dependencies
  index.html         # Unified ChatGPT-style frontend
  README.md          # This file

Setup and Run

1. Install dependencies

pip install -r requirements.txt

2. Configure your API key

Edit config.py and set your OpenAI API key (or use a local model).

3. Start the server

uvicorn main:app --reload

4. Open the Frontend

Open index.html in your browser to start chatting.

Example Output

// POST /chat  {"question": "What is Python used for?"}
{
  "question": "What is Python used for?",
  "answer": "Python is a versatile programming language used for web development, data science, machine learning, automation, scripting, and building APIs. It is known for its readability and large ecosystem of libraries.",
  "sources": ["knowledge_base.txt:chunk_3", "knowledge_base.txt:chunk_7"],
  "response_time_ms": 1250
}

Core Concepts

• RAG (Retrieval-Augmented Generation): Combines document retrieval with LLM generation for grounded answers.
• Text Splitting: Large documents are chunked into overlapping segments for better retrieval.
• Vector Store (FAISS): Embeds text chunks into vectors for semantic similarity search.
• Prompt Templates: LangChain's PromptTemplate structures the LLM input with context injection.
• Conversation Memory: ConversationBufferMemory maintains chat history across turns.