RAG API Chatbot

A Retrieval-Augmented Generation (RAG) chatbot that answers questions about public APIs using a CSV knowledge base, LangChain, OpenAI, and Streamlit.

About the Project

This project is a chatbot that leverages Retrieval-Augmented Generation (RAG) to answer user questions about public APIs. It uses a CSV file as its knowledge base, retrieves relevant API information using vector search (FAISS), and generates accurate, context-aware answers using OpenAI's GPT-3.5-turbo model. The chatbot can be run both as a command-line interface (CLI) and as a web app via Streamlit. Link for streamlit application: https://rag-chatbot-gpt.streamlit.app/

Setup Instructions

1. Clone the Repository

git clone https://github.com/<your-username>/<your-repo>.git
cd <your-repo>

2. Create and Activate a Virtual Environment

python3 -m venv .venv
source .venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

4. Set Up Your OpenAI API Key

For Local Development:

• Create a .env file in the project root:

  OPENAI_API_KEY=sk-...
  

For Streamlit Cloud:

• Go to your app’s settings on Streamlit Cloud.
• Add your OpenAI API key in the Secrets section:

  OPENAI_API_KEY = "sk-..."
  

5. Run the Chatbot

Command-Line Interface:

python main.py

Streamlit Web App:

streamlit run app.py

• Open the provided local URL in your browser.

Project Structure & File Descriptions

File/Folder	Purpose
app.py	Streamlit web app for interactive chatbot experience.
main.py	Command-line interface for the chatbot.
vectors.py	Loads the CSV, builds the vector store, and exposes the retriever for RAG.
public_apis_clean.csv	The knowledge base: a CSV file containing public API details.
requirements.txt	Python dependencies for the project.
.env	(Not committed) Stores your OpenAI API key for local development.
.streamlit/secrets.toml	(Not committed) Stores your OpenAI API key for Streamlit Cloud deployment.
.gitignore	Ensures secrets and unnecessary files are not committed to git.

About Each File

app.py

• The main Streamlit app.
• Loads the retriever and LLM, formats context, and provides a chat interface.
• Reads the OpenAI API key from Streamlit secrets.

main.py

• CLI version of the chatbot.
• Reads the OpenAI API key from the environment (via .env).
• Useful for quick local testing.

vectors.py

• Loads the CSV knowledge base.
• Creates document embeddings using OpenAI.
• Builds a FAISS vector store for fast similarity search.
• Exposes a retriever object for use in both CLI and web app.

public_apis_clean.csv

• The knowledge base for the chatbot.
• Contains details about various public APIs (name, description, auth, HTTPS, CORS, link, etc.).

requirements.txt

• Lists all Python dependencies required for the project.

.env

• Not committed to git.
• Stores your OpenAI API key for local development.

.streamlit/secrets.toml

• Not committed to git.
• Stores your OpenAI API key for Streamlit Cloud deployment.

.gitignore

• Ensures that sensitive files (like .env and .streamlit/secrets.toml) and unnecessary files are not tracked by git.

Notes

• The chatbot uses FAISS for in-memory vector search, which works on both local and Streamlit Cloud deployments.
• You can expand the knowledge base by adding more rows to public_apis_clean.csv.

Citations

• Use of LLMs
  • ChatGPT for creating embeddings and computing the API keys
  • Gemini for creating comprehensive README with appropriate formatting

Enjoy your RAG-powered API chatbot!