Documentation-AI-Assistant

This repo is about how to create a documentation chatbot

Chatbot Prototype Setup Guide

1. Create a Virtual Environment

python3 -m venv venv
source venv/bin/activate

2. Install Python Dependencies

pip install -r requirements.txt 

3. Set Gemini API Key

• Create a .env file in the project root directory

GEMINI_API_KEY=api_key_here

4. Adjust Documentation Path

• In docs_db.py, you need to update the path to your local Markdown files.

5. Build the Vector Database

• Run docs_db.py to create the FAISS vector index:

python docs_db.py 

• You should see the message "FAISS vectorstore saved".

6. Run the Application

streamlit run app.py 

Notes

• Vector Embedding: [HuggingFaceEmbeddings] (https://huggingface.co/docs/transformers/quicktour)for vector embeddings

• Chunk Size Adjustment: In docs_db.py or rag_backend.py, you can adjust the chunk size if retrieval accuracy is low.

• Model Configuration:: You can change which Gemini model is used in app.py