WhatsApp Chat BOT

Purpose

This is a ChatBot (Helper) designed to answer questions posted in on a WhatsApp group. The ChatBot understands your queries and provides the closest possible answers.

Example: Run below queries

• Google drive links
• Training Meeting Invite
• Baljeet contact details, please
• Recommended dev environment
• Quick_Chicken_Byte

Environment

• Linux
• Python3
• ChromaDB
• ngrok

Architecture

How to run?

1. Export WhatsApp messages to text file
2. Run 'parser_querymate.py' which generates csv from text file

   python parser_querymate.py

3. RAG csv data using 'producer_querymate.py' and store in ChromaDB file

   python producer_querymate.py

4. Run 'consumer_querymate.py' with sample quries

   python consumer_querymate.py

5. Run 'ui_querymate.py' UI service on localhost to communicate with CromaDB

   streamlit run ui_querymate.py

6. Run ngrok and expose UI interface to cloud

   ngrok http http://localhost:8501

Get sources from github

git clone [email protected]:bhagavansprasad/QueryMate.git

TODO

1. Image search
2. Replies are not upto the mark
3. Works on offline messages of WhatsApp
4. Optimization
5. User different LLMs to make the search better
6. Port it to Hugging Interface