Skip to content

Latest commit

 

History

History
29 lines (20 loc) · 2.05 KB

readme.md

File metadata and controls

29 lines (20 loc) · 2.05 KB

YelpBot

Deployed @ Yelp Chatbot on Streamlit

Currently down due to API credit limits! Check out the demo again at the start of the next month or clone the repo and update the .streamlit/secrets.toml file with own API key. .

Screenshot

This chatbot interfaces with a LangChain agent designed to answer questions about a subset of users, businesses, and reviews on Yelp. The agent uses retrieval-augment generation (RAG) over both structured and unstructured data. Data is hosted by Neo4js Graph Database Management System. Dataset provided by Yelp.

Credits

To Do

  1. Restrict to domain-specific queries. ("I can only answer questions pertaining to Yelp data.")
  2. ✨ Update yelp_review_chain/Experience tool, to integrate business identifying information into the embedding for the reviews.
  3. ✨ Scale (< 10% total data currently deployed due to github file size limitations.) Goal: Full Yelp datasets, including tips.csv. Update 11.04.2024 - Need to readjust expectations due to space limitations of Neo4js.
  • yelp_agent.py: Complete TripTimes Tool integration. ✅ 29.03.2024
  • Deploy on streamlit community cloud. ✅ 30.03.2024
  • Enhance TripTimes tool: add query filter for location specification. ✅ 31.03.2024
  • yelp_bulk_csv_write.py: Fix issues with specifying data types that are affecting aggregation queries. ✅ 31.03.2024
  • Add Memory i.e. making the YelpBot a conversational bot. ✅ 09.04.2024
  • New Tool: NearestBusiness (formerly ProximityFinder). (find closest business by trip time). tools/nearestbusiness_chain.py ✅ 11.04.2024