Wandbot v1.2.0

What's Changed

This release introduces a number of exciting updates and improvements:

• Parallel LLM Calls: Replaced the llama-index with the LECL, enabling parallel LLM calls for increased efficiency.
• ChromaDB Integration: Transitioned from FAISS to ChromaDB to leverage metadata filtering and speed.
• Query Enhancer Optimization: Improved the query enhancer to operate with a single LLM call.
• Modular RAG Pipeline: Split the RAG pipeline into three distinct modules: query enhancement, retrieval, and response synthesis, for improved clarity and maintenance.
• Parent Document Retrieval: Introduced parent document retrieval functionality within the retrieval module to enhance contextuality.
• Sub-query Answering: Added sub-query answering capabilities in the response synthesis module to handle complex queries more effectively.
• API Restructuring: Redesigned the API into separate routers for retrieval, database, and chat operations.

Evaluation Result

wandbot version	Comment	response accuracy
1.0.0	our baseline wandbot	53.78 %
1.1.0	improvement over baseline; in production for the longest	72.45 %
1.2.0	our new enhanced wandbot	81.63 %

Full Changelog: v1.1.0...v1.2.0

Wandbot v1.1.0

What's Changed

• Feat v1.1 Updates to chat client, retrieval, and evauations by @parambharat in #54
• add more info on evaluation by @ayulockin in #50
• fix: add app logging to stream tables by @parambharat in #51
• Zendesk to replit by @ArtsiomWB in #52

Full Changelog: v1.0.0...v1.1.0

Wandbot v1.0.0

What is Wandbot?

wandbot is our W&B GPT-4 powered bot live in Discord, internal Slack, ChatGPT (yes) and Discourse DM’s (coming soon). wandbot has been helping internal and external users for over 7 months now, the code is fully open source and the Replit deployment is public, making it a great “build in public” story of how we’re approaching AI-powered app development.
wandbot stack

For the nerds:

GPT-4
OpenAI’s ada embeddings
Cohere re-ranker
LlamaIndex
faiss index
fastapi
Replit Deployments