v1.6.0

• Separated embeddings from index so they can be cached separate from FAISS index
• Split add_doc method out so that chunks can be added separately

v1.4.1

Updated to newest langchain version syntax

v1.4.0

• Revised to new callback syntax
• Added fallback for non-async LLMs
• Added check for empty citation chain results
• Unit tests working again

v1.3.0

Added ability to do async callbacks

v1.2.0

Made agents more robust to syntax problems

v1.1.1

• Minor improvements to agents
• Added tests for pyzotero

v1.1.0

• Made contexts be a specific dataclass
• Improved README
• Improvements to zotero integration from @MilesCranmer and @luuvy757
• Added agent tests

v1.0.0

Improved speed via batching async ops
Added agent approach to combine scraping and evidence gathering
Tighter integration with paperscraper
Added filter on query to improve performance with large libraries
Zotera Integration from @MilesCranmer
Improved caching

v0.1.0

Added support for new chat models in langchain

v0.0.21

Updated to use ChatGPT turbo pricing