Improve performance with contextual compression, a technique where retrieved documents are compressed, and irrelevant information is filtered out.

Temporarily remove epub file loader for now

Update document_retriever.py

Create ruff.toml

Disable tokenizer transformer parallelism to avoid deadlocks

Update chat_profile.py

Use all-mpnet-base-v2 vector embedding for highest performance

Update app icon

Update icons

Remove top bar

Update gitignore

Update README

Update README

Add new app icon. Refactor and cleanup.

Update app icon README

Handle epub loader

Refactor

Update app

Refactor

Add txt and docx file loader

Update app.py

Update app.py

Update README

Update README

Update UI

Refactor app with better Document retrieval embedding and better chat streaming

Refactor app with better Document retrieval embedding and better chat streaming

Update API key config

Update app

Fix configuration

Deploy

Refactor app using RunnableWithMessageHistory with better Prompt tuning

Update app

Handle secrets

Update config

Update app

Handle secrets management

Add theme config

Deploy app

Refactor app

Refactor app

Extract chat profile to class

Refactor to extract token stream handler

Deploy

Update chat

Merge branch 'main' of github.com:vinhnx/InkChat

Update pysqlite3-binary package to run on cloud

Update client

More proper client

Using langchain_openai to stream message