v0.8

• added slack bot variant to the mix

v0.7.1

• switch to gpt-4 model

v0.7

• split agent and ui
• added ragas notebook and baseline

v0.6.2

Align assistant API impl with latest changes

v0.6.1

• settle on mmr instead of reranking

v0.6

• Update data sources for Camel 4.x releases (includes Camel Quarkus)
• Introduced multi store lookups to overcome some limitations of the current docs
• Harmonised storage of vector and contents to settle on ~3k characters per entry

v0.5.1

• Added three more knowledge sources (Spring Boot Starter, Spring Boot Developer, Fuse on Openshift)

v0.5

• Fix the empty message bug (streamlit session state replay)
• Store the entire conversation when receiving feedback

v0.4

Move to langchain agent based implementation to support a conversational style and memory while retaining the previous data sources

v0.3

Rephrase initial prompt to improve the quality of the input