search

deepset-ai / haystack

:mag: LLM orchestration framework to build customizable, production-ready LLM applications. Connect components (models, vector DBs, file converters) to pipelines or agents that can interact with your data. With advanced retrieval methods, it's best suited for building RAG, question answering, semantic search or conversational agent chatbots.
https://haystack.deepset.ai
Apache License 2.0
15.21k stars 1.76k forks source link

Add InMemoryChatMessageStore in haystack-experimental #7919

Closed julian-risch closed 1 week ago

julian-risch commented 1 month ago

Add an InMemoryChatMessageStore similar to InMemoryDocumentStore that can work with ChatMessageRetriever and ChatMessageWriter.

julian-risch commented 1 month ago

Here is a branch with InMemoryChatMessageStore: https://colab.research.google.com/drive/1eDGOR241q9Ek35BvETp1ihiA5yJH9dkK?usp=sharing I merged the ChatMessageWriter and ChatMessageRetriever and here is an example notebook using the three: https://github.com/deepset-ai/haystack-experimental/compare/main...chat_message_store

Should be separated in case we open PRs. Right now the purpose of the branches is to facilitate a discussion about alternative implementations of memory

Cc @vblagoje

vblagoje commented 1 month ago

Should we create RAG+chat to accent the difference between the old and the new approach? I can do that part...

julian-risch commented 1 week ago

closed by https://github.com/deepset-ai/haystack-experimental/pull/49