search

kingjulio8238 / memary

Human Memory for AI Agents.
https://www.memarylabs.com
MIT License
1.14k stars 79 forks source link

Memory Retention #46

Open rajib76 opened 1 week ago

rajib76 commented 1 week ago

The code can be further optimized by reducing some of the hardcodings. In this pull request, i moved the user to .env. I plan to look at some more areas to improve and create a pull request for the same.

kingjulio8238 commented 1 week ago

@rajib76 thanks for the request to make memary more flexible. It would be great to fix any more existing hardcodings within this PR!

rajib76 commented 1 week ago

I have added the commit to fix the typo in README. Can this now be merged? I will need some time to look at the rest of the hardcoding, it may require a little more coding. For example thinking to create a graphmemorystore class so that we give flexibility to users to use any type of graph db and not just NEO4J. Currently the KG supports only NEO4J.

rajib76 commented 1 week ago

Added the NEO4J_DB also in .env. With this I think I have taken care of most of the hard codings. There are other opportunities to optimize the code which will do slowly

kingjulio8238 commented 1 week ago

@rajib76 Agreed, It would be great to expand support to other graph providers! Once you have finished with that update this PR and we'll merge the contribution.

Thanks again! Excited to see updates

rajib76 commented 1 week ago

Added the forgetting feature using ebbinghaus law