You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As related to my investigations of the Python redis memory connector in #6084 and #6088, the redis memory connector by default only goes *=> to do the search. This will return all the keys in the store. However, there should be an option for the user to allow them to provide a key to filter on, much like in native redis. Although you can pass in the collection name, this allows for greater granularity and accuracy of RAG, especially for large applications that have similar values and many k-v pairs stored.
We found that when we had key value pairs with similar or nearly identical values, even when passing the collection name in, we got poor performance, unless we modified the filter.
The text was updated successfully, but these errors were encountered:
As related to my investigations of the Python redis memory connector in #6084 and #6088, the redis memory connector by default only goes *=> to do the search. This will return all the keys in the store. However, there should be an option for the user to allow them to provide a key to filter on, much like in native redis. Although you can pass in the collection name, this allows for greater granularity and accuracy of RAG, especially for large applications that have similar values and many k-v pairs stored.
We found that when we had key value pairs with similar or nearly identical values, even when passing the collection name in, we got poor performance, unless we modified the filter.
The text was updated successfully, but these errors were encountered: