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
What do you think about exposing createIndex/createIndexIfMissing for the RedisVectorStore class?
I want to execute index creation on demand, even before the first similarity search execution.
Checking if the index exists on every execution of similaritySearch is suboptimal (it doubles the number of calls with the additional execution of $this->client->ftinfo($this->redisIndex);).
I would also like to have the chance to configure the indexing method: HNSW or FLAT. FLAT is good for reduced datasets, HNSW scales much better.
The text was updated successfully, but these errors were encountered:
I will take a look and propose something. I would find it useful to enable more configurations for the index; besides the length and the algorithm, distance and float size are good ones, too. Enabling hash data modeling would complicate the current implementation, maybe we can evaluate it in a later iteration. I will share something when I have a working sample.
Hi,
What do you think about exposing createIndex/createIndexIfMissing for the RedisVectorStore class?
I want to execute index creation on demand, even before the first similarity search execution.
Checking if the index exists on every execution of similaritySearch is suboptimal (it doubles the number of calls with the additional execution of $this->client->ftinfo($this->redisIndex);).
I would also like to have the chance to configure the indexing method: HNSW or FLAT. FLAT is good for reduced datasets, HNSW scales much better.
The text was updated successfully, but these errors were encountered: