Luca Garulli comments

Results 366 comments of


                                            Luca Garulli

Rebuilding the vector index makes it 0 dim

Are you using the old HSNW or the new LSM Vector?

Optimize RAM usage with many databases open

Yes, we still need to do those things to improve startup time + initial RAM used

Creating a vector index after nodes already exists doesn't make vectorNeighbors return anything

@ExtReMLapin please get the latest main branch, I've just pushed some fixes to the LSM vector and optimization (60% less space on disk). I've also just added a test case...

Creating a vector index after nodes already exists doesn't make vectorNeighbors return anything

Is there a way you can upload this database as a zip? Or a test case to reproduce the same content?

Cannot create a node with cypher when there is an LSM vector index existing for it

How do you store the array of floats in the vertex/document? Using a list of object ?

Command randomly fails (resending it works) : `com.arcadedb.exception.CommandParsingException`

I've improved the exception to contain the failed query and parameters, so you can check what's wrong. Please let me know if happens again and please attach the new exception.

Vector search is much faster than hnswlib, but returns different neighbors

@tae898 I think the issue is the lazy loading of vectors. We leave the vectors in the documents/vertices, so when the find neighbors is called, the jvector graph (serialized in...

SQL query issue with optimization

This fix slowed down many query execution, reopening it.

Feature Request: FULL_TEXT / CONTAINSTEXT ordering

We need to add a score to the result and then order by score before providing the result. The score algorithm could be super simple: +1 for each matched word.

Feature Request: FULL_TEXT / CONTAINSTEXT ordering

Working on this isse, the concept of score is already implemented inside the full text index, we just need to expose it.