search

seunginah / semanticvectors

Automatically exported from code.google.com/p/semanticvectors
0 stars 0 forks source link

getVector method in VectorStoreReader could be more efficient. #1

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
Current implementation of getVector does a linear scan of the objects in
the vector store. This lookup could easily be made logarithmic. 

You won't notice a performance hit unless you have a lot of query terms,
especially since the vector search using cosine similarity is itself linear.

Original issue reported on code.google.com by maryl...@gmail.com on 25 Oct 2007 at 9:03

GoogleCodeExporter commented 9 years ago

Original comment by dwidd...@gmail.com on 25 Oct 2007 at 9:06

GoogleCodeExporter commented 9 years ago 
This is fixed to some extent with the VectorStoreReaderRAMCache: if i. you have 
lots
of queries to make, and ii. your index is small enough to fit in memory, then 
using
VectorStoreRAMCache will give this performance improvement.

It would still be nice to have something that works from disk as well.

Original comment by dwidd...@gmail.com on 5 Jun 2008 at 7:36

GoogleCodeExporter commented 9 years ago 
After a few months and no complaints / requests, I'm guessing that 
VectorStoreRAM
solves any real problems, reducing single-query latency is not a priority.

Original comment by dwidd...@gmail.com on 12 Jul 2009 at 12:55