Save and load usearch options from index header page #101
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
var77
force-pushed
the
feature/save-usearch-options-in-header
branch
from
0ecf4b8 to
994917a
Compare
var77
changed the title
Ngalstyan4
approved these changes
Sep 5, 2023
src/hnsw/build.c
Outdated
| assert(error == NULL); | ||
| opts.dimensions = usearch_dimensions(buildstate->usearch_index, &error); | ||
| assert(error == NULL); | ||
| opts.expansion_add = usearch_expansion_add(buildstate->usearch_index, &error); |
There was a problem hiding this comment.
Could all of these be retrieved from usearch_metadata?
There was a problem hiding this comment.
There should be done some changes in usearch to return this data from metadata but yes technically it is possible.
There was a problem hiding this comment.
I have added this params to usearch_metadata in this PR
| @@ -39,7 +39,7 @@ EXPLAIN (COSTS FALSE) SELECT ROUND(l2sq_dist(v, :'v777')::numeric, 2) FROM sift_ | |||
|
|
|||
| SELECT ROUND(l2sq_dist(v, :'v777')::numeric, 2) FROM sift_base1k order by v <-> :'v777' LIMIT 10; | |||
| INFO: began scanning with 0 keys and 1 orderbys | |||
| INFO: starting scan with dimensions=128 M=16 efConstruction=128 ef=64 | |||
| INFO: starting scan with dimensions=128 M=16 efConstruction=64 ef=32 | |||
There was a problem hiding this comment.
Can you comment on what happened here?
There was a problem hiding this comment.
I have changed the ef search and ef construction to other values from the default ones we have now, so it would it more visible that the values are being loaded from the file header
var77
force-pushed
the
feature/save-usearch-options-in-header
branch
3 times, most recently
from
bff3ce7 to
89a05af
Compare
var77
force-pushed
the
feature/save-usearch-options-in-header
branch
from
89a05af to
36a4307
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
With this changes we will keep the necessary usearch options (
dimensions,ef,ef_construction,m,metric_kind) in our index header when building index and in scans read the options from index header instead of index options.Also we will add
efandef_constructionparams in usearch index header, and make it reload the values when loaded from file as you can see in this PR .By doing this we can omit all the parameters when building index from file, as we will get the required values from the usearch header and save them in our index header to be used on table scans.
Currently the tests are failing, as the old versions of index files are being used, they need to be replaced in the bucket with newer versions.