Support out-of-core and distributed IVF_PQ ingestion #531
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
NikolaosPapailiou
approved these changes
Oct 17, 2024
| + "".join(random.choices(string.ascii_letters, k=10)) | ||
| ) | ||
| if index_type == "IVF_PQ": | ||
| PARTIAL_WRITE_ARRAY_DIR = storage_formats[storage_version][ |
There was a problem hiding this comment.
Why not follow the same dir structure? The random number part is required in order to support parallel or failed ingestions.
There was a problem hiding this comment.
Ah, so I did not realize it was for parallel ingestions, I thought just failed ingestions. And so I added logic to delete the group before use in C++. I did this because I didn't want Python and C++ to have to communicate about what the name of the dir was. But I'll update to follow the pattern in this follow-up PR: #554
jparismorgan
changed the title
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Outdated suggestions cannot be applied.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.
What
Here we update to support out-of-core and distributed ingestion in IVF_PQ. We do this by updating the IVF_PQ API. The general design is to use the IVF_FLAT code path, but use these IVF_PQ C++ functions:
create()train()ingest_parts()consolidate_partitions()Note that we also use
create_temp_data_group()because to support re-ingestion with a new temp data group.In the future we may also want to move
compute_partition_indexes()in C++, but for now we leave it in Python.We still are able to support C++-only use of the index with this approach, though to make things easier we also support
ingest()so that a user only needs to callcreate(),train(), and theningest().Testing