NVIDIA / NVIDIA/cuvs

[FEA] Standardize `cuVS::neighbors::index`

Open
#222 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

feature request
Dominant language
Cuda
Stars
854
Forks
236
Avg merge
3d 3h
Merged PRs (30d)
62

Description

Currently, all of the Ann indexes inherit from a common base class, but the base class doesn't have any virtual methods or state attached to it, so any APIs coded to the index class can't even assume common functions like build and search. This is making it hard to code new APIs like tiered index and replication / sharding APIs that are agnostic to the index type.

We have been working on generalizing the various index implementations so that we can enable this functionality. This issue is primarily concerned with taking it across the finish line and having all of the Ann index implementations inherit at least their build(), search() and extend() capabilities from the common base (or mixins if more appropriate given the differences).

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reviewing the common base class and the various ANN index implementations. Compare how build(), search(), and extend() are currently exposed across the implementations, then determine how the common base or mixins can provide those capabilities. Done means all ANN indexes share these capabilities so APIs such as tiered indexing and replication or sharding can remain index-agnostic.

Written by the indexing model from the issue text.

Assessment

Domain
backend-api-design, machine-learning, search
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.