feat: add support for colbert-style multivector datasets

[AstraBert]

This PR adds support for loading ColBERT-style multivectors from a dataset.

The dataset, denoted by the multivector format, should encompass:

• A vectors.npy file containing all sub-vectors stacked together
• A offsets.npy file containing num_points + 1 vector boundaries as int32/int64

We parse both files and load the i-th vector as vectors[offsets[i]:offsets[i+1]].

Here is an example configuration:

collection:
   # other fields
   vectors:
    - name: colbert
      size: 128
      multivector:
        comparator: max_sim
        count: 1                 # ignored for this source
      source:
        type: dataset
        name: my-colbert-corpus
        format: multivector
        path: my-colbert-corpus  # directory containing vectors.npy + offsets.npy

All Submissions:

• Contributions should target the dev branch. Did you create your branch from dev?
• Have you followed the guidelines in our Contributing document?
• Have you checked to ensure there aren't other open Pull Requests for the same update/change?

New Feature Submissions:

1. Does your submission pass tests?
2. Have you formatted your code locally using cargo +nightly fmt --all command prior to submission?
3. Have you checked your code using cargo clippy --all --all-features command?

Changes to Core Features:

• Have you added an explanation of what your changes do and why you'd like us to include them? (see above)
• Have you written new tests for your core changes, as applicable?
• Have you successfully ran tests with your changes locally?

[AstraBert]

Clippy failures seem pre-existing? It didn't catch anything locally

[xzfc]

Thanks.

Clippy failures seem pre-existing?

That's right.

It didn't catch anything locally

These appear in Rust 1.98 which released less than a week ago. Perhaps you want to upgrade.