feat: add SDK support for search records with similarity

[jfcalvo]

Description

This PR suggest an implementation to give support to /api/me/datasets/{dataset_id}/records/search endpoint with vector similarity searching.

Here is a full example working with similarity search from python SDK using no real 100% fake data:

rg.init(api_key=owner.api_key)
workspace = Workspace.create(name="test")

feedback_dataset.add_vector_settings(VectorSettings(name="vector", dimensions=2))

feedback_dataset.add_records(
    [
        FeedbackRecord(fields={"text": "hello"}, vectors={"vector": [1, 1]}),
        FeedbackRecord(
            fields={"text": "hello"},
            vectors={"vector": [2, 2]},
            responses=[ResponseSchema(status="discarded")],
        ),
        FeedbackRecord(
            fields={"text": "hello"},
            vectors={"vector": [4, 4]},
            responses=[ResponseSchema(status="submitted", values={"question": {"value": "answer"}})],
        ),
    ]
)

remote = feedback_dataset.push_to_argilla("test_find_similar_records", workspace=workspace)
only_submitted_and_discarded_records = remote.filter_by(response_status=["submitted", "discarded"])

records_with_scores = only_submitted_and_discarded_records.find_similar_records(
    vector_name="vector",
    value=[1, 1],
    max_results=3,
)
# Even if the used vector is the same that contains the record 0, the search will skip it since it has no submitted or discarded responses.

Closes #4020

Type of change

• New feature (non-breaking change which adds functionality)

How Has This Been Tested

Tested locally

Checklist

• I added relevant documentation
• follows the style guidelines of this project
• I did a self-review of my code
• I made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• I filled out the contributor form (see text above)
• I have added relevant notes to the CHANGELOG.md file (See https://keepachangelog.com/)

[alvarobartt]

As already mentioned here I think we should use a client schema indeed, not the SDK one

[alvarobartt]

WDYT about using pydantic schemas in the SDK instead of just the kwargs? @frascuchon @gabrielmbmb

[alvarobartt]

WDYT about using pydantic schemas in the SDK instead of kwargs? @frascuchon @gabrielmbmb

[github-actions]

The URL of the deployed environment for this PR is https://argilla-quickstart-pr-4023-ki24f765kq-no.a.run.app

[codecov]

Codecov Report

Attention: 5 lines in your changes are missing coverage. Please review.

Files	Coverage Δ
.../argilla/client/feedback/schemas/remote/records.py	88.88% <100.00%> (+0.11%)	⬆️
src/argilla/client/sdk/commons/errors.py	90.90% <100.00%> (+0.16%)	⬆️
src/argilla/client/sdk/v1/datasets/api.py	92.23% <100.00%> (+0.78%)	⬆️
src/argilla/client/sdk/v1/datasets/models.py	100.00% <100.00%> (ø)
src/argilla/server/search_engine/elasticsearch.py	93.10% <ø> (+20.68%)	⬆️
src/argilla/server/search_engine/opensearch.py	48.21% <ø> (-42.86%)	⬇️
src/argilla/client/feedback/dataset/base.py	86.20% <75.00%> (-0.35%)	⬇️
...c/argilla/client/feedback/dataset/local/dataset.py	85.97% <60.00%> (-0.90%)	⬇️
.../argilla/client/feedback/dataset/remote/dataset.py	92.14% <85.71%> (-0.31%)	⬇️

... and 10 files with indirect coverage changes

📢 Thoughts on this report? Let us know!.

[alvarobartt]

Left some minor comments, I'll complete the review once those are addressed, nice job!