feat: SentenceTransformersDocumentEmbedder and SentenceTransformersTextEmbedder can accept and pass any arguments to SentenceTransformer.encode

[oroszgy]

Enhanced SentenceTransformersDocumentEmbedder and SentenceTransformersTextEmbedder to accept an additional parameter, which is passed directly to the underlying SentenceTransformer.encode method for greater flexibility in embedding customization.

Related Issues

No related issues

Proposed Changes:

Recent embedding methods requires extra parameters when used with SentenceTransformer (e.g. task, prompt, prompt_name). One specific example is jina-embeddings-v3 which supports task-specific embeddings through the task paramater.

How did you test it?

Extended corresponding unit tests and make them pass.

Checklist

• I have read the contributors guidelines and the code of conduct
• I have updated the related issue with new insights and changes
• I added unit tests and updated the docstrings
• I've used one of the conventional commit types for my PR title: fix:, feat:, build:, chore:, ci:, docs:, style:, refactor:, perf:, test: and added ! in case the PR includes breaking changes.
• I documented my code
• I ran pre-commit hooks and fixed any issue

[CLAassistant]

All committers have signed the CLA.

[coveralls]

Pull Request Test Coverage Report for Build 13161311307

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• 4 unchanged lines in 2 files lost coverage.
• Overall coverage increased (+0.001%) to 91.3%

---

Files with Coverage Reduction	New Missed Lines	%
components/embedders/sentence_transformers_document_embedder.py	2	96.77%
components/embedders/sentence_transformers_text_embedder.py	2	96.3%

Totals
Change from base Build 13158724398:	0.001%
Covered Lines:	8952
Relevant Lines:	9805

---

💛 - Coveralls

[anakin87]

Hello and thanks for the PR!

While I understand the need of passing task to jina-embeddings-v3, I see that most of the parameters for encode can be already set in the __init__ of the Embedders, while some of them are not acceptable to make sure that these components run correctly (e.g., output_value, convert_to_numpy, ...).

So I need to think a bit about covering the use case you proposed... I'll let you know!

[oroszgy]

Thanks for the feedback! Looking forward to hearing back from you.

[anakin87]

In general, this approach is OK.

I've left some comments.

Plus,

• let's test that encode_kwargs are passed to the embedding backend using mocks (this test can be an inspiration:

  haystack/test/components/embedders/test_sentence_transformers_document_embedder.py

  Line 298 in 5ae9488

  def test_embed_metadata(self):

  )
• merge main from your branch: I recently merged a PR (fix: fix test failures with Transformers models in PRs from forks #8809) that should make tests pass

[oroszgy]

@anakin87 Thanks for the feedback! Went through all of them, and fixed my PR accordingly.

[anakin87]

some final things to fix

[oroszgy]

Sorry for my oversight, now should be good.

[anakin87]

Nice.

Thanks for your contribution! 💙