fix(qdrant): respect falsy run() overrides in retrievers

[joshinitin2016883-sketch]

Related Issues

• None for Qdrant. It's the same bug as Falsy run() overrides are silently ignored in SentenceTransformersSimilarityRanker #3854 (in SentenceTransformersSimilarityRanker), fixed the same way as fix: FastEmbedder reordered the top_k fallback to happen after validation #3745–fix: NvidiaRanker reordered the top_k fallback to happen after validation #3749.

Proposed Changes:

QdrantEmbeddingRetriever, QdrantSparseEmbeddingRetriever and QdrantHybridRetriever resolve their run() / run_async() overrides with or:

scale_score=scale_score or self._scale_score,
return_embedding=return_embedding or self._return_embedding,
score_threshold=score_threshold or self._score_threshold,

or falls back on any falsy value, so valid overrides are replaced by the init values:

retriever = QdrantEmbeddingRetriever(document_store, scale_score=True, return_embedding=True, score_threshold=0.5)
retriever.run(query_embedding=[0.1, 0.2], scale_score=False, return_embedding=False, score_threshold=0.0)
# the store receives scale_score=True, return_embedding=True, score_threshold=0.5

This change falls back to the init value only when the argument is None, which is how QdrantHybridRetriever already resolves rrf_k and rrf_weights in the same file.

I left top_k, group_by and group_size unchanged: none of them has a meaningful falsy value.

How did you test it?

• Regression tests: test_run_falsy_runtime_values_override_init and test_run_async_falsy_runtime_values_override_init in each of the three retriever test files. All 6 fail before the change (assert True is False) and pass after.
• Unit tests: 142 passed. Integration tests (in-memory Qdrant): 157 passed, 1 skipped.
• ruff check, ruff format --check and mypy (using the integration's config) are clean.

Hatch is blocked by Windows Application Control on my machine, so these ran from a venv with the integration installed in editable mode rather than through hatch run.

Notes for the reviewer

• Prepared with Claude Code running locally: it found this while checking where the Falsy run() overrides are silently ignored in SentenceTransformersSimilarityRanker #3854 pattern also occurs, wrote the change and tests, and ran the checks above on my machine. I reviewed the diff and the results before opening this PR.

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

🤖 Generated with Claude Code

[github-actions]

Coverage report (qdrant)

This PR does not seem to contain any modification to coverable code.

[bogdankostic]

Thanks @joshinitin2016883-sketch! Looks already good in general, just left one suggestion on making the code more readable.

[bogdankostic]

Let's put the runtime value first, not only here but thoughout the file.

Suggested change

	scale_score=self._scale_score if scale_score is None else scale_score,
	return_embedding=self._return_embedding if return_embedding is None else return_embedding,
	score_threshold=self._score_threshold if score_threshold is None else score_threshold,
	scale_score=scale_score if scale_score is not None else self._scale_score,
	return_embedding=return_embedding if return_embedding is not None else self._return_embedding,
	score_threshold=score_threshold if score_threshold is not None else self._score_threshold,