docs: replace old haystack core imports with haystack_integrations paths by julian-risch · Pull Request #3545 · deepset-ai/haystack-core-integrations

julian-risch · 2026-07-03T12:45:56Z

Related Issues

Part of https://github.com/deepset-ai/haystack-private/issues/449

Proposed Changes:

Replace stale from haystack.components.embedders import SentenceTransformers... imports with the new from haystack_integrations.components.embedders.sentence_transformers import ... paths in the usage examples of retriever docstrings (these generate the API reference pages): arcadedb, elasticsearch (x2), faiss, opensearch, pgvector, pinecone, supabase, valkey, and the transformers TransformersZeroShotTextRouter
Add a # Requires: pip install sentence-transformers-haystack comment above the new imports in docstring examples
Same import replacement in example scripts of alloydb, astra, azure_ai_search, langfuse, llama_cpp, mongodb_atlas, pgvector, pinecone, pyversity, qdrant, valkey, vespa, and weave (plus HuggingFaceAPIChatGenerator in the langfuse chat example)
Rename TransformersSimilarityRanker → SentenceTransformersSimilarityRanker in the weave hybrid-retrieval example (old core class name and old import path)
Remove explicit warm_up() calls on SentenceTransformers embedders in docstrings/examples — their run() auto-invokes warm-up

How did you test it?

python -m py_compile on all changed files
hatch fmt --check passes on the changed sources; a scan confirms zero remaining old-path imports of migrated classes in the repo

Notes for the reviewer

I chose not to mention the extra new dependency in example files but only in the docstring examples. Happy to discuss other options.

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:.

🤖 Generated with Claude Code

SentenceTransformers embedders/rankers moved from Haystack core to this repo ahead of Haystack 3.0, but retriever docstrings and example scripts still imported them from haystack.components. Update docstring usage examples (which generate the API reference pages) and example scripts to the new haystack_integrations import paths, note the required sentence-transformers-haystack package, and drop explicit warm_up() calls now that the embedders auto-warm-up on run(). Also rename TransformersSimilarityRanker to SentenceTransformersSimilarityRanker in the weave example. Related to deepset-ai/haystack-private#449 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

github-actions · 2026-07-03T12:47:00Z