refactor(pinecone): make get_metadata_field_min_max more consistent; use async DocumentStore mixin tests

[Aftabbs]

Related Issues

• fixes Use async DocumentStore mixin tests in Pinecone #3051

Proposed Changes:

Replaces 17 duplicate async test methods in TestDocumentStoreAsync with the generic mixin classes from haystack.testing.document_store_async (available since haystack-ai 2.28.0).

Mixins added:

• WriteDocumentsAsyncTest
• CountDocumentsAsyncTest
• FilterDocumentsAsyncTest
• CountDocumentsByFilterAsyncTest
• CountUniqueMetadataByFilterAsyncTest
• DeleteDocumentsAsyncTest
• DeleteAllAsyncTest
• DeleteByFilterAsyncTest
• UpdateByFilterAsyncTest
• GetMetadataFieldsInfoAsyncTest
• GetMetadataFieldMinMaxAsyncTest
• GetMetadataFieldUniqueValuesAsyncTest

Preserved (Pinecone-specific):

• test_embedding_retrieval — exercises _embedding_retrieval_async with real vectors
• test_close — verifies the async index lifecycle (close_async / re-initialize)
• test_sentence_window_retriever — end-to-end retrieval with SentenceWindowRetriever

Version bump: haystack-ai>=2.26.1 → >=2.28.0 (required for haystack.testing.document_store_async module).

Note: test_count_not_empty_async is overridden to work around a missing @staticmethod on the mixin method in haystack v2.28.0 (fixed in haystack main, tracked in deepset-ai/haystack#11174).

How did you test it?

• Integration tests run against a live Pinecone index via CI (pytest -v -m integration)
• Mixin-inherited tests provide equivalent coverage for the removed methods
• Pinecone-specific tests (embedding, close lifecycle, sentence window) still run

Notes for the reviewer

The class-level document_store fixture delegates to the conftest's document_store_async fixture (which already wraps write_documents_async, delete_documents_async, and delete_all_documents_async with asyncio.sleep for Pinecone's eventual consistency). This keeps the sleep behaviour intact for all mixin tests without duplicating fixture code.

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: refactor:

[github-actions]

Coverage report (pinecone)

Click to see where and how coverage changed

File	Statements	Missing	Coverage	Coverage (new stmts)	Lines missing
integrations/pinecone/src/haystack_integrations/document_stores/pinecone
document_store.py					869, 872-873
Project Total

This report was generated by python-coverage-comment-action

[anakin87]

I ran the integration tests locally (they do not run on PRs from forks for security reasons)

I got the following errors

FAILED tests/test_document_store_async.py::TestDocumentStoreAsync::test_get_metadata_field_min_max_empty_collection_async - ValueError: No values found for metadata field 'priority'
FAILED tests/test_document_store_async.py::TestDocumentStoreAsync::test_delete_documents_empty_document_store_async - pinecone.exceptions.exceptions.NotFoundException: (404)
Reason: Not Found
HTTP response headers: <CIMultiDictProxy('Date': 'Mon, 27 Apr 2026 12:56:18 GMT', 'Content-Type': 'application/json', 'Content-Length': '55', 'Connection': 'keep-alive', 'x-pinecone-request-latency-ms': '25', 'x-envoy-upstream-service-time': '25', 'x-pinecone-response-duration-ms': '26', 'Server': 'envoy')>
HTTP response body: {"code":5,"message":"Namespace not found","details":[]}
FAILED tests/test_document_store_async.py::TestDocumentStoreAsync::test_filter_simple_async - AssertionError
FAILED tests/test_document_store_async.py::TestDocumentStoreAsync::test_filter_compound_async - AssertionError
FAILED tests/test_document_store_async.py::TestDocumentStoreAsync::test_write_documents_async - NotImplementedError: Default write_documents_async() behaviour depends on the Document Store implementation, as we don't enforce a default behaviour when no policy is set. Override this test in your custom test class.
FAILED tests/test_document_store_async.py::TestDocumentStoreAsync::test_write_documents_duplicate_fail_async - Failed: DID NOT RAISE <class 'haystack.document_stores.errors.errors.DuplicateDocumentError'>
FAILED tests/test_document_store_async.py::TestDocumentStoreAsync::test_write_documents_duplicate_skip_async - AssertionError

I suggest you also run them locally (a free Pinecone API key is needed) and update this PR to make them pass

[github-actions]

Heads-up for maintainers

This PR is from a fork and touches integrations whose integration tests require API keys.
Those tests are skipped in CI because fork PRs don't have access to repo secrets for security reasons.

Affected integrations:

• pinecone

Please run the integration tests locally (hatch run test:integration inside each folder) before approving.

[anakin87]

FAILED tests/test_document_store_async.py::TestDocumentStoreAsync::test_filter_simple_async - AssertionError: assert [Document(id=1a46c9a3f25735c2d8402227a895d59168b60004ea88bcdebaae45e4c64dc5db, content: 'A Foo Document 0', meta: {'chapter': 'intro', 'date': '1969-07-21T20:17:40', 'name': 'name_0', 'number': 2, 'page': '100'}, embedding: vector of size 768), Document(id=d919935b9a5b055ccc496154697127f40888631a1d5dc776b3c620fb843d7f09, content: 'A Foo Document 2', meta: {'chapter': 'intro', 'date': '1969-07-21T20:17:40', 'name': 'name_2', 'number': 2, 'page': '100'}, embedding: vector of size 768), Document(id=f3b69f5129585bbd1c529979ea301d812fcbab89e25b1a55d2b2c91afac0ab83, content: 'A Foo Document 1', meta: {'chapter': 'intro', 'date': '1969-07-21T20:17:40', 'name': 'name_1', 'number': 2, 'page': '100'}, embedding: vector of size 768)] == [Document(id=1a46c9a3f25735c2d8402227a895d59168b60004ea88bcdebaae45e4c64dc5db, content: 'A Foo Document 0', meta: {'name': 'name_0', 'page': '100', 'chapter': 'intro', 'number': 2, 'date': '1969-07-21T20:17:40'}, embedding: vector of size 768), Document(id=d919935b9a5b055ccc496154697127f40888631a1d5dc776b3c620fb843d7f09, content: 'A Foo Document 2', meta: {'name': 'name_2', 'page': '100', 'chapter': 'intro', 'number': 2, 'date': '1969-07-21T20:17:40'}, embedding: vector of size 768), Document(id=f3b69f5129585bbd1c529979ea301d812fcbab89e25b1a55d2b2c91afac0ab83, content: 'A Foo Document 1', meta: {'name': 'name_1', 'page': '100', 'chapter': 'intro', 'number': 2, 'date': '1969-07-21T20:17:40'}, embedding: vector of size 768)]

  At index 0 diff: Document(id=1a46c9a3f25735c2d8402227a895d59168b60004ea88bcdebaae45e4c64dc5db, content: 'A Foo Document 0', meta: {'chapter': 'intro', 'date': '1969-07-21T20:17:40', 'name': 'name_0', 'number': 2, 'page': '100'}, embedding: vector of size 768) != Document(id=1a46c9a3f25735c2d8402227a895d59168b60004ea88bcdebaae45e4c64dc5db, content: 'A Foo Document 0', meta: {'name': 'name_0', 'page': '100', 'chapter': 'intro', 'number': 2, 'date': '1969-07-21T20:17:40'}, embedding: vector of size 768)

  Full diff:
    [
  -     Document(id=1a46c9a3f25735c2d8402227a895d59168b60004ea88bcdebaae45e4c64dc5db, content: 'A Foo Document 0', meta: {'name': 'name_0', 'page': '100', 'chapter': 'intro', 'number': 2, 'date': '1969-07-21T20:17:40'}, embedding: vector of size 768),
  -     Document(id=d919935b9a5b055ccc496154697127f40888631a1d5dc776b3c620fb843d7f09, content: 'A Foo Document 2', meta: {'name': 'name_2', 'page': '100', 'chapter': 'intro', 'number': 2, 'date': '1969-07-21T20:17:40'}, embedding: vector of size 768),
  -     Document(id=f3b69f5129585bbd1c529979ea301d812fcbab89e25b1a55d2b2c91afac0ab83, content: 'A Foo Document 1', meta: {'name': 'name_1', 'page': '100', 'chapter': 'intro', 'number': 2, 'date': '1969-07-21T20:17:40'}, embedding: vector of size 768),
  +     Document(id=1a46c9a3f25735c2d8402227a895d59168b60004ea88bcdebaae45e4c64dc5db, content: 'A Foo Document 0', meta: {'chapter': 'intro', 'date': '1969-07-21T20:17:40', 'name': 'name_0', 'number': 2, 'page': '100'}, embedding: vector of size 768),
  +     Document(id=d919935b9a5b055ccc496154697127f40888631a1d5dc776b3c620fb843d7f09, content: 'A Foo Document 2', meta: {'chapter': 'intro', 'date': '1969-07-21T20:17:40', 'name': 'name_2', 'number': 2, 'page': '100'}, embedding: vector of size 768),
  +     Document(id=f3b69f5129585bbd1c529979ea301d812fcbab89e25b1a55d2b2c91afac0ab83, content: 'A Foo Document 1', meta: {'chapter': 'intro', 'date': '1969-07-21T20:17:40', 'name': 'name_1', 'number': 2, 'page': '100'}, embedding: vector of size 768),
    ]
FAILED tests/test_document_store_async.py::TestDocumentStoreAsync::test_filter_compound_async - AssertionError: assert [Document(id=8dc93d7bb8817ed727b8fb0fabfbed6a5c130f7c4d4df05c5615eb7ce6b7bcec, content: 'A Foo Document 0', meta: {'chapter': 'intro', 'date': '1969-07-21T20:17:40', 'name': 'name_0', 'number': 2, 'page': '100'}, embedding: vector of size 768)] == [Document(id=8dc93d7bb8817ed727b8fb0fabfbed6a5c130f7c4d4df05c5615eb7ce6b7bcec, content: 'A Foo Document 0', meta: {'name': 'name_0', 'page': '100', 'chapter': 'intro', 'number': 2, 'date': '1969-07-21T20:17:40'}, embedding: vector of size 768)]

  At index 0 diff: Document(id=8dc93d7bb8817ed727b8fb0fabfbed6a5c130f7c4d4df05c5615eb7ce6b7bcec, content: 'A Foo Document 0', meta: {'chapter': 'intro', 'date': '1969-07-21T20:17:40', 'name': 'name_0', 'number': 2, 'page': '100'}, embedding: vector of size 768) != Document(id=8dc93d7bb8817ed727b8fb0fabfbed6a5c130f7c4d4df05c5615eb7ce6b7bcec, content: 'A Foo Document 0', meta: {'name': 'name_0', 'page': '100', 'chapter': 'intro', 'number': 2, 'date': '1969-07-21T20:17:40'}, embedding: vector of size 768)

  Full diff:
    [
  -     Document(id=8dc93d7bb8817ed727b8fb0fabfbed6a5c130f7c4d4df05c5615eb7ce6b7bcec, content: 'A Foo Document 0', meta: {'name': 'name_0', 'page': '100', 'chapter': 'intro', 'number': 2, 'date': '1969-07-21T20:17:40'}, embedding: vector of size 768),
  +     Document(id=8dc93d7bb8817ed727b8fb0fabfbed6a5c130f7c4d4df05c5615eb7ce6b7bcec, content: 'A Foo Document 0', meta: {'chapter': 'intro', 'date': '1969-07-21T20:17:40', 'name': 'name_0', 'number': 2, 'page': '100'}, embedding: vector of size 768),
    ]

I get these failures. Can you run tests locally and fix them? Ping me when works.

[Aftabbs]

Fixed the lint error (B905: zip() without strict= parameter) in commit 19cdde5. The assert_documents_are_equal override now also handles float32 precision loss (commit 6414430) — Pinecone stores embeddings as float32, so round-tripped values differ from the float64 originals; the override checks only embedding presence rather than exact values. Ready for re-review.

[anakin87]

I did some final improvements. Now running all tests locally before merging.

[anakin87]

Thank you