feat(supabase): add SupabaseGroongaDocumentStore and SupabaseGroongaRetriever

[ShubhamGond105]

Related Issues

• fixes SupabaseGroongaDocumentStore and SupabaseGroongaRetriever #3083

Proposed Changes:

Adds SupabaseGroongaDocumentStore and SupabaseGroongaRetriever for full-text search powered by PGroonga, a PostgreSQL extension for fast multilingual full-text search exposed by Supabase.

• SupabaseGroongaDocumentStore: stores and searches documents using PGroonga full-text search — no embeddings needed
• SupabaseGroongaRetriever: retrieves documents using plain text queries
• Follows the same structure as the merged PR feat: Add Supabase integration with PgVector document store and embedding retriever  #3164 (SupabasePgvectorDocumentStore)
• Can be used alongside SupabasePgvectorEmbeddingRetriever in hybrid search pipelines

How test

19 unit tests covering:

• Default and custom init parameters
• write_documents, delete_documents, filter_documents, count_documents
• Serialization round-trip (to_dict / from_dict)
• Retriever init, run, and serialization
• Empty query handling
• DuplicatePolicy (OVERWRITE, SKIP, FAIL) handling

Notes for the reviewer

• Followed the same structure as PR feat: Add Supabase integration with PgVector document store and embedding retriever  #3164 as a blueprint
• Used supabase-py client for all database operations
• PGroonga must be enabled on the Supabase project as a prerequisite
• Happy to make any adjustments based on feedback

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: feat:

[ShubhamGond105]

Hi @davidsbatista,
I'm actively working on fixing the remaining CI issues:

count argument type error — need to use CountMethod.exact instead of "exact" string
result.data union-attr error in _groonga_retrieval — need stronger type guard
Will push fixes shortly.
Happy to get any other feedback in the meantime!

[davidsbatista]

Can you confirm that the client doesn't support async operations? You did not add any do the SupabaseGroongaDocumentStore

[davidsbatista]

Can you also define an async def run_async(...) ? Does the client supports it?

[davidsbatista]

Hi @ShubhamGond105, thanks for your contribution. I left some comments that need to be addressed; also, the type-checking issues need to be fixed.

In the future, please make sure the CI is green before asking for a review, unless the reason why is not green is out of your control and needs our action.

Thanks once again!

[github-actions]

Coverage report (supabase)

Click to see where and how coverage changed

File	Statements	Missing	Coverage	Coverage (new stmts)	Lines missing
integrations/supabase/src/haystack_integrations/components/retrievers/supabase
groonga_retriever.py					133-135
integrations/supabase/src/haystack_integrations/document_stores/supabase
groonga_document_store.py					93-94, 97, 125-126, 141-142, 172-179, 196-197, 216-217, 223-227, 238-239, 260-261, 273, 285-314
Project Total

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

[ShubhamGond105]

"Hi @davidsbatista, I've addressed all the feedback points:

Moved client initialization to warm_up()
Added DocumentStore base class
Fixed default DuplicatePolicy to FAIL
Implemented filter support
Added run_async() to retriever
Fixed all mypy and lint errors

All tests passing locally. Please take a look when you get a chance!"

[davidsbatista]

Tests are failing on the CI; we need to have them fixed

[davidsbatista]

There are no integration tests:
- Zero integration tests exist; the checklist requires at minimum inheriting DocumentStoreBaseTests
- Add a Docker Compose service with PGroonga, or add @pytest.mark.integration tests that skip when the service is unavailable

There are no integration tests at all. Please also note that Integration tests should inherit from DocumentStoreBaseTests (and optionally DocumentStoreBaseExtendedTests) to cover all standard document store behaviours. This is completely absent.

Please, read through this article https://docs.haystack.deepset.ai/docs/creating-custom-document-stores to understand in more detail what's missing in your PR

Please make sure you go through this, we need to have integration tests.

[davidsbatista]

Suggested change

	supabase_key: Secret = Secret.from_env_var("SUPABASE_SERVICE_KEY"),
	supabase_key: Secret = Secret.from_env_var("SUPABASE_SERVICE_KEY", strict=False)

[davidsbatista]

There are still a few unresolved issues, and no integration tests - please go through them and let me know if you need any help.

When in doubt, look at how other DocumentStores work (e.g., retrievers logic, filters, etc.) and replicate it - ElasticSearch and OpenSearch document stores have good patterns to follow

[davidsbatista]

Hi @ShubhamGond105 - do you still have time to work on this?

[ShubhamGond105]

Yes but I need more time to fix it.