feat: add metadata exploration methods to MongoDBAtlasDocumentStore

[Amanbig]

Planned Work

• Implement count_documents_by_filter
  (following the behavior of OpenSearchDocumentStore)

• Implement count_unique_metadata_by_filter
  (aligned with existing implementations in other document stores)

• Implement get_metadata_fields_info
  (using sample-based inference, similar to OpenSearch)

• Implement get_metadata_field_min_max

• Implement get_metadata_field_unique_values
  (consistent with the MongoDB implementation in add the following operations to MongoDBDocumentStore #2640)

• Add async counterparts for all new methods

• Add integration tests for the new functionality

---

Related Issues

• Fixes add the following operations to MongoDBDocumentStore #2640

---

Proposed Changes

This work adds missing metadata and aggregation operations to MongoDBDocumentStore, bringing it in line with other Haystack document stores.
The implementation follows patterns already used in OpenSearch and MongoDB integrations to keep behavior consistent across backends.

---

How Did You Test It?

• Added integration tests for all new methods
• Ensured parity with MongoDB and OpenSearch implementations

---

Notes for the Reviewer

• Logic mirrors the MongoDB implementation from add the following operations to MongoDBDocumentStore #2640 where applicable
• Sync and async implementations share core logic

---

Checklist

• I have read the [Contributing Guidelines](https://github.com/deepset-ai/haystack-core-integrations/blob/main/CONTRIBUTING.md) and the [Code of Conduct](https://github.com/deepset-ai/haystack-core-integrations/blob/main/CODE_OF_CONDUCT.md)
• I have updated the related issue with progress and insights
• I added tests and updated docstrings
• I used a conventional commit type (feat:)

[CLAassistant]

All committers have signed the CLA.

[davidsbatista]

The following methods need to be implemented

• get_metadata_fields_info_async()
• get_metadata_field_min_max_async()
• get_metadata_field_unique_values_async()

[Amanbig]

Thanks for the review, @davidsbatista!
I agree — the async and sync implementations share common logic. I’ll refactor the duplicated parts into a shared helper and update both methods accordingly.

I’ll also go ahead and implement the remaining async methods:

• get_metadata_fields_info_async()
• get_metadata_field_min_max_async()
• get_metadata_field_unique_values_async()

Will push the updates shortly.