Explicitly Export PartitionMetadata to Satisfy Type Checkers

[George Berdovskiy (george-sentry)]

What

At Sentry, we use this Python package to write Kafka consumers and producers. Some of our references the PartitionMetadata class, which is defined but not explicitly exported. As a result, when we attempt to import this class (and possibly others), type checking fails with an error like this one.

$ python -m mypy
src/sentry/monitors/tasks/clock_pulse.py:11: error: Module "confluent_kafka.admin" does not explicitly export attribute "PartitionMetadata"  [attr-defined]
tests/sentry/monitors/tasks/test_clock_pulse.py:6: error: Module "confluent_kafka.admin" does not explicitly export attribute "PartitionMetadata"  [attr-defined]
Found 2 errors in 2 files (checked 7783 source files)

This requires working around the type checker (specifically mypy) with # type: ignore[attr-defined], defeating the purpose of using types in the first place.

Checklist

• Contains customer facing changes? Yes, the class PartitionMetadata is now explicitly exported.
• Did you add sufficient unit test and/or integration test coverage for this PR? No, because runtime behavior is unaffected.

References

N/A

Test & Review

I tested these changes by pointing the Sentry monolith to this version of confluent-kafka instead of the published one. Follow these instructions to set up Sentry locally.

$ brew install librdkafka

$ export CPATH="$(brew --prefix)/include${CPATH:+:$CPATH}"
$ export LIBRARY_PATH="$(brew --prefix)/lib${LIBRARY_PATH:+:$LIBRARY_PATH}"
$ export PKG_CONFIG_PATH="$(brew --prefix)/lib/pkgconfig${PKG_CONFIG_PATH:+:$PKG_CONFIG_PATH}"

$ uv pip install -e ~/Desktop/Repositories/confluent-kafka-python --python .venv/bin/python

$ python -m mypy
Success: no issues found in 7783 source files

Open Questions

• Is there a good reason for PartitionMetadata to not be publicly exported?
• Are there any other types that should be explicitly exported?

[confluent-cla-assistant]

🎉 All Contributor License Agreements have been signed. Ready to merge.
✅ george-sentry
_{Please push an empty commit if you would like to re-run the checks to verify CLA status for all contributors.}

[Copilot]

Pull request overview

This PR makes PartitionMetadata explicitly re-exported from confluent_kafka.admin so type checkers (e.g., mypy with strict re-export rules) accept from confluent_kafka.admin import PartitionMetadata without requiring # type: ignore[attr-defined].

Changes:

• Split the _metadata imports so PartitionMetadata is explicitly re-exported via from ... import PartitionMetadata as PartitionMetadata.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

This adds an explicit re-export for PartitionMetadata, but the rest of the re-exported symbols in this module are still implicit from a type-checker perspective. To avoid having to special-case additional symbols later (and to make the public surface area unambiguous), consider defining an explicit __all__ for the admin package (or consistently using from ... import X as X for all intended re-exports).