Fix mypy errors attributed to pytorch_lightning.loggers.logger.py

[jxtngx]

What does this PR do?

Fix mypy errors attributed to pytorch_lightning.loggers.logger.py for issue #13445

Fixes #<issue_number>

Does your PR introduce any breaking changes? If yes, please list them.

Before submitting

• Was this discussed/approved via a GitHub issue? (not for typos and docs)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure your PR does only one thing, instead of bundling different changes together?
• Did you make sure to update the documentation with your changes? (if necessary)
• Did you write any new necessary tests? (not for typos and docs)
• Did you verify new and existing tests pass locally with your changes?
• Did you list all the breaking changes introduced by this pull request?
• Did you update the CHANGELOG? (not for typos, docs, test updates, or minor internal changes/refactors)

PR review

Anyone in the community is welcome to review the PR.
Before you start reviewing, make sure you have read the review guidelines. In short, see the following bullet-list:

• Is this pull request ready for review? (if not, please submit in draft mode)
• Check that all items from Before submitting are resolved
• Make sure the title is self-explanatory and the description concisely explains the PR
• Add labels and milestones (and optionally projects) to the PR so it can be classified

[jxtngx]

@awaelchli @carmocca et al

regarding logger.merge_dicts:

I've looked at the references and the hierarchy of logger.merge_dicts; it seems to only exist to fulfill a test, and for backwards compatibility in loggers.base.merge_dicts. Neither base.merge_dicts or logger.merge_dicts is used by the base Logger, or any of the third party loggers. The way logger.merge_dicts is implemented violated it's own type hints, causing the following errors:

1. line 409: error: Argument 2 to "merge_dicts" has incompatible type "Optional[Callable[[Sequence[float]], float]]"; expected "Optional[Mapping[str, Callable[[Sequence[float]], float]]]
2. line 411: error: Incompatible types in assignment (expression has type "float", target has type "Dict[Any, Any]")

I updated the agg_key_funcs to be more generic with Optional[Mapping]. Using a defaultdict as d_out during recursion fixed the above errors; the defaultdict is returned as a dict with dict(d_out), satisfying the intended return type.

[awaelchli]

We shouldn't annotate self. self has always the pseudo type "Self"

[jxtngx]

When I remove that annotation, mypy complains:

pytorch_lightning/loggers/logger.py:42: error: Function is missing a type annotation for one or more arguments.

Rather than annotate Self, I commented with # type: ignore[no-untyped-def] to exclude the error in the latest commit.

[carmocca]

This self for the local function definiton is not necessarily the "self" from the actual instance.

[akihironitta]

@JustinGoheen Other than my comments below, it looks good to me!

Also, the following PRs are waiting on checks, is there anything more I need to do before the remaining checks can run?

Let's wait for reviews from others!

[otaj]

Oh, things started failing as ExperimentWriter is undefined. I'll take a closer look at it on Monday

[jxtngx]

@otaj I had set a conditional import of ExperimentWriter; I removed that condition in favor of a hardcoded import.

[otaj]

@JustinGoheen Great :)