Add run_name argument to the MLFlowLogger constructor

[i-aki-y]

What does this PR do?

This PR adds a run_name argument to the MLFlowLogger constructor.
As is shown in #7571, the run_name is used to identify the specific execution (run) in mlflow.
By default, the run_name is the same value as the run_id that is a UUID.
So, mlflow allows the user to set more informative run_name in mlflow.start_run method argument^[1].

The current MLFlowLogger does not support setting the run_name explicitly.
(As I pointed out in #7571, there is a workaround, but it is hard to notice, I think.)
So, in this PR, the run_name argument is added to the MLFlowLogger.

To make this PR, I referred to the current implementation of mlflow.start_run^[1]

I would appreciate it if you could review this PR.

[1] https://www.mlflow.org/docs/latest/python_api/mlflow.html#mlflow.start_run

Fixes #7571

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 update the CHANGELOG? (not for typos, docs, test updates, or internal minor changes/refactorings)

PR review

Anyone in the community is free to review the PR once the tests have passed.
Before you start reviewing make sure you have read 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

Did you have fun?

Make sure you had fun coding 🙃
👍

[codecov]

Codecov Report

Merging #7622 (d0b3900) into master (20f6337) will decrease coverage by 4%.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master   #7622    +/-   ##
=======================================
- Coverage      92%     88%    -4%     
=======================================
  Files         197     198     +1     
  Lines       12878   12942    +64     
=======================================
- Hits        11899   11380   -519     
- Misses        979    1562   +583     

[justusschock]

can we maybe lift this to the init already? So that self.tags is always a dict?

[i-aki-y]

Thanks. I have a concern about this. The current self.tags looks to be a public attribute (it has no prefix '_'), so I worried that the user could set None to it after the initialization.

Note that the document does not say the tags are public, so I'm not sure it is intentional:
https://pytorch-lightning.readthedocs.io/en/latest/api/pytorch_lightning.loggers.mlflow.html#mlflow-logger

Another minor reason why I put this here is to minimize the side effect. If the run_name is not set explicitly, the self.tags still be 'None' after this PR. this is the same behavior as the current version.

I could not judge which is better, so any suggestions and comments are welcome!

[tchaton]

LGTM !