deprecate metrics pkg 1/n

[Borda]

What does this PR do?

This is the initial PR with deprecation process, here we drop docs and forward the user to the new package

In follow-up PR we will ... (suggesting one metric ~ one PR)

1. connect particular metric to torchmetrics - import the relevant metric from TM and replace all function body PL metric by a call to the TM version of the same metric
2. add a deprecation warning - just after the func signature or class init add a warning
3. test deprecation warning
4. drop particular metric test

see for example:
https://github.com/PyTorchLightning/pytorch-lightning/blob/master/pytorch_lightning/metrics/functional/classification.py#L375-L415

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 🙃

[pep8speaks]

Hello @Borda! Thanks for updating this PR.

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2021-03-15 12:06:35 UTC

[tchaton]

LGTM !

[carmocca]

@Borda with the fixes in #8002 this warning appears every time pl is imported. Should avoid that

[Borda]

Do you mean imported in the same running session?
we can have an internal counter to make sure it is raised only once...

[carmocca]

I mean the following:

$ python -c 'import pytorch_lightning'
/Users/carlosmocholi/git/pytorch-lightning/pytorch_lightning/metrics/__init__.py:43: LightningDeprecationWarning: `pytorch_lightning.metrics.*` module has been renamed to `torchmetrics.*` and split off to its own package (https://github.com/PyTorchLightning/metrics) since v1.3 and will be removed in v1.5
  rank_zero_deprecation(

[dlangerm]

I'm getting major log spam with this warning on windows and it appears there is no way to turn it off.

import warnings
# from pytorch_lightning.utilities.warnings import LightningDeprecationWarning
# Even specifying LightningDeprecationWarning below does not work
warnings.filterwarnings("ignore", category=DeprecationWarning) 

import pytorch_lightning

The above code causes the deprecation warning to print, so when I have a data loader with multiple workers, this message spams the terminal.

[carmocca]

Yeah, we should remove it as it's too intrusive. Would you like to open a PR?