simplify skip-if tests >> 0/n

[Borda]

What does this PR do?

simplify chaining skip-if in pytests

@pytest.mark.skipif(**skipif_args(min_gpus=2, min_torch="99"))
def test_skip():
    ...

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
• Check that target branch and milestone match!

Did you have fun?

Make sure you had fun coding 🙃

[codecov]

Codecov Report

Merging #5920 (0f658a4) into master (40d5a9d) will decrease coverage by 1%.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master   #5920    +/-   ##
=======================================
- Coverage      93%     92%    -1%     
=======================================
  Files         159     159            
  Lines       11380   11589   +209     
=======================================
+ Hits        10624   10660    +36     
- Misses        756     929   +173     

[carmocca]

Why not make it @skipIfNoGPUs instead of @pytest.mark.skipif(**_SKIPIF_NO_GPUS)?

[Borda]

Why not make it @skipIfNoGPUs instead of @pytest.mark.skipif(**_SKIPIF_NO_GPUS)?

the main point was about chaining them together and on the same line...

[carmocca]

the main point was about chaining them together and on the same line...

IMO I'd rather have different conditions in different lines, as is usual in other projects

[Borda]

the main point was about chaining them together and on the same line...

IMO I'd rather have different conditions in different lines, as is usual in other projects

so to have grouped under unitest class, since we have for an example test file where each of the included tests has 5 skips if and all the same for all tests so in the skip is most of the code and not clear reading...
cc: @awaelchli @SeanNaren

[Borda]

Ok, so there are let's say three ways to go:

1. 😄 just to standardize the skip if content so we avoid typos and misalignment of the condition and skip reason so the usage would be just

   @pytest.mark.skipif(**_SKIPIF_PT_LE_1_6)
   @pytest.mark.skipif(**_SKIPIF_NO_APEX)
   @pytest.mark.skipif(**_SKIPIF_NO_GPUS)
   def test_...(...):
       ...

2. 🚀 chaining the skip conditions so it would be a single line and easier to read

   @pytest.mark.skipif(**(_SKIPIF_PT_LE_1_6 | _SKIPIF_NO_APEX | _SKIPIF_NO_GPUS))
   def test_...(...):
       ...

3. 🎉 creating custom wrapper with all these cases as argument (which would increase contribution level as a user needs to read/learn this custom parameterization)

   @lightning_mark_skipif(min_pt=1.6, apex=True, min_gpus=2)
   def test_...(...):
       ...

4. 👀 ready-to-use skipifs

   @gpu_required
   @torch_ge_1_6_required
   @apex_required
   def test():
       ...

5. 👍 creating custom wrapper with all these cases as argument similar to 3)

   @lightning_mark_skipif(_SKIPIF_PT_LE_1_6, _SKIPIF_NO_APEX, _SKIPIF_NO_GPUS)
   def test_...(...):
       ...

6 @

@lightning_mark_skipif(_SKIPIF_PT_LE_1_6, _SKIPIF_NO_APEX, _SKIPIF_NO_GPUS)
def test_...(...):
    ...

less vote a bit... 1: 😄 2: 🚀 3: 🎉 4: 👀 5: 👍

cc: @PyTorchLightning/core-contributors

[Borda]

@tadejsv @SeanNaren @SkafteNicki @akihironitta @rohitgr7 it is not exactly like the promise above as it turned out that the overwriting pytest marks would be tricky and no very sure of any accident unwanted interaction with other pytest marks such as parameterized...

[SkafteNicki]

LGTM :]

[carmocca]

So why don't we provide the skipif ourselves?

return pytest.mark.skipif(condition=any(conditions), reason=reason)

does it not work?

[SeanNaren]

Nice. Should clean up some of these really stacked decorators finally, so nice work!