Pure package & base tests

[Borda]

What does this PR do?

Add test for the package without extras

TODO

• add doctest to run Trainer with single epoch

Before submitting

• Was this discussed/approved via a Github issue? (no need for typos and docs improvements)
• 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? Otherwise, we ask you to create a separate PR for every change.
• Did you make sure to update the documentation with your changes?
• Did you write any new necessary tests?
• Did you verify new and existing tests pass locally with your changes?
• If you made a notable change (that affects users), did you update the CHANGELOG?

PR review

Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.

Did you have fun?

Make sure you had fun coding 🙃

[awaelchli]

great idea! this should really help prevent bugs like #2410 sneaking in!

[codecov]

Codecov Report

Merging #2418 into master will decrease coverage by 0%.
The diff coverage is n/a.

@@          Coverage Diff           @@
##           master   #2418   +/-   ##
======================================
- Coverage      89%     89%   -0%     
======================================
  Files          69      69           
  Lines        5477    5475    -2     
======================================
- Hits         4857    4855    -2     
  Misses        620     620           

[Borda]

@awaelchli any idea why Windows does not have any new PyTorch?

ERROR: Could not find a version that satisfies the requirement torch>=1.3 (from -r ./requirements/base.txt (line 4)) (from versions: 0.1.2, 0.1.2.post1, 0.1.2.post2)
ERROR: No matching distribution found for torch>=1.3 (from -r ./requirements/base.txt (line 4))

on the other hand, complete testing works fine... maybe because it is cached?

EDIT: resolved :]

[Borda]

this seems like a Win issue with PT...

pytorch_lightning\core\__init__.py:339: in <module>
    from pytorch_lightning.core.decorators import data_loader
pytorch_lightning\core\decorators.py:4: in <module>
    import torch
C:\hostedtoolcache\windows\Python\3.7.7\x64\lib\site-packages\torch\__init__.py:81: in <module>
    ctypes.CDLL(dll)
C:\hostedtoolcache\windows\Python\3.7.7\x64\lib\ctypes\__init__.py:364: in __init__
    self._handle = _dlopen(self._name, mode)
E   OSError: [WinError 126] The specified module could not be found

EDIT: resolved with installing CPU versions...
even CPU based machine installs CUDA version... :]

[mergify]

This pull request is now in conflict... :(

[Borda]

@awaelchli @williamFalcon mind check the final version?

[awaelchli]

can we not leave it here?
save_hyperparameters will only save in_dim and out_dim, right?
the user should still be able to assign whatever they want to self.

[Borda]

yes, he does but we do not want to recommend this practice, right? @williamFalcon

[williamFalcon]

people can do whichever but it’s better to use the save call for hparams tracking

[awaelchli]

if you want, you can move these to the top of the file, they don't necessarily need to be inside the doctest.

[Borda]

but they are not called outside test flake8 may complain about unused impprts

[awaelchli]

it seems since you removed circle CI, we cannot preview docs anymore directly in the CI artifacts. Maybe we can add it back later.

[Borda]

it seems since you removed circle CI, we cannot preview docs anymore directly in the CI artifacts. Maybe we can add it back later.

well we are exporting build-docs with GH action but it is zip format, we may check if there is a way to browse it online or add CircleCI just for the docs? not sure how many people really used it... @PyTorchLightning/core-contributors?

EDIT: Just for the record CircleCI was dropped in #2378 #2412