Raise ModuleNotFoundError when unavailable

[akihironitta]

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 to update the docs?
• Did you write any new necessary tests?

What does this PR do?

Fixes #265.

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 🙃

[pep8speaks]

Hello @akihironitta! Thanks for updating this PR.

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

Comment last updated at 2020-10-22 16:14:00 UTC

[codecov]

Codecov Report

Merging #290 into master will decrease coverage by 0.11%.
The diff coverage is 77.77%.

@@            Coverage Diff             @@
##           master     #290      +/-   ##
==========================================
- Coverage   82.06%   81.94%   -0.12%     
==========================================
  Files          97       98       +1     
  Lines        5441     5501      +60     
==========================================
+ Hits         4465     4508      +43     
- Misses        976      993      +17     

Flag	Coverage Δ
#cpu	23.46% <16.66%> (+0.03%)	⬆️
#pytest	23.46% <16.66%> (+0.03%)	⬆️
#unittests	81.40% <77.77%> (-0.09%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
pl_bolts/datasets/imagenet_dataset.py	20.11% <50.00%> (+0.95%)	⬆️
pl_bolts/datasets/ssl_amdim_datasets.py	76.31% <100.00%> (+1.99%)	⬆️
pl_bolts/utils/semi_supervised.py	96.82% <100.00%> (+0.10%)	⬆️
pl_bolts/models/__init__.py	100.00% <0.00%> (ø)
pl_bolts/models/vision/unet.py	98.03% <0.00%> (ø)
pl_bolts/models/vision/__init__.py	100.00% <0.00%> (ø)
pl_bolts/models/vision/segmentation.py	65.38% <0.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d61e211...1e60395. Read the comment docs.

[Borda]

@akihironitta could you please explore @awaelchli recommendation about importing without try/catch

[akihironitta]

@Borda Thanks for reviewing as always!
As for optional imports, Lightning-AI/pytorch-lightning#2266 (comment) suggests:

if importlib.util.find_spec("some_package") is not None:
    from some_package import my_function

and after that, Lightning-AI/pytorch-lightning#2877 tried implementing pytorch_lightning.utilities.is_some_package_available(), but it's closed without getting merged.

Is there a preferred way of importing optional packages without try/catch? Should I wait for Lightning-AI/pytorch-lightning#2877 to get merged or directly use Lightning-AI/pytorch-lightning#2266 (comment)?

[awaelchli]

@akihironitta
You could do this:

SKLEARN_AVAILABLE = importlib.util.find_spec("some_package") is not None

and then wherever you need it:

if SKLEARN_AVAILABLE:
    # import something, warn or raise

This is my recommendation when you need a variable telling you if the package is available without the need to actually import it.
You can also use ModuleNotFoundError, nothing wrong with that, but when a variable like IS_MODULE_AVAILABLE is required anyway, I think using importlib is cleaner.

Hope this helps.

[akihironitta]

@awaelchli Thank you for the suggestion! I'll start using importlib here.