handle ImportError and OSError in extras.pytorch

[cosmicBboy]

Signed-off-by: Niels Bantilan [email protected]

TL;DR

The introduction of "soft" dependencies in flytekit.extras to support pytorch types leads to an OSError when trying to import torch: #1032. This PR adds OSError as a potential exception raised when installing pytorch.

Type

• Bug Fix
• Feature
• Plugin

Are all requirements met?

• Code completed
• Smoke tested
• Unit tests added
• Code documentation added
• Any pending items have an associated Issue

Complete description

For context, see this slack thread: https://unionai.slack.com/archives/C01H0FN1NJX/p1661365911337069

Flytekit has a plugin system, hosted as separate packages in this directory of the flytekit repo. However, since Flyte is a ML- and Data-aware orchestration tool, it would make sense for the core ML frameworks to have first class support without having to install any plugins.

Therefore, we're establishing a new pattern for plugins in flytekit.extras that has "soft" dependencies on packages. For example, if the user has pytorch installed, the type plugins should automatically be registered. If the user doesn't, it should gracefully handle errors that might result from importing torch.

Tracking Issue

NA

Follow-up issue

NA
OR
https://github.com/flyteorg/flyte/issues/

[codecov]

Codecov Report

Merging #1141 (c42638b) into master (c4ba45a) will decrease coverage by 0.01%.
The diff coverage is 0.00%.

@@            Coverage Diff             @@
##           master    #1141      +/-   ##
==========================================
- Coverage   68.35%   68.34%   -0.02%     
==========================================
  Files         287      287              
  Lines       25911    25915       +4     
  Branches     2900     2901       +1     
==========================================
  Hits        17711    17711              
- Misses       7721     7725       +4     
  Partials      479      479              

Impacted Files	Coverage Δ
flytekit/extras/pytorch/__init__.py	0.00% <0.00%> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.