ppai: fix lint errors

[davidcavazos]

Description

Fixes#7786

Checklist

• I have followed Sample Guidelines from AUTHORING_GUIDE.MD
• README is updated to include all relevant information
• Tests pass: nox -s py-3.6 (see Test Environment Setup)
• Lint pass: nox -s lint (see Test Environment Setup)
• These samples need a new API enabled in testing projects to pass (let us know which ones)
• These samples need a new/updated env vars in testing projects set to pass (let us know which ones)
• Please merge this PR for me once it is approved.
• This sample adds a new sample directory, and I updated the CODEOWNERS file with the codeowners for this sample

[dandhlee]

TypeError: 'type' object is not subscriptable D:

[leahecole]

TypeError: 'type' object is not subscriptable D:

Ah crap - this might be a backwards compatibility thing. If that's the case and we need to maintain compatibility with older versions of Python and that means we have to use Any, that's an okay justification.

[dandhlee]

Few more issues:

 File "/workspace/people-and-planet-ai/timeseries-classification/create_datasets.py", line 32, in <module>
train_eval_split: list[int],
TypeError: 'type' object is not subscriptable

• 404 on models

Is the 404 error flaky? Or is it an actual error?

[davidcavazos]

It looks like the Dataflow pipeline failed with File "train_model.py", line 129, in get_image NameError: name 'beam' is not defined [while running 'Get image-ptransform-59']. That's weird because we're passing save_main_session=True to the pipeline, so it should have all the imports (including beam).

[davidcavazos]

It looks like there's some sort of exception on saving the global context for the Beam library itself. It looks like it's something that changed on a more recent version because it wasn't an issue before. Importing beam explicitly inside those functions seems to fix it.

[davidcavazos]

It's not finding the AutoML model to do predictions. The model was a permanent resource and it somehow got deleted. I'm going to make a fallback so if it doesn't find the model in AutoML, it'll use a mock implementation to simply test that all methods are called correctly.