[BUG] Fix issue with plot_prediction_actual_by_variable unsupported operand type(s) for *: 'numpy.ndarray' and 'Tensor'

[fnhirwa]

fixes #1822

This PR fixes the issue with plot_prediction_actual_by_variable where the input np.ndarray was being multiplied to torch.Tensor in TorchNormalizer __call__

[codecov]

Codecov Report

Attention: Patch coverage is 75.00000% with 3 lines in your changes missing coverage. Please review.

Please upload report for BASE (main@cc6a3d7). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
pytorch_forecasting/data/encoders.py	75.00%	3 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1903   +/-   ##
=======================================
  Coverage        ?   86.28%           
=======================================
  Files           ?       96           
  Lines           ?     7775           
  Branches        ?        0           
=======================================
  Hits            ?     6709           
  Misses          ?     1066           
  Partials        ?        0           

Flag	Coverage Δ
cpu	86.28% <75.00%> (?)
pytest	86.28% <75.00%> (?)

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

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[fkiraly]

we should add a test for this, if it fixes a bug

[fkiraly]

Can you explain why we are changing tests, and not adding one? I was expecting a new tests.

In either case, I may need an explanation on your changes to the tests, that would be appreciated.

[fnhirwa]

Can you explain why we are changing tests, and not adding one? I was expecting a new tests.

In either case, I may need an explanation on your changes to the tests, that would be appreciated.

In the old test cases we were defaulting to torch.Tensor as a testing example

pytorch-forecasting/tests/test_data/test_encoders.py

Line 82 in cc6a3d7

kwargs.setdefault("data", torch.rand(100))

With this modification we can cover both np.ndarray, pd.Series and torch.Tensor as inputs to the Normalizer.

[fkiraly]

oh, you just moved it to the mark.parametrize, and added code to translate string to the object, is that correct?

Minor, non-blocking comment: my gut feeling says pytest.fixture in this case, for the str-to-object code, but it is clear enough what is happening (so, not blocking)