[ENH] RDSTClassifier and transformer tests

[baraline]

Reference Issues/PRs

What does this implement/fix? Explain your changes.

• Add a classifier warper for RDST with a ridge classifier
• Delete some docstring on RDST transformer that were not up-to-date with the current implementation.
• Debug some test functions
• Add sanity checks and debug edges cases for RDST Transformer

Does your contribution introduce a new dependency? If yes, which one?

No.

What should a reviewer concentrate their feedback on?

Check if the implementation of RDSTClassifier is matching the standards of aeon.

Did you add any tests for the change?

I added a default test parameter set for the get_test_params method.

PR checklist

For all contributions

• I've added myself to the list of contributors.
• Optionally, I've updated aeon's CODEOWNERS to receive notifications about future changes to these files.
• The PR title starts with either [ENH], [MNT], [DOC], or [BUG] indicating whether the PR topic is related to enhancement, maintenance, documentation, or bug.
• For new estimators, I've added the estimator to the online documentation.

[MatthewMiddlehurst]

Will get back to you Re: MacOS if i can figure it out

[MatthewMiddlehurst]

Macs are miserable. But either the latest commit solves the different results or my VM does not work as intended.

Tried for a while to get a UEA Macbook working then found an old setup for Catch22 which used to have the same issue!

[baraline]

I was getting numba type error for passing a RandomState object to random_dilated_shapelet_extraction rather than an integer on my windows (what i changed in the latest commit). Hopefully it's this ? If not, I'm a bit out of ideas ...

[MatthewMiddlehurst]

Nope, I would just remove/skip/comment out the test for now. Setting up a Mac will take a little bit more work on my end unless anyone else can get one.

[baraline]

I added some checks and the end of the fit method to avoid cases where we have no shapelets or NaN values in the shapelets values. So now more informative errors will be raised if these cases happen. Tests for the classifier seem to be OK on MacOS.

[MatthewMiddlehurst]

Also there appears to be an out of index error on one of the numba functions, see https://github.com/aeon-toolkit/aeon/actions/runs/5565449870/jobs/10165816706 (numba does not check for this by default)

[baraline]

This is due to a mistake in the data given to the function in the test for. This should now work as intended, I needed to transpose the array to have the expected shape from the testing data. This should not impact normal execution as the correct shape is given by the sliding_mean_std_one_series function in the transform.

[baraline]

Is the readthedocs issue due to something wrong with the docs ? Only see a timeout on the errors

[MatthewMiddlehurst]

LGTM.

Docs issue is not related to this PR, happening on other PRs and we have to figure out why!