Modify `PlotterMixin.transform_plot` to give a dictionary to the call to `plot`

[ulupo]

Reference issues/PRs Fixes a bug introduced in #453.

Types of changes

• [x] Bug fix (non-breaking change which fixes an issue)
• [ ] New feature (non-breaking change which adds functionality)
• [ ] Breaking change (fix or feature that would cause existing functionality to change)

Description In #453, automatic titles were introduced for several (but not all) plot methods. However, this leads to incorrectly reported sample indices when coming from transform_plot of the latter's internal logic. This PR keeps titles but fixes them by modifying the transform_plot method of PlotterMixins so that plot is called not on an array of length 1 but on a dictionary with a single key:

        Xt = self.transform(X[sample:sample+1])
        self.plot(Xt, sample=0, **plot_params).show()

        return Xt

becomes

        Xt = self.transform(X[sample:sample+1])
        self.plot({sample: Xt[0]}, sample=sample, **plot_params).show()

        return Xt

Checklist

• [x] I have read the guidelines for contributing.
• [x] My code follows the code style of this project. I used flake8 to check my Python changes.
• [x] My change requires a change to the documentation.
• [x] I have updated the documentation accordingly.
• [x] I have added tests to cover my changes.
• [x] All new and existing tests passed. I used pytest to check this on Python tests.

[ulupo]

Another "solution" would be to make the outputs of transform in the code for transform_plot into dictionaries with the single key sample (which could be non-zero). Not sure if more or less elegant! It would at least remove the need for sample_orig completely...

[wreise]

Another "solution" would be to make the outputs of transform in the code for transform_plot into dictionaries with the single key sample (which could be non-zero). Not sure if more or less elegant! It would at least remove the need for sample_orig completely...

I don't understand why you wouldn't need sample_orig. Wouldn't you need to pass the key to the VectorisationTransformer.plot method anyway?

[ulupo]

@wreise to make it clearer, this is the proposed fix: the code for transform_plot would be

        Xt = self.transform(X[sample:sample+1])
        self.plot({sample: Xt[0]}, sample=sample, **plot_params).show()

        return Xt

Then you have to change nothing else, I claim. At least in the majority of cases...

[wreise]

@wreise to make it clearer, this is the proposed fix: the code for transform_plot would be

        Xt = self.transform(X[sample:sample+1])
        self.plot({sample: Xt[0]}, sample=sample, **plot_params).show()

        return Xt

IMO, this looks better, even though we feed a dict instead of a ndarray. The "problem" appears only when joining transform on a single sample and plot, and it should be fixed there.

Then you have to change nothing else, I claim. At least in the majority of cases...

Maybe I can test that?

[ulupo]

Maybe I can test that?

Oh please!! :) Feel free to take over this PR. I'm happy to proceed with the dict route.