New tutorial: on how to customise metrics with own user-defined functions

[annahedstroem]

Would be great with a tutorial that shows:

• how to employ user-defined explanation functions into the tests
• how to customise the metric tests like perturb_func, similarity_func etc

Maybe call the tutorial something like: tutorial-customise-metrics-with-user-defined-functions

[p16i]

Regarding similarity_func, I'm not sure to what extent the current API allows, but we could show how to compose a number of similarity_func's together. For example, in randomization tests, I have seen people used both cosine similarity and SSIM. So, with this composition, we could have results for both metrics with one experiment.

[annahedstroem]

@heytitle like the suggestion! would you be interested in contributing with a tutorial including that? :-) feel free to make a PR with it in that case!