Dice score as metric

[MrShevan]

What does this PR do?

Fixes #992

Before submitting

• Was this discussed/approved via a Github issue? (no need for typos and docs improvements)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure to update the docs?
• Did you write any new necessary tests?

PR review

Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.

Did you have fun?

Make sure you had fun coding 🙃

[justusschock]

Hey, looking pretty good already! Thanks a lot for your implementation!
Left some minor comments. Most of them are just to use immutable types if possible :)

The only one that concerns me a bit is the one about testing :)

[codecov]

Codecov Report

Merging #1021 (508fbe1) into master (6b96ebf) will decrease coverage by 0%.
The diff coverage is 80%.

@@          Coverage Diff           @@
##           master   #1021   +/-   ##
======================================
- Coverage      95%     95%   -0%     
======================================
  Files         180     181    +1     
  Lines        7828    7873   +45     
======================================
+ Hits         7435    7468   +33     
- Misses        393     405   +12     

[justusschock]

Also: should we update the jaccard implementation to rely on this implementation? Calculating them like this should be faster than computing a whole conf map and J = D/(2-D)

cc @SkafteNicki

[SkafteNicki]

@justusschock I would be fine on relying on the Jaccard implementation for this, however I think we should be careful extrapolating the J = D/(2-D) to all cases. I am not completely sure that this formula holds when you take class averaging, multidim averaging etc. into account. That at least need to be checked.

[MrShevan]

@justusschock, May I ask you for help?) I see that all failing checks caused by installing of dependencies. Could this be related to using scipy? Should I add something in PR to avoid it?

[justusschock]

@Borda mind having a look on failing audio deps?

[justusschock]

One more thing left to do from my side and then it's ready to go :)

Can you raise a deprecationwarning in the older implementation asking users to switch to the newer one?

[MrShevan]

@justusschock thank you for your comment :) I've added deprecation warning.

[justusschock]

@Borda there are plenty of tests that did not run on this PR. Why did you merge it like that?

[MrShevan]

@justusschock Did I forget something? I can open a new one and add the necessary tests. This will be inside v0.9

[justusschock]

No, it's fine now, I restarted the tests and they passed, they just weren't finished when this was merged :D

[MrShevan]

Oh, i got it :) thanks for all your advice @Borda @justusschock @SkafteNicki! I get a lot of experience working on your project

[Borda]

Why did you merge it like that?

I did auto-merge so all suppose to be fine at that time... or what I missed?

[justusschock]

Almost all our CI configs are not strictly required, but they didn't even run (canceled for some reason)