Fix auc calculation and add tests

[maximsch2]

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?

What does this PR do?

Fixes #195
Closes #196

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 🙃

[codecov]

Codecov Report

Merging #197 (92582c7) into master (2584345) will increase coverage by 0.02%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #197      +/-   ##
==========================================
+ Coverage   96.63%   96.65%   +0.02%     
==========================================
  Files         180      180              
  Lines        5760     5738      -22     
==========================================
- Hits         5566     5546      -20     
+ Misses        194      192       -2     

Flag	Coverage Δ
Linux	79.88% <100.00%> (-0.05%)	⬇️
Windows	79.88% <100.00%> (-0.05%)	⬇️
cpu	96.65% <100.00%> (+0.02%)	⬆️
gpu	96.65% <ø> (+0.02%)	⬆️
macOS	96.65% <100.00%> (+0.02%)	⬆️
pytest	96.65% <100.00%> (+0.02%)	⬆️
python3.6	95.55% <100.00%> (+0.01%)	⬆️
python3.8	96.61% <100.00%> (+0.02%)	⬆️
python3.9	96.51% <100.00%> (+0.02%)	⬆️
torch1.3.1	95.55% <100.00%> (+0.01%)	⬆️
torch1.4.0	95.67% <100.00%> (+0.01%)	⬆️
torch1.8.1	96.51% <100.00%> (+0.02%)	⬆️

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

Impacted Files	Coverage Δ
torchmetrics/utilities/data.py	96.42% <ø> (+0.97%)	⬆️
torchmetrics/functional/classification/auc.py	86.95% <100.00%> (-0.55%)	⬇️
__w/1/s/torchmetrics/average.py
__w/1/s/torchmetrics/retrieval/retrieval_recall.py
...s/torchmetrics/retrieval/mean_average_precision.py
...chmetrics/functional/classification/stat_scores.py
__w/1/s/torchmetrics/utilities/distributed.py
...1/s/torchmetrics/functional/retrieval/precision.py
...ics/functional/classification/matthews_corrcoef.py
...chmetrics/classification/precision_recall_curve.py
... and 172 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2584345...92582c7. Read the comment docs.

[maximsch2]

Stable sort that we are trying to do doesn't work: https://colab.research.google.com/gist/maximsch2/4341c3b01482f8b38ea0f0ed70affd33/stable-sort-in-pytorch.ipynb

[SkafteNicki]

@maximsch2 I see your point that no matter what we do, we will never get stable sorting on cpu until torch v1.9 is released. Could you please add a note to the affected metrics, that users can expect slightly different results on cpu vs gpu.
Also add a TODO that this is something we probably should revisit in the future.

[pep8speaks]

Hello @maximsch2! Thanks for updating this PR.

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2021-04-23 14:17:17 UTC