Add __repr__ and __str__ to Metrics baseclass

[MathijsdeBoer]

Description

When training a model using MONAI metrics for experiment tracking, I tend to log which metrics I am using. Unfortunately, just sending the metrics objects to Tensorboard will result in a list like [CustomMetric1, CustomMetric2, <monai.metrics.Metric ...>, etc.]

Adding __repr__ and __str__ methods to the base class will solve this small annoyance. The current implementation will only return the class name, but if a certain metric would wish to report more data for its __repr__ string, this can be easily overridden in any subclass.

Types of changes

• Non-breaking change (fix or new feature that would not break existing functionality).
• Breaking change (fix or new feature that would cause existing functionality to change).
• New tests added to cover the changes.
• Integration tests passed locally by running ./runtests.sh -f -u --net --coverage.
• Quick tests passed locally by running ./runtests.sh --quick --unittests --disttests.
• In-line docstrings updated.
• Documentation updated, tested make html command in the docs/ folder.

[MathijsdeBoer]

Ideally one would exist for the Loss classes, too. But the baseclass for them is torch.nn.modules.loss._Loss, so either one could make a subclass of that class, add the __repr__ and __str__ methods to it, and point all existing loss classes to the new one instead, or these methods could be added to the baseclass in PyTorch with a pull request in that project.

[MathijsdeBoer]

I always forget to do the signoff for these pull requests, followed the rebase method on the details page to fix it!

[KumoLiu]

/build

[KumoLiu]

/build

[KumoLiu]

/build