Add ckpt_path option to LightningModule.test()

[yukw777]

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?
• If you made a notable change (that affects users), did you update the CHANGELOG?

What does this PR do?

Fixes #2046.
If ckpt_path is "best" (default), it loads the best weights saved by ModelCheckpoint for the test loop.
If ckpt_path is a path to a checkpoint file, it loads the weights from the file for the test loop.
If ckpt_path is None, it uses the weights from the end of training for the test loop.
If model parameter is set, ckpt_path is ignored.

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 🙃

[yukw777]

@williamFalcon @Borda I finally found some time to work on this, and I saw that you guys have already finalized the changelog entry for 0.8, so I wanted to check with you which version this should go. Obviously, I think it'd be great if we could squeeze this in with 0.8, but I don't want to create unnecessary delays in the release, so let me know!

[SkafteNicki]

This is an awesome addition, maybe we should add some documentation about this behavior to:
https://github.com/PyTorchLightning/pytorch-lightning/blob/master/docs/source/test_set.rst