[Doc] Fix broken data docs; reenable nitpicky mode

[peytondmurray]

Why are these changes needed?

This PR fixes a broken link and type annotation in Ray Train, and re-enables nitpicky mode in the docs.

There were two issues affecting the documentation:

1. The autosummary of ray.data.dataset.Schema was incorrectly referencing ray.data.Schema
2. The docstring of Schema contained entries (types and names) under the Attribute section which were already documented using docstrings of their respective properties. The build error that you get when you build the documentation unhelpfully suggested using noindex, but instead we just need to either use the docstrings in Attributes or the docstrings under the property methods, not both.

Related issue number

Reverts #46241. Blocked by #46280.

Checks

• I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
• I've run scripts/format.sh to lint the changes in this PR.
• I've included any doc changes needed for https://docs.ray.io/en/master/.
  • I've added any new APIs to the API Reference. For example, if I added a
    method in Tune, I've added it in doc/source/tune/api/ under the
    corresponding .rst file.
• I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
• Testing Strategy
  • Unit tests
  • Release tests
  • This PR is not tested :(

[can-anyscale]

thanks for fixing!

[bveeramani]

What happens if we exclude ".*"?

[peytondmurray]

We get a bunch of false error messages from trouble that Sphinx has with type generics, which are used as annotations in various places in the code base. There's definitely other ways to deal with this issue, but this is the most straightforward right now.

In the future it might be worth investigating ways to remove this rule, because we'll be better protected against invalid class references, but for now I think this is a good start.

[can-anyscale]

request change since the name of the api is incorrect (see comments); let me know otherwise

[bveeramani]

Thanks!