[RLlib] Enhance env-rendering callback.

[sven1977]

Enhance example env-rendering callback.

• Allow passing in worker indices on which exclusively to render/log episodes.
• Fix bug where for short rollout-fragment lengths, it could happen that the callback resets the best/worst stats after each sample() call (even those calls in which no episode is finished), thus messing up the best/worst stats and possibly logging/recording the wrong episodes.

Why are these changes needed?

Related issue number

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 :(

[simonsays1980]

LGTM.