[RLlib] Policies get/set_state fixes and enhancements.

[sven1977]

Policies currently do not properly return their exploration state when calling Policy.get_state(). This PR adds the exploration state to the return value of policy.get_state(). Exploration.get_info() has been renamed into get_state() (backward compatible). A new Exploration.set_state() method has been added, which is used by Policy.set_state().

This is in preparation of:

• making policies addable to/deletable from a worker's policy_map in-flight
• self-play with 100s of policy snapshots
• league-based training

This may also fix:
#16065

Why are these changes needed?

Related issue number

Checks

• 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 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 :(