Unify env step interface

[kengz]

Unify env step interface

• update env.step interface to be consistent with gym return: state, reward, done, info
• this includes only interface changes and no logic changes.
• methods changed and propagated include:
  • env.step
  • env.space_step
  • control env.step
• additionally, also reorder ENV_DATA_NAMES to follow the same convention state, reward, done, and propagate accordingly