newer version of gym requires rewards to be float

[haruishi43]

Description

Newer versions of gym>=0.20.0 requires reward returned by step() to be float. The new passive_env_checker.py checks if the environment complies with this or not. This error could not be avoided with writing custom wrappers that converts the original int reward to float.

Error log:

  ...
  File "/home/ubuntu/.pyenv/versions/3.8.8/lib/python3.8/site-packages/gym/utils/passive_env_checker.py", line 278, in passive_env_step_check
    assert isinstance(
AssertionError: The reward returned by `step()` must be a float

Type of change

Please select all relevant options:

• [x] Bug fix (non-breaking change which fixes an issue)
• [ ] New feature (non-breaking change which adds functionality)
• [ ] MGMT (non-breaking change to deployment, CI, etc.

Checklist

• [x] My code follows the style guidelines of this project
• [x] I have performed a self-review of my own code
• [x] I have commented my code, particularly in hard-to-understand areas
• [x] I have made corresponding changes to the documentation
• [x] I have added tests that prove my fix is effective or that my feature works

[haruishi43]

Might need to update nes_py as well. And I noticed that the reward range is a tuple of ints, which might still raise the error. https://github.com/Kautenja/nes-py/blob/bd1b06448e29675b82eaf2953fa83886250dc67f/nes_py/nes_env.py#L303

[Kautenja]

Version 8.2.0 of nes-py resolves this issue for all downstream environments. Thanks for bringing it to my attention and making the suggestion!