Change kwargs=None optional parameter to **kwargs

[TimOgden]

When running the t2t_decoder.py script from magenta, this line of code will result in this traceback:

Traceback (most recent call last):
  File "C:\Users\togde\Anaconda3\envs\magenta\Scripts\t2t_decoder-script.py", line 33, in <module>
    sys.exit(load_entry_point('magenta', 'console_scripts', 't2t_decoder')())
  File "C:\Users\togde\Anaconda3\envs\magenta\Scripts\t2t_decoder-script.py", line 25, in importlib_load_entry_point
    return next(matches).load()
  File "C:\Users\togde\Anaconda3\envs\magenta\lib\site-packages\importlib_metadata\__init__.py", line 194, in load
    module = import_module(match.group('module'))
  File "C:\Users\togde\Anaconda3\envs\magenta\lib\importlib\__init__.py", line 127, in import_module
    return _bootstrap._gcd_import(name[level:], package, level)
  File "<frozen importlib._bootstrap>", line 1006, in _gcd_import
  File "<frozen importlib._bootstrap>", line 983, in _find_and_load
  File "<frozen importlib._bootstrap>", line 967, in _find_and_load_unlocked
  File "<frozen importlib._bootstrap>", line 677, in _load_unlocked
  File "<frozen importlib._bootstrap_external>", line 728, in exec_module
  File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed
  File "c:\users\togde\programmingprojects\magenta\magenta\tensor2tensor\t2t_decoder.py", line 19, in <module>
    from magenta.tensor2tensor import models  # pylint: disable=unused-import
  File "c:\users\togde\programmingprojects\magenta\magenta\tensor2tensor\models.py", line 18, in <module>
    from magenta.models.score2perf import score2perf_hparams
  File "c:\users\togde\programmingprojects\magenta\magenta\models\score2perf\__init__.py", line 17, in <module>
    from magenta.models.score2perf import score2perf
  File "c:\users\togde\programmingprojects\magenta\magenta\models\score2perf\score2perf.py", line 30, in <module>
    from tensor2tensor.models import transformer
  File "C:\Users\togde\Anaconda3\envs\magenta\lib\site-packages\tensor2tensor\models\__init__.py", line 51, in <module>
    from tensor2tensor.models.research import rl
  File "C:\Users\togde\Anaconda3\envs\magenta\lib\site-packages\tensor2tensor\models\research\rl.py", line 27, in <module>
    from tensor2tensor.envs import tic_tac_toe_env
  File "C:\Users\togde\Anaconda3\envs\magenta\lib\site-packages\tensor2tensor\envs\__init__.py", line 23, in <module>
    from tensor2tensor.envs import tic_tac_toe_env
  File "C:\Users\togde\Anaconda3\envs\magenta\lib\site-packages\tensor2tensor\envs\tic_tac_toe_env.py", line 244, in <module>
    register()
  File "C:\Users\togde\Anaconda3\envs\magenta\lib\site-packages\tensor2tensor\envs\tic_tac_toe_env.py", line 240, in register
    "tensor2tensor.envs.tic_tac_toe_env:TicTacToeEnv", version="v0")
  File "C:\Users\togde\Anaconda3\envs\magenta\lib\site-packages\tensor2tensor\rl\gym_utils.py", line 360, in register_gym_env
    return env_name, gym.make(env_name)
  File "C:\Users\togde\Anaconda3\envs\magenta\lib\site-packages\gym\envs\registration.py", line 676, in make
    return registry.make(id, **kwargs)
  File "C:\Users\togde\Anaconda3\envs\magenta\lib\site-packages\gym\envs\registration.py", line 520, in make
    return spec.make(**kwargs)
  File "C:\Users\togde\Anaconda3\envs\magenta\lib\site-packages\gym\envs\registration.py", line 133, in make
    _kwargs = self.kwargs.copy()
AttributeError: 'NoneType' object has no attribute 'copy'

The tensor2tensor.rl library is automatically being imported at the start of the script, which calls the register_gym_env function in gym_utils.py with the optional kwargs argument not being supplied. Thus when gym is registering the environment, self.kwargs.copy() will throw an errors since kwargs is None.

This pull request changes the optional parameter into the more pythonic **kwargs argument that will be an empty dictionary if not supplied. This has been tested and does indeed fix the issue with magenta.

[bq-xiao]

(venv) D:\pyworkspace\tensor2tensor>python tensor2tensor\bin\t2t_trainer.py --registry_help
Traceback (most recent call last):
  File "tensor2tensor\bin\t2t_trainer.py", line 24, in <module>
    from tensor2tensor import models  # pylint: disable=unused-import
  File "D:\pyworkspace\tensor2tensor\tensor2tensor\models\__init__.py", line 51, in <module>
    from tensor2tensor.models.research import rl
  File "D:\pyworkspace\tensor2tensor\tensor2tensor\models\research\rl.py", line 27, in <module>
    from tensor2tensor.envs import tic_tac_toe_env
  File "D:\pyworkspace\tensor2tensor\tensor2tensor\envs\__init__.py", line 23, in <module>
    from tensor2tensor.envs import tic_tac_toe_env
  File "D:\pyworkspace\tensor2tensor\tensor2tensor\envs\tic_tac_toe_env.py", line 244, in <module>
    register()
  File "D:\pyworkspace\tensor2tensor\tensor2tensor\envs\tic_tac_toe_env.py", line 240, in register
    "tensor2tensor.envs.tic_tac_toe_env:TicTacToeEnv", version="v0")
  File "D:\pyworkspace\tensor2tensor\tensor2tensor\rl\gym_utils.py", line 360, in register_gym_env
    return env_name, gym.make(env_name)
  File "D:\pyworkspace\tensorflow-demo\venv\lib\site-packages\gym\envs\registration.py", line 610, in make
    _kwargs = spec_.kwargs.copy()
AttributeError: 'NoneType' object has no attribute 'copy'

The same error, when change the function parameter like this:

register_gym_env(class_entry_point, version="v0", **kwargs):

It works fine!!! 👍

[sohrab-crowdian]

I faced this issue as well.

@hawkinsp or someone else at Google, kindly consider accepting this merge request.

[George-Ogden]

fixes #1922

[brianroland]

Seems helpful. Any update?

[brianroland]

Nice @fengtony686 !

[Varun9198]

@fengtony686 can you please merge this pull request.

[khaledadrani]

The issue is still there? Can it be merged please?

[colmantse]

sorry has it been merged yet? facing the same issue.