"TypeError: argument of type 'LazyCompletionHelp' is not a container or iterable" when attempting to run "spktrain experiment=qm9_atomwise"

[JEAM-professional]

I am following the simple examples in the Installation page (https://schnetpack.readthedocs.io/en/latest/getstarted.html) to get acquainted with schnetpack. When attempting to follow the first one, "Training your first models -> Example1: QM9", the following command

spktrain experiment=qm9_atomwise

produces the following error log for me:

Traceback (most recent call last):
  File "/opt/homebrew/Cellar/python@3.14/3.14.5/Frameworks/Python.framework/Versions/3.14/lib/python3.14/argparse.py", line 1748, in _check_help
    formatter._expand_help(action)
    ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^
  File "/opt/homebrew/Cellar/python@3.14/3.14.5/Frameworks/Python.framework/Versions/3.14/lib/python3.14/argparse.py", line 665, in _expand_help
    if '%' not in help_string:
       ^^^^^^^^^^^^^^^^^^^^^^
TypeError: argument of type 'LazyCompletionHelp' is not a container or iterable

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/.../.venv/bin/spktrain", line 5, in <module>
    cli.train()
    ~~~~~~~~~^^
  File "/.../.venv/lib/python3.14/site-packages/hydra/main.py", line 85, in decorated_main
    args_parser = get_args_parser()
  File "/Users/jearias/.venv/lib/python3.14/site-packages/hydra/_internal/utils.py", line 547, in get_args_parser
    parser.add_argument(
    ~~~~~~~~~~~~~~~~~~~^
        "--shell-completion",
        ^^^^^^^^^^^^^^^^^^^^^
    ...<2 lines>...
        help=LazyCompletionHelp(),  # type: ignore
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    )
    ^
  File "/opt/homebrew/Cellar/python@3.14/3.14.5/Frameworks/Python.framework/Versions/3.14/lib/python3.14/argparse.py", line 1562, in add_argument
    self._check_help(action)
    ~~~~~~~~~~~~~~~~^^^^^^^^
  File "/opt/homebrew/Cellar/python@3.14/3.14.5/Frameworks/Python.framework/Versions/3.14/lib/python3.14/argparse.py", line 1750, in _check_help
    raise ValueError('badly formed help string') from exc
ValueError: badly formed help string

It gets past this if I switch to another virtual environment running on python3.12 instead of one running on python3.14.

[stefaanhessmann]

Hi @JEAM-professional,
thanks for letting us know! It seems like hydra is not yet adapted to the newest python version. For now, I would recommend to just use the 3.12 version. We will monitor this and fix it, if it remains an issue.