[python-package] use positional args for np.arange in _InnerPredictor

[wagner-austin]

Summary

Change np.arange calls from keyword arguments to positional arguments in _InnerPredictor to fix mypy call-overload errors. Numpy's type stubs have overloads for positional args but not for (start=int, stop=int, step=int) keyword args.

Mypy Error (before fix)

python-package/lightgbm/basic.py:1320: error: No overload variant of "arange" matches argument types "int", "Any", "int"  [call-overload]
python-package/lightgbm/basic.py:1536: error: No overload variant of "arange" matches argument types "int", "int", "int"  [call-overload]

Related Issue

Fixes: #3867

Test Plan

• Ran pre-commit run mypy --files python-package/lightgbm/basic.py
• Verified mypy no longer reports call-overload errors for lines 1320 and 1536

[jameslamb]

Oh wow, thank you for this! I'm sad to see us lose keyword arguments here as they do add a bit of clarity, but it really does seem that's what numpy expects.

From help(np.arange) on numpy==2.2.5:

arange(...)
    arange([start,] stop[, step,], dtype=None, *, device=None, like=None)

    Return evenly spaced values within a given interval.

    ``arange`` can be called with a varying number of positional arguments:

    * ``arange(stop)``: Values are generated within the half-open interval
      ``[0, stop)`` (in other words, the interval including `start` but
      excluding `stop`).
    * ``arange(start, stop)``: Values are generated within the half-open
      interval ``[start, stop)``.
    * ``arange(start, stop, step)`` Values are generated within the half-open
      interval ``[start, stop)``, with spacing between values given by
      ``step``

And it even looks like these stubs were recently updated to now call that first argument start_or_stop: numpy/numpy#30147

It also looks like these were the only uses in lightgbm where np.arange() is called with keyword arguments. All others use positional arguments:

git grep -E 'np\.arange'

Thanks!