fix to_scalar_or_list when v.ndim == 0

[ankokumoyashi]

• test code

import plotly.graph_objs as go
import numpy as np
go.Scatter(
    y=[np.array(1), np.array(2)],
    x=[1, 1]
)

• before fix

sh-4.2# python test.py
Traceback (most recent call last):
  File "test.py", line 5, in <module>
    x=[1, 1]
  File "/root/review/plotly.py/plotly/graph_objs/_scatter.py", line 2567, in __init__
    self['y'] = y if y is not None else _v
  File "/root/review/plotly.py/plotly/basedatatypes.py", line 2841, in __setitem__
    self._set_prop(prop, value)
  File "/root/review/plotly.py/plotly/basedatatypes.py", line 3077, in _set_prop
    val = validator.validate_coerce(val)
  File "/root/review/plotly.py/_plotly_utils/basevalidators.py", line 382, in validate_coerce
    v = to_scalar_or_list(v)
  File "/root/review/plotly.py/_plotly_utils/basevalidators.py", line 56, in to_scalar_or_list
    return [to_scalar_or_list(e) for e in v]
  File "/root/review/plotly.py/_plotly_utils/basevalidators.py", line 56, in <listcomp>
    return [to_scalar_or_list(e) for e in v]
  File "/root/review/plotly.py/_plotly_utils/basevalidators.py", line 58, in to_scalar_or_list
    return [to_scalar_or_list(e) for e in v]
TypeError: iteration over a 0-d array

• after fix

sh-4.2# python test.py
sh-4.2#

[jonmmease]

Thanks for the fix @ankokumoyashi! We should add a small test for this to protect against future regressions. I think an additional test in _plotly_utils/tests/validators/test_dataarray_validator.py would be a good place for it. Basically to test that [np.array(1), np.array(2)] is coerced into [1, 2].

I can add it in a few days if you don't have time, thanks again for the contribution!

[ankokumoyashi]

I added test. Thank you for your kind instruction!

[jonmmease]

Great! Thanks for the contribution @ankokumoyashi!