Accept NumPy scalars in integer and string properties

[RaphaelFakhri]

Link to issue

Closes #5827

Description of change

Integer properties now accept NumPy integer scalars, and non-strict string properties now accept any number, including NumPy scalars. IntegerValidator checks numbers.Integral instead of int, and StringValidator checks numbers.Number instead of (int, float), which matches how NumberValidator already validates numbers.

Demo

import numpy as np
import pandas as pd
import plotly.graph_objects as go

df = pd.DataFrame({"year": [2020, 2021], "v": [1, 2]})
fig = go.Figure([go.Scatter(y=[1], name=y) for y in df.year.unique()])
fig.update_layout(xaxis_nticks=np.int64(5))
print([t.name for t in fig.data], fig.layout.xaxis.nticks)

Before: ValueError: Invalid value of type 'numpy.int64' received for the 'name' property of scatter

After: ['2020', '2021'] 5. The figure serializes the same with the json and orjson engines ("name":"2020", "nticks":5).

Testing strategy

The existing integer and string validator tests now include NumPy scalars: scalar acceptance, acceptance within min/max bounds, rejection outside min/max bounds, a list of NumPy integers for array_ok properties, and NumPy integer and float32 conversion to strings. Out-of-range NumPy integers are still rejected.

Additional information (optional)

Strings, booleans, NaN and infinite values are validated as before: numbers.Integral excludes floats, and the string check still runs before the number check.

Guidelines

• I have reviewed the pull request guidelines and the Code of Conduct and confirm that this PR follows them.
• I have added an entry to the changelog if needed (not required for documentation PRs).

[camdecoster]

Thanks for the update. I'll review and follow up.