`DeprecationWarning` for assigning to `numpy.ndarray.shape`

[keewis]

(caught by xarray's upstream-dev CI)

numpy is deprecating assignment to the shape property of numpy.ndarray instances in numpy>=2.5.0, which netCDF4 appears to be doing. To reproduce:

import netCDF4
import numpy as np
import warnings

warnings.filterwarnings("error")  # raise on all warnings

with netCDF4.Dataset("test.nc", mode="w") as ds:
    x = ds.createDimension("x", 3)
    y = ds.createDimension("y", 4)
    var = ds.createVariable("a", "f4", ("y", "x"), fill_value=np.nan)
    var.set_always_mask(False)
    var.set_auto_mask(False)
    
    new_data = np.array([[ 1.76405235,  0.40015721,  0.97873798],
       [ 2.2408932 ,         np.nan, -0.97727788],
       [ 0.95008842, -0.15135721, -0.10321885],
       [ 0.4105985 ,  0.14404357,  1.45427351]], dtype="f4")
    var[...] = new_data

(not sure if both set_always_mask and set_auto_mask is necessary, but the warning is raised in the assignment of the data)

Result:

Traceback (most recent call last):
  File "<python-input-6>", line 18, in <module>
    var[...] = new_data
    ~~~^^^^^
  File "src/netCDF4/_netCDF4.pyx", line 5633, in netCDF4._netCDF4.Variable.__setitem__
DeprecationWarning: Setting the shape on a NumPy array has been deprecated in NumPy 2.5.
As an alternative, you can create a new view using np.reshape (with copy=False if needed).

Edit: the version of netCDF4 is 1.7.4