Weights are not properly applied for masked and clipped arrays in `combine()`

[monodera]

I'm trying to combine images with ccdproc.combine().

The following is an example to show the issue. Note that I set the mask = [[1, 0, 0], [0, 0, 0]] for the first image.

import numpy as np
from astropy.nddata import CCDData
from ccdproc import combine

ccd1 = CCDData([[1, 1, 1], [2, 2, 2]], mask=[[1, 0, 0], [0, 0, 0]], unit="adu")
ccd2 = CCDData([[1, 1, 1], [2, 2, 2]], unit="adu")

weights = np.array([1, 1])
weights = None

combined_ccd = combine(
    [ccd1, ccd2],
    method="average",
    weights=weights,
)

print(combined_ccd.data)

With weights = None, the result is [[1., 1., 1.], [2., 2., 2.]] as expected.

On the other hand, with weights = np.array([1, 1]), the result becomes [[0.5 1. 1. ], [2., 2., 2.]], while I still expect the result is [[1, 1, 1], [2, 2, 2]].

I think that the combine() does not remove the masked elements from the weights array, so smaller value are returned.

A similar behavior is also seen in the clipped arrays.

My Python environment is the following:

• Python 3.11.12
• ccdproc 2.4.3
• Numpy 2.2.5
• Astropy 7.0.2