cumsum broken with NaNs

[dcherian]

import numpy_groupies as npg
from numpy import array, nan


a = array([[5.0, 6.0, 7.0, 8.0, 9.0], [0.0, 0.0, 0.0, nan, 0.0]])
#mask = np.isnan(a)
#a[mask] = 0

result = npg.aggregate_numpy.aggregate(
    array([1, 1, 1, 0, 0]),
    a,
    func="cumsum",
    axis=-1,
)
#result[mask] = np.nan
result

array([[ 5., 11., 18.,  8., 17.],
       [nan, nan, nan, nan, nan]])

We just need to do the nan replacement as with nancumsum