update dtype aliases in imageio

[rdenham]

For numpy 1.24.0, the deprecation for the aliases np.object, np.bool, np.float, np.complex, np.str, and np.int expired (introduces NumPy 1.20).

This means that in recent versions of numpy you'll get errors like:

AttributeError: module 'numpy' has no attribute 'bool'

Perhaps the authors could update the mapping in imageio.py to look like:

dataTypeMapping = [
    (numpy.uint8, gdalconst.GDT_Byte),
    (numpy.bool_, gdalconst.GDT_Byte),
    (numpy.int16, gdalconst.GDT_Int16),
    (numpy.uint16, gdalconst.GDT_UInt16),
    (numpy.int32, gdalconst.GDT_Int32),
    (numpy.uint32, gdalconst.GDT_UInt32),
    (numpy.single, gdalconst.GDT_Float32),
    (numpy.float_, gdalconst.GDT_Float64)
]

So that the original alias numpy.bool becomes numpy.bool_. Same for numpy.float. I think bool and float would also work.

Probably there are other ways too, but this approach seems to be the one with minimal effort and should not change the behaviour of the existing code.

[neilflood]

Thanks @rdenham
Good point. Yes, we will update with something like that. Stay tuned......

[neilflood]

@rdenham
There is now a new release of rios-1.4.14, which fixes this problem. Thank you for letting us know.

Just FYI, python-fmask had similar problems, so there is now a new version 0.5.8 of that, too.

[rdenham]

thanks. Closed by #52

[gillins]

Also a tuiview 1.2.11 with a fix FWIW.