Converting Pillow .tiff image to numpy uint8 produces TypeError on Windows 10

[mvanlierBCG]

What did you do?

Loading a .tiff image:

img = Image.open(img_name)

Convert the loaded image to a NumPy array:

image_array = np.array(img, dtype='uint8')

What did you expect to happen?

That I am able to covert the loaded image to a NumPy array.
In OS: Ubuntu 18.04.5 the environment, same version names don't throw an error and it works.

What actually happened?

TypeError:
int() argument must be a string, a bytes-like object or a number, not 'TiffImageFile'

What are your OS, Python and Pillow versions?

• OS: Windows 10
• Python: 3.7
• Pillow: 7.0.0

from PIL import Image
import numpy as np
img = Image.open(img_name)
image_array = np.array(img, dtype='uint8')

Tiff image
test_tiff.zip

[cgohlke]

Cannot reproduce:

Python 3.7.6 (tags/v3.7.6:43364a7ae0, Dec 19 2019, 00:42:30) [MSC v.1916 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> from PIL import Image
>>> import numpy as np
>>> img = Image.open('test_tiff.tiff')
>>> image_array = np.array(img, dtype='uint8')
>>> image_array.shape
(1028, 2448, 3)
>>> np.__version__
'1.16.6'

[kmilos]

This is a compressed TIFF file, and I suspect, the same libtiff build problem as #4237

But Pillow should really handle this better, i.e. I don't think #3863 is fully closed.

[radarhere]

Could we get a complete traceback for the error?

Could you check if your Windows and Ubuntu systems are using different versions of libtiff? That could also be helpful.

>>> from PIL import Image
>>> Image.core.libtiff_version
'LIBTIFF, Version 4.1.0\nCopyright (c) 1988-1996 Sam Leffler\nCopyright (c) 1991-1996 Silicon Graphics, Inc.'
>>> 

[radarhere]

I'm not able to reproduce this on our GitHub Actions or AppVeyor jobs, so closing due to lack of response.

[zhangsan5213]

I just got the same thing on conda 4.9.2 python 3.7

I'm using also python 3.7 on my laptop and it worked fine. My PC was not good and threw me the error above.