imread() fails on HTTPS, code example from documentation

[MartinMeo]

Hey folks,

I am brand new with imageio. It looks like np.stack() is throwing
ValueError('all input arrays must have the same shape')
when reading a .gif file via HTTPS from a URL. I tried two different image files with the same result.

The code is directly from documentation examples so why doesn't everyone get this error on day one.
examples reading-from-fancy-sources

I'd appreciate any ideas or insights into the issue.

import imageio.v3 as iio
import io

# from HTTPS
web_image = "https://upload.wikimedia.org/wikipedia/commons/d/d3/Newtons_cradle_animation_book_2.gif"
# web_image = "https://media.giphy.com/media/BPJmthQ3YRwD6QqcVD/giphy.gif"
frames = iio.imread(web_image, index=None)

Traceback (most recent call last):
File "/Users/martinmeo/SW_DEV/PYTHON/2023/IMAGE-IO-LIB/demo.py", line 7, in
frames = iio.imread(web_image, index=None)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/martinmeo/miniconda3/envs/PATH311/lib/python3.11/site-packages/imageio/v3.py", line 54, in imread
return np.asarray(img_file.read(**call_kwargs))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/martinmeo/miniconda3/envs/PATH311/lib/python3.11/site-packages/imageio/plugins/pillow.py", line 234, in read
image = np.stack([im for im in iterator], axis=0)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "<array_function internals>", line 200, in stack
File "/Users/martinmeo/miniconda3/envs/PATH311/lib/python3.11/site-packages/numpy/core/shape_base.py", line 464, in stack
raise ValueError('all input arrays must have the same shape')
ValueError: all input arrays must have the same shape

HW: MacBook Pro M2
OS: macOS Ventura 13.4.1 (c)
conda 23.3.1
python: 3.11.3
imageio: 2.31.1

[Pandede]

Try to set mode to RGB or RGBA

frames = iio.imread(web_image, index=None, mode='RGB')

The GIF cannot be read properly as the first frame has a different dimension to the others, therefore numpy unable to stack the frames with different shape:

from collections import Counter

from PIL import Image, ImageSequence

web_image = "Newtons_cradle_animation_book_2.gif"
gif_image = Image.open(web_image)

mode_counter = Counter(
    frame.mode
    for frame in ImageSequence.Iterator(gif_image)
)
print(mode_counter)  # Counter({'RGBA': 35, 'P': 1})

[Pandede]

This bug is also mentioned in PR of Pillow: python-pillow/Pillow#6150

[FirefoxMetzger]

Fixed by #1036