Repair bad pixels doesn't work if "cv2" isn't installed

[bmilde]

I'm getting this without having opencv-python installed:

Traceback (most recent call last): File "convert_images3.py", line 22, in convert_image bad_pixels = rawpy.enhance.find_bad_pixels([image_path1, image_path2, ...]) File "/usr/local/lib/python3.10/dist-packages/rawpy/enhance.py", line 80, in find_bad_pixels coords.extend(_find_bad_pixel_candidates(raw, isCandidate)) File "/usr/local/lib/python3.10/dist-packages/rawpy/enhance.py", line 112, in _find_bad_pixel_candidates coords = _find_bad_pixel_candidates_bayer2x2(raw, isCandidateFn) File "/usr/local/lib/python3.10/dist-packages/rawpy/enhance.py", line 178, in _find_bad_pixel_candidatesbayer2x2 med = median(rawslice) TypeError: median() got an unexpected keyword argument 'footprint'

The error is resolved when https://pypi.org/project/opencv-python/ is installed (pip install opencv-python), but if the code is supposed to run without it as well, then this should be fixed.

[letmaik]

The change to footprint was done recently in https://github.com/letmaik/rawpy/pull/201. Which version of scikit-image are you using?

[bmilde]

I have 0.18.3

pip3 show scikit-image

Name: scikit-image Version: 0.18.3 Summary: Image processing in Python Home-page: https://scikit-image.org Author: Author-email: License: Modified BSD Location: /usr/lib/python3/dist-packages Requires: Required-by:

[letmaik]

The footprint keyword was introduced in scikit-image 0.19 (released Dec 2021) where selem got deprecated and finally removed in 0.20. You will have to either use the previous rawpy version or update your scitkit-image version to 0.19 or higher.