ExtractorProcessor.__call__() does not retain the order given by the members filter

[carlocastoldi]

Description of the problem:
When I use an ExtractorProcessor (e.g. Unzip), I would expect that if I specify a list of files i want to extract, the order of the extracted paths was retained.
Instead, ExtractorProcessor uses os.walk(self.extract_dir)

pooch/pooch/processors.py

Line 126 in 5860444

for path, _, files in os.walk(self.extract_dir):

and then filters the extracted files based on the given members, resulting in an arbitrary order.

I would like to suggest to invert the check: loop over the members (if any), and filter on the extracted files.

Also: why not use pathlib.Path?

Example

>>> pooch.Unzip(
                extract_dir=".",
                members=[
                        'E11.5/E11-5_DevCCF_Annotations_31.5um.nii.gz',
                        'E11.5/E11-5_MRI-adc_31.5um.nii.gz',
                ])

Expected output

'/home/castoldi/brainglobe_workingdir/kim_dev_mouse/downloads/./E11.5/E11-5_DevCCF_Annotations_31.5um.nii.gz']
'/home/castoldi/brainglobe_workingdir/kim_dev_mouse/downloads/./E11.5/E11-5_MRI-adc_31.5um.nii.gz',

Actual output

'/home/castoldi/brainglobe_workingdir/kim_dev_mouse/downloads/./E11.5/E11-5_MRI-adc_31.5um.nii.gz',
'/home/castoldi/brainglobe_workingdir/kim_dev_mouse/downloads/./E11.5/E11-5_DevCCF_Annotations_31.5um.nii.gz']

System information

• Operating system: Ubuntu 24.04
• Python installation (Anaconda, system, ETS): system
• Version of Python: 3.12.3
• Version of this package: 1.8.2