If you have several input images and use function mine_patches, the header will be added to csv file for each one. Check if file is empty or header is already there.

Coming from Spyros

When patch_size is defined in microns, the equivalent pixel conversion is done using the level of 0, and this does not get scaled for any other value. Thus, when a user selects level!=0, the final output is incorrect.

Reference: mlcommons/GaNDLF#459

When processing histology images, it would sense to be able to extract X x Y patches in terms of microns rather than pixels. This would require the following:

• Reading the magnification level(s) in the image
• Define X and Y from above in terms of M and N pixels
• Proceed extracting patches in terms of M x N pixels as usual

For example, if a WSI has MPP definition (these are defined as headers "tiffslide.mpp-x" and "tiffslide.mpp-y" from TiffSlide) as A_x along the x-axis and A_y along the y-axis, the patch size (in terms of pixels) that would need to get extracted would be: [M , N] ==[X / A_x , Y / A_y]

In cohesive downstream analyses 2 solutions can be done to ensure that models only process patches in the same shape:

• Resize extract patches of size [X / A_x , Y / A_y] to a single standard shape (such as [300, 300])
• Specify a single patch size

Is your feature request related to a problem? Please describe.
Currently, the only way to run OPM is to do a git clone and follow setup steps.

Describe the solution you'd like
It would make it much easier to work with if this is presented in PyPI.

Describe alternatives you've considered
Publish on conda is also a solution.

Additional context
A few examples:

• https://docs.microsoft.com/en-us/azure/devops/pipelines/artifacts/pypi?view=azure-devops&tabs=yaml
• https://stackoverflow.com/a/57957608/1228757
• https://docs.microsoft.com/en-us/azure/devops/pipelines/tasks/package/twine-authenticate?view=azure-devops#publish-python-distribution-to-azure-artifacts-feed
• https://github.com/FETS-AI/LabelFusion

There is a warning coming from opm.utils:

(/workspaces/OPM/venv) @sarthakpati ➜ /workspaces/OPM (pr/grenkoca/19 ✗) $ python patch_miner.py -h
/workspaces/OPM/opm/utils.py:4: FutureWarning: The `skimage.morphology.selem` module is deprecated and will be removed in scikit-image 1.0 (`skimage.morphology.selem` has been moved to `skimage.morphology.footprints`).
  from skimage.morphology.selem import disk

Put this check 2 places:

• OPM/opm/patch_manager.py

  Line 19 in 8abc8db

  self.slide_object = openslide.open_slide(filename)

• OPM/opm/patch_manager.py

  Line 46 in 8abc8db

  self.label_map_object = openslide.open_slide(path)

Coming from Spyros.