Feature Request: Restart option for the feature extraction pipeline

[GeorgeBatch]

• TIA Toolbox version: develop branch
• Python version: 3.11
• Operating System: linux

Feature Request

I am computing features from a bunch of Whole Slide Images using code like in https://github.com/TissueImageAnalytics/tiatoolbox/blob/develop/examples/11-import-foundation-models.ipynb

I think a restart option would be very beneficial when computing features for many slides.

Currently, the save directory is checked. If it already exists, the process terminates with an Error. This is good to prevent overwriting, but restarting would be also good to have, if asked for. To implement the restart option, I think the overwriting prevention would need to be handled on slide level instead of checking the existence of a directory, which means that instead of numbered features, real names would probably need to be used when saving features and positions npy files.

[shaneahmed]

The new engine design #578 allows you to overwrite existing results. We will update the design to skip the WSIs which have been processed already.

[GeorgeBatch]

@shaneahmed, great! One option to overwrite, and one to resume would be great to have!