Draft: Add Fast Sketch Cleanup plugin

Note: it requires Pillow and openvino modules available in the system or the dependencies to work.

Still to do:

• Make sure the training runs in a different thread/process (can block the user, just be responsive during that)
  • That should fix the issue of it not running the first time (trying to make it made the code a mess)
• Cut the image in parts, run them all through the model and merge it to return to the user, so the static models will work on all images
• Find out the model size in the plugin and act accordingly
• Add GPU as option
• Only show the devices that are actually available for the user - I mean grey out the unavailable ones
• Add invert as option?
• Add a yaml file with stuff like whether it needs inverting, and minimum divisions and context
• Eventually, include the model in the plugin itself
• Ask users: should it work differently (right now it takes the canvas as input, and adds another layer on top as output)
• Clean code
• T.b.d.

Test Plan

Prerequisites:

1. Build Krita, or download the plugin.
2. Have both required packages: Pillow and openvino (in imports as: PIL and openvino):
   • either available in your system: pip3 install openvino Pillow
   • or build a version of Krita that contains them, using dependencies from this MR: dkazakov/krita-deps-management!40 (merged)
3. Download a model, for example from here (TSC model): https://drive.google.com/file/d/1hensU0_zjwTZTnKo9M0qb_Jfbukl68eH/view?usp=drive_link or here (exp 06 model): https://drive.google.com/file/d/1IrxErCbrvHJMPIaPoSw2NHbq40RnHZSR/view?usp=drive_link
4. Use an input image, for example this one: fish_scaled

Test:

1. Go to Tools -> Scripts -> Fast Sketch Cleanup.
2. Select CPU (most likely) as backend
3. Click on the Run button (currently you might need to do it twice)
4. Wait
5. Look at the effect.

Formalities Checklist

• I confirmed this builds.
• I confirmed Krita ran and the relevant functions work.
• I tested the relevant unit tests and can confirm they are not broken. (If not possible, don't hesitate to ask for help!)
• I made sure my commits build individually and have good descriptions as per KDE guidelines.
• I made sure my code conforms to the standards set in the HACKING file.
• I can confirm the code is licensed and attributed appropriately, and that unattributed code is mine, as per KDE Licensing Policy.
• Does the patch add a user-visible feature? If yes, is there a documentation MR ready for it at Krita Documentation Repository?

Reminder: the reviewer is responsible for merging the patch, this is to ensure at the least two people can build the patch. In case a patch breaks the build, both the author and the reviewer should be contacted to fix the build. If this is not possible, the commits shall be reverted, and a notification with the reasoning and any relevant logs shall be sent to the mailing list, kimageshop@kde.org.