Add fast NL means denoiser node

[adegerard]

This PR adds a Non-local Means Denoising algorithm.

Note: the "Color strength" input is only used when the input image is RGB/RGBA: is it possible to hide this field when the connected image is a grayscale image?

[joeyballentine]

is it possible to hide this field when the connected image is a grayscale image?

Yes actually. You would use an "if_group" with a type condition that matches RGB/RGBA, similar to this:

chaiNNer/backend/src/packages/chaiNNer_pytorch/pytorch/restoration/upscale_face.py

Line 125 in 880441d

if_group(Condition.type(0, 'PyTorchModel { arch: "CodeFormer" }'))(

[adegerard]

Yes actually. You would use an "if_group" with a type condition that matches RGB/RGBA, similar to this:

OK. I'll modify my code.

[adegerard]

Done.

[RunDevelopment]

Formatting on tests currently fail. To fix the formatting, run the command npm run lint-fix - after installing our dev dependencies of course (npm ci and pip install -r requirements.txt). The tests fail because you added a new test case (the .chn file) but didn't update the test snapshots. Run npm test -- -u to update them.

[adegerard]

Formatting on tests currently fail. To fix the formatting, run the command npm run lint-fix - after installing our dev dependencies of course (npm ci and pip install -r requirements.txt). The tests fail because you added a new test case (the .chn file) but didn't update the test snapshots. Run npm test -- -u to update them.

I didn't know. I am a beginner. Would it be possible to document the different steps before creating a pull request ? Thanks.

[RunDevelopment]

Would it be possible to document the different steps before creating a pull request ?

Good point. I'll add it to our contributing docs.

[RunDevelopment]

Thank you @adegerard!