Fix VAEEncodeForInpaint to support WAN VAE tuple downscale_ratio

[ChrisFab16]

• Handle WAN VAE downscale_ratio which is a tuple (function, min, max) instead of integer
• Fixes TypeError when using VAEEncodeForInpaint with Qwen Image/WAN VAEs
• Backward compatible with standard VAEs that use integer downscale_ratio

[rattus128]

are these actually min/max?

It looks like x and y. This code looks like it might be solving the same problem:

    def spacial_compression_decode(self):
        try:
            return self.upscale_ratio[-1]
        except:
            return self.upscale_ratio

    def spacial_compression_encode(self):
        try:
            return self.downscale_ratio[-1]
        except:
            return self.downscale_ratio

Can you just call vae.spacial_compression_encode():

        x = (pixels.shape[1] // vae.spacial_compression_encode()) * vae.spacial_compression_encode()
        y = (pixels.shape[2] // vae.spacial_compression_encode()) * vae.spacial_compression_encode()

[ChrisFab16]

Thanks for the feedback, @rattus128!
I updated VAEEncodeForInpaint to use vae.spacial_compression_encode() for the downscale ratio and offsets, matching the existing VAE helper used elsewhere. This removes the manual tuple handling and works for both WAN-style tuples and standard ints. Tested locally with mock VAEs; both tuple and int downscale cases behave as expected.

[rattus128]

If you check the diff here, I think your IDE or git or something has done a full whitespace reformat of the file. You will need to reduce the diff back to just the content changes with whitespace consistency to surrounds.

[comfy-pr-bot]

Test Evidence Check

⚠️ Warning: Test Explanation Missing

If this PR modifies behavior that requires testing, a test explanation is required. PRs lacking applicable test explanations may not be reviewed until added. Please add test explanations to ensure code quality and prevent regressions.

⚠️ Warning: Visual Documentation Missing

If this PR changes user-facing behavior, visual proof (screen recording or screenshot) is required. PRs without applicable visual documentation may not be reviewed until provided.

You can add it by:

• GitHub: Drag & drop media directly into the PR description
• YouTube: Include a link to a short demo