add avif parsing function

[audioscavenger]

add avif parsing function. added as much comments as I could. more details here

[comfyanonymous]

do you have a test file for this or a way to create one?

[audioscavenger]

You are right, sorry I should have included a test rig.

• Attached is a basic json and output files, and also the *-2AVIF *-2WEBP files converted from png with imagick + exiftool.
• I have made the getAvifMetadata and getWebpMetadata more resilient to errors by adding a simple test if the data[key] is undifined.
• I have left the comments and removed all the debug console.log AVIF-validation.zip If you don't want the outputs, here is just the json for testing: AVIF-validation.json

I have also completely updated the post that explains all this at avif parsing explanations

All seems to work perfectly, i tested different workflows with various prompt sises, and converted lots of png to avif and webp, all load perfectly!

[audioscavenger]

hi, I see that the number of PR is piling up... sorry that I cannot help you more. Is there anyone that can back you up for testing PR's? I still use this code myself and haven't found a bug since I submitted the last update. Please let me know if you need more details on how it reads the AVIF. I basically implemented what you did with Webp, but with different chunk sizes and dynamic position detections, based off a meticulous ascii analysis of different AVIF files + the RFC I could find.

[mcmonkey4eva]

We're gonna be working to improve PR handling here very soon

[mcmonkey4eva]

Could you please rebuild the PR? It has a merge conflict caused by the recent addition of audio file support.

[mcmonkey4eva]

frontend is replaced by https://github.com/Comfy-Org/ComfyUI_frontend now