Feature request: save image with metadata
Closed this issue · 3 comments
Hi, thanks for making this ComfyUi addon.
As you mentioned yourself, ComfyUi's metadata is quite a mess. I wasn't able to reconstruct the prompt and settings from an image I made earlier. It seems reconstructing the complete workflow is the only option.
With that being said: would it be possible to add a node to this addon which saves an image with metadata in the selected format? That way, I can first use ComfyUi to reconstruct the image workflow, then add the SaveImageWithMetadata node and select the metadata format. I then have an image I should be able to load using LoadImageWithMetadata.
My goal eventually is to have a folder containing images with just the following metadata:
- positive prompt
- negative prompt
- checkpoint name
This will be enough for me to run an upscaler script for each of the image files in the folder.
Sounds reasonable to have this step added to the initial workflow, as you might have a clear view of what goes into the prompts and we don't need to reconstruct this later by going through the node tree.
I saw this a few days back on reddit, maybe this is already what you need:
https://www.reddit.com/r/comfyui/comments/15pz7ak/node_to_save_image_and_option_to_save_text_prompt/
https://www.reddit.com/r/comfyui/comments/15tjz0b/biggest_let_down_of_comfyui_you_loose_generation/
The later seems to be more 'professional' but i did not have a close look at any of these. This one seems to save in A1111 format which should automatically work with my node.
Github link: https://github.com/giriss/comfy-image-saver
Seems to have a minor compatibility issue at the moment. This causes CFG scale to not be readable by my node.
For anyone looking for a all in one solution (save and load nodes) made by the original author of the prompt reader, please use this node collection here:
https://github.com/receyuki/comfyui-prompt-reader-node
I guess this one can be loaded via comfy manager as well.