CivitAI serves a re-encoded, metadata-free optimized rendition as the recipe
preview, so the ComfyUI workflow embedded in the original image was dropped:
imported recipes reported has_workflow=false and never offered "Send Workflow
to ComfyUI" even when the source image carried one.
Recover the workflow from the original rendition and carry it to the save step
as data, so the stored preview stays the small optimized image:
- ExifUtils: embed a caller-supplied workflow during optimize_image's single
encode pass, and add embed_workflow() to patch WebP EXIF in place (used by
the verbatim skip_optimize branch and as a safety net).
- RecipePersistenceService.save_recipe: embed metadata["workflow"] before
detecting has_workflow.
- analyze_remote_image: return the workflow recovered from the original
rendition it already downloads for EXIF parsing.
- RecipeManagementHandler: add _fetch_original_media() and workflow helpers;
_do_import_from_url reuses them, and _do_import_remote_recipe fetches the
original only when CivitAI reports a ComfyUI payload (meta.comfy) so
workflow-less images pay no extra bandwidth.
- Batch URL imports and the import modal forward the recovered workflow.
Verified against the reported image: has_workflow flips from false to true and
the recovered workflow matches the original (25 nodes, same graph id).
Add Load Image Metadata (LoraManager) to extract reusable prompts,
model references, LoRA stacks, and sampling settings from images.
Prefer saved A1111-style parameters by default, with optional workflow
and subgraph sampler selection. Resolve local model and LoRA names,
report missing resources, and recover extraction failures with explicit
defaults and readable diagnostics.
Include parser, resource-resolution, and node regression tests, plus
usage documentation.
Recipes imported by drag & drop / file-picker record no source_path and
were rejected by re-import. Fall back to the recipe's own saved image,
which still carries the original embedded generation metadata.
Re-import now re-parses that original metadata instead of the appended
recipe JSON block, so parser upgrades produce fresh results. The
already-optimized preview image is kept verbatim: only its WebP EXIF
chunk is rewritten in place to replace the recipe metadata block, and
the recipe JSON is rewritten with the new analysis plus carried-over
user edits.
Add support for storing checkpoint information in image EXIF metadata. The checkpoint data is simplified and includes fields like model ID, version, name, hash, and base model. This allows for better tracking of AI model checkpoints used in image generation workflows.
- Removed the extract_user_comment and update_user_comment methods to streamline the ExifUtils class.
- Cleaned up unnecessary imports and reduced code complexity, focusing on essential functionality for image metadata extraction.
- Replaced the download function for images from Twitter to Civitai in recipe_routes.py.
- Updated metadata extraction from user comments to a more comprehensive image metadata extraction method in ExifUtils.
- Enhanced the appending of recipe metadata to utilize the new metadata extraction method.
- Added a new utility function to download images from Civitai.
- Updated the recipe saving logic to accept a workflow JSON input instead of individual fields like name, tags, and metadata.
- Implemented parsing of the workflow to extract generation parameters and Lora stack, improving the recipe creation process.
- Enhanced error handling for missing workflow data and invalid Lora formats.
- Removed deprecated code related to individual field handling, streamlining the recipe saving functionality.
- Updated the front-end widget to send the workflow JSON directly, simplifying the data preparation process.
- Updated the logic in ExifUtils to first identify KSampler nodes and store references to positive and negative prompt nodes.
- Added a helper function to follow these references and extract text content from CLIP Text Encode nodes.
- Implemented a fallback mechanism to extract prompts using traditional methods if references are not available.
- Improved code readability with additional comments and structured handling of node data.
- Introduced a new API endpoint to save recipes directly from the LoRAs widget.
- Implemented logic to handle recipe data, including image processing and metadata extraction.
- Enhanced error handling for missing fields and image retrieval.
- Updated the ExifUtils to extract generation parameters from images for recipe creation.
- Added a direct save option in the widget, improving user experience.
- Moved the recipe metadata appending logic to occur after the JSON creation for better workflow.
- Adjusted the user comment formatting in ExifUtils to include a newline for improved readability.
- Reduced the maximum height of the recipe modal bottom section for better layout consistency.
- Removed print statements for initialization and setup in RecipeRoutes to reduce console clutter and improve logging practices.
- Updated image optimization parameters in RecipeRoutes to enhance image quality by increasing the target width.
- Modified user comment handling in ExifUtils to ensure proper formatting when appending recipe metadata, improving metadata consistency.
- Implemented image optimization in RecipeRoutes, resizing and converting uploaded images to WebP format while preserving metadata.
- Updated ExifUtils to support EXIF data handling for WebP images, ensuring compatibility with various image formats.
- Added a new method for optimizing images, allowing for better performance and quality in image uploads.
- Updated ExifUtils to handle both JPEG/TIFF and non-JPEG/TIFF images for extracting UserComment from EXIF data, improving compatibility with various image formats.
- Introduced A1111MetadataParser to support parsing of images with A1111 metadata format, extracting prompts, negative prompts, and LoRA information.
- Enhanced error handling and logging for metadata parsing processes, ensuring better traceability and debugging capabilities.
- Introduced a new RecipeParserFactory to streamline the parsing of recipe metadata from user comments, supporting multiple formats.
- Removed legacy metadata extraction logic from RecipeRoutes, delegating responsibilities to the new parser classes.
- Enhanced error handling for cases where no valid parser is found, ensuring graceful responses.
- Updated the RecipeScanner to improve the handling of LoRA metadata and reduce logging verbosity for better performance.
- Added functionality to extract and process existing recipe metadata from images, including LoRA details and Civitai information.
- Updated ExifUtils to manage recipe metadata more effectively, including appending and removing metadata from user comments.
- Improved the ImportManager to utilize recipe metadata for setting default recipe names and tags when importing shared recipes.
- Introduced new endpoints for sharing recipes and downloading shared images in RecipeRoutes.
- Implemented logic to process recipe images and append metadata to EXIF data.
- Updated RecipeCard component to handle sharing via API calls, providing user feedback during the process.
- Enhanced error handling for missing recipe IDs and failed API responses.
Will Miao