- Add append: to autotag skip-tagged filter alongside prepend:
- Update tag chip styling to dark fill with white text
- Add v3.1.0 changelog entry covering all changes since v3.0.26
Add two WD14 classifier models as alternatives to JoyCaption LLM taggers.
WD14 models are ~400MB ONNX classifiers that output Danbooru tags with
confidence scores — much faster inference, no prompt needed.
Backend:
- Add wd14-swinv2 and wd14-vit entries to MODELS dict
- Add check/download/load/generate methods for ONNX-based WD14 inference
- Add general_threshold and character_threshold params to generate_tags()
- Filter "prepend:*" tags from exclude prefixes and skip-tagged logic
Frontend:
- Add WD14 radio buttons with download status in both admin and gallery
- Toggle prompt textarea vs threshold sliders based on model selection
- Pass thresholds to API for batch and single-image tagging
- Extract _setModelStatus() helper using safe DOM methods
Dependencies: onnxruntime (optional, with clear error message)
- Remove unused `patch` import from test_output_dir.py
- Add debug log message to ImportError handler instead of bare pass
- Invalidate gallery file listing cache alongside output dir cache
when the user changes gallery_root_path in settings
_find_comfyui_output_dir() now checks GalleryConfig.MONITORING_DIRECTORIES
first before falling back to filesystem auto-detection. Also invalidates
the cached directory when the user changes the setting via save_settings().
Closes#108
- Move utils imports to module-level with try/except for relative
(ComfyUI) and absolute (test runner) import contexts in prompts.py.
Fixes 500 errors on rating update, prompt save, and prompt edit.
- Add missing None for rating parameter in image scan's save_prompt
call, which was passing the notes string as rating causing a type
comparison error during image scanning.
- Black (auto-fix formatting on commit)
- Flake8 (block on syntax errors and undefined names)
- Bandit (security scan, non-blocking to match CI)
- Tests (run on pre-push only)
- Apply Black formatting fixes caught by the new hooks
- Fix XSS in showFullPrompt() across admin.js, gallery.js, metadata.html
by replacing HTML string interpolation with DOM manipulation (createElement
+ textContent), eliminating unsafe user content injection
- Fix stack trace exposure in py/api error responses by replacing str(e)
with generic messages and adding logger.exception() for server-side
traceability
- Fix AttributeError: 'str' has no attribute 'as_posix' on thumbnail_rel_path
in get_output_images and find_duplicates (thumbnail_rel_path is already a
string from f-string construction, not a Path object)
- Fix NameError: total_images undefined in generate_thumbnails (was defined as
total_media, causing progress logging and response to fail)
- Add URL encoding for paths in scan_output_dir for filenames with spaces
- Exclude thumbnails directory from image monitor to prevent linking thumbnail
files to prompts via fallback mechanism
- Bump version to 3.0.31
Add ability to configure a custom image scan directory in Settings:
- New "Image Scan Directory" field in Settings modal
- "Currently Monitoring" display shows active directory path
- Settings API now returns/saves gallery_root_path and monitored_directories
- Image monitor checks GalleryConfig before auto-detecting
- Restart notification when gallery path changes
Closes#76
Resolve conflicts in py/api.py by keeping feature/button's thumbnail
path generation approach (using rel_path.with_suffix('') and string
formatting) which was tested and working correctly.
- Add Google-style docstrings to main node files (prompt_manager.py, prompt_manager_text.py)
- Enhance database module docstrings with detailed parameter and return documentation
- Add comprehensive docstrings to all utils modules with usage examples
- Document API endpoints and configuration classes thoroughly
- Update module-level docstrings in all __init__.py files
- Ensure consistent naming (ComfyUI_PromptManager) across all documentation
- Follow Python best practices for docstring formatting
Frontend improvements:
- Fixed "Check thumbnails at startup" modal to use EventSource properly
- Shows current file being processed with directory path
- Displays file type (image/video) and action (generating/skipping)
- Real-time progress updates with ETA calculation
- Detailed status messages in scrollable log area
- Increased modal width from max-w-md to max-w-2xl for long filenames
- Path truncation for files over 60 characters
- Console logging for debugging when debug mode enabled
- Better EventSource error handling with connection state logging
Backend improvements:
- Verbose scanning progress with directory counts
- File type breakdown in start message (X images, Y videos)
- More frequent progress updates (every 5 files instead of 10)
- Include current file path, type, and action in progress events
- Send error notifications during processing (first 5 errors)
- Enhanced completion message with error count and processing rate
- Detailed logging throughout the process
The thumbnail generation now provides real-time, detailed feedback in both
the settings panel and the startup prompt modal, making it clear which files
are being processed and the overall progress.
Fixes issue where PromptManager Gallery showed identical prompts for all
images when using workflows like Flux models that generate multiple images
with the same base workflow.
**Problem:**
- Gallery relied solely on PNG metadata extraction for prompt display
- Images generated from same workflow (e.g., Flux models) contained
identical workflow metadata but should show different prompts
- Navigation between images didn't update prompt information correctly
**Solution:**
1. **Enhanced metadata extraction:**
- Improved parseWorkflowData() to handle more ComfyUI node types
- Added support for PromptManagerText, UNETLoader, DualCLIPLoader
- Better negative prompt detection patterns
- Enhanced sampling parameter extraction
2. **New API endpoint:**
- Added GET /prompt_manager/images/prompt/{image_path}
- Queries database for actual prompt used during generation
- Provides accurate prompt data linked to specific images
3. **Hybrid metadata loading:**
- Primary: Query database for linked prompt data
- Fallback: Extract from PNG metadata if no database record
- Combines database prompts with PNG technical parameters
**Changes:**
- py/api.py: Added get_image_prompt() endpoint and route
- web/gallery.html: Enhanced loadImageMetadata() and parseWorkflowData()
**Result:**
- Gallery now displays correct, unique prompts for each image
- Prompt information updates properly when navigating between images
- Generation data refreshes correctly with prev/next navigation
- Maintains backward compatibility for images without database records
Resolves user-reported issue with Flux.1_Krea_Dev_workflow.json where
all images showed identical prompts in gallery view.
Fix issue where every image appears twice in gallery on Windows systems
due to case-insensitive filesystem and path handling differences.
- Deduplicate files when searching with both lowercase/uppercase extensions
- Use pathlib.as_posix() to ensure forward slashes in URLs on all platforms
- Add normalized path tracking to prevent duplicate file collection
The root cause was:
1. Files were found twice (once for *.png and once for *.PNG) on Windows
2. URLs with backslashes caused path inconsistencies
Fixes gallery showing 1922 images when only 961 files exist on disk.
Fix AttributeError when processing PNG files with newer ComfyUI workflow format
where node inputs are stored as lists instead of dictionaries.
- Add _get_node_inputs() helper to normalize both dict and list input formats
- Add _find_text_in_node() to extract text from nodes using multiple strategies
- Update _extract_positive_prompt_from_comfyui_data() to use safe accessors
- Maintain full backward compatibility with existing workflow formats
Fixes error: "'list' object has no attribute 'get'" when scanning images
from workflows using the newer ComfyUI format.
Vito Sansevero