- Add cache entry validator service for data integrity checks
- Add cache health monitor service for periodic health checks
- Enhance model cache and scanner with validation support
- Update websocket manager for health status broadcasting
- Add initialization banner service for cache health alerts
- Add comprehensive test coverage for new services
- Update translations across all locales
- Refactor sync translation keys script
Replace recursive directory traversal with first-level-only symlink scanning
to fix severe performance issues on large model collections (220K+ files).
- Rename _scan_directory_links to _scan_first_level_symlinks
- Only scan symlinks directly under each root directory
- Skip traversal of normal subdirectories entirely
- Update tests to reflect first-level behavior
- Add test_deep_symlink_not_scanned to document intentional limitation
Startup time reduced from 15+ minutes to seconds for affected users.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Update package.json test script to run both JS and Vue tests
- Simplify LoraCyclerLM output by removing redundant lora name fallback
- Extend Vitest config to include TypeScript test files
- Add Vue testing dependencies and setup for component testing
- Implement comprehensive test suite for BatchQueueSimulator component
- Add test setup file with global mocks for ComfyUI modules
- Change `model_type` field to `sub_type` for checkpoint models to improve naming consistency
- Add `sub_type="embedding"` for embedding models to properly categorize model subtypes
- Maintain backward compatibility with existing metadata structure
- Add optional `media_type_hint` parameter to `_get_file_extension_from_content_or_headers` method
- When `media_type_hint` is "video" and no extension can be determined from content/headers/URL, default to `.mp4`
- Pass image metadata type as hint in both `process_example_images` and `process_example_images_batch` methods
- Add unit tests to verify media type hint behavior and priority
- Remove backward compatibility code for `model_type` in `ModelScanner._build_cache_entry()`
- Update `CheckpointScanner` to only handle `sub_type` in `adjust_metadata()` and `adjust_cached_entry()`
- Delete deprecated aliases `resolve_civitai_model_type` and `normalize_civitai_model_type` from `model_query.py`
- Update frontend components (`RecipeModal.js`, `ModelCard.js`, etc.) to use `sub_type` instead of `model_type`
- Update API response format to return only `sub_type`, removing `model_type` from service responses
- Revise technical documentation to mark Phase 5 as completed and remove outdated TODO items
All cleanup tasks for the model type refactoring are now complete, ensuring consistent use of `sub_type` across the codebase.
This commit resolves the semantic confusion around the model_type field by
clearly distinguishing between:
- scanner_type: architecture-level (lora/checkpoint/embedding)
- sub_type: business-level subtype (lora/locon/dora/checkpoint/diffusion_model/embedding)
Backend Changes:
- Rename model_type to sub_type in CheckpointMetadata and EmbeddingMetadata
- Add resolve_sub_type() and normalize_sub_type() in model_query.py
- Update checkpoint_scanner to use _resolve_sub_type()
- Update service format_response to include both sub_type and model_type
- Add VALID_*_SUB_TYPES constants with backward compatible aliases
Frontend Changes:
- Add MODEL_SUBTYPE_DISPLAY_NAMES constants
- Keep MODEL_TYPE_DISPLAY_NAMES as backward compatible alias
Testing:
- Add 43 new tests covering sub_type resolution and API response
Documentation:
- Add refactoring todo document to docs/technical/
BREAKING CHANGE: None - full backward compatibility maintained
Move filter preset creation, deletion, application, and storage logic
from FilterManager into a dedicated FilterPresetManager class to
improve separation of concerns and maintainability.
- Add FilterPresetManager with preset CRUD operations
- Update FilterManager to use preset manager via composition
- Handle EMPTY_WILDCARD_MARKER for wildcard base model filters
- Add preset-related translations to all locale files
- Update filter preset UI styling and interactions
Use union type "AUTOCOMPLETE_TEXT_PROMPT,STRING" to enable input mode
compatibility with STRING outputs while preserving autocomplete widget
functionality via widgetType option.
Fixes issue where text inputs could not receive connections from
STRING-type outputs after changing from built-in STRING to custom
AUTOCOMPLETE_TEXT_PROMPT type.
Affected nodes:
- Prompt (LoraManager)
- Text (LoraManager)
Moves onboarding_completed and dismissed_banners from localStorage
to backend settings (settings.json) to survive incognito/private
browser modes.
Fixes#786
Introduce a new TextLM node to the Lora Manager extension, providing a simple text input with autocomplete functionality for tags and styles. The node is integrated into the module's import system and node class mappings, enabling users to utilize autocomplete features for efficient prompt creation.
- Add aliases column to tags table to store comma-separated alias lists
- Update FTS schema to version 2 with searchable_text field containing tag names and aliases
- Implement schema migration to rebuild index when upgrading from old schema
- Modify search logic to match aliases and return canonical tag with matched alias info
- Update index building to include aliases in searchable text for FTS matching
This enables users to search for tag aliases (e.g., "miku") and get results for the canonical tag (e.g., "hatsune_miku") with indication of which alias was matched.
Remove all autocomplete.txt parsing logic and fallback code, simplifying
the service to use only TagFTSIndex for Danbooru/e621 tag search
with category filtering.
- Remove WordEntry dataclass and _words_cache, _file_path attributes
- Remove _determine_file_path(), get_file_path(), load_words(), save_words(),
get_content(), _parse_csv_content() methods
- Simplify search_words() to only use TagFTSIndex, always returning
enriched results with {tag_name, category, post_count}
- Remove GET/POST /api/lm/custom-words endpoints (unused)
- Keep GET /api/lm/custom-words/search for frontend autocomplete
- Rewrite tests to focus on TagFTSIndex integration
This reduces code by 446 lines and removes untested pysssss plugin
integration. Feature is unreleased so no backward compatibility needed.
- Centralize cache path resolution in new py/utils/cache_paths.py module
- Migrate legacy cache files to organized structure: {settings_dir}/cache/{model|recipe|fts|symlink}/
- Automatically clean up legacy files after successful migration with integrity verification
- Update Config symlink cache to use new path and migrate from old location
- Simplify service classes (PersistentModelCache, PersistentRecipeCache, RecipeFTSIndex, TagFTSIndex) to use centralized migration logic
- Add comprehensive test coverage for cache paths and automatic cleanup
Update the placeholder text in the PromptLM class to include guidance for quick tag search functionality. The new placeholder now reads "Enter prompt... /char, /artist for quick tag search", providing users with immediate cues on how to utilize tag search features directly within the input field. This improves usability by making advanced functionality more discoverable.
- Add TagFTSIndex service for fast SQLite FTS5-based tag search (221k+ tags)
- Implement command-mode autocomplete: /char, /artist, /general, /meta, etc.
- Support category filtering via category IDs or names
- Return enriched results with post counts and category badges
- Add UI styling for category badges and command list dropdown
Detects symlink changes at any depth, not just at root level. Uses two-tier validation:
- Fingerprint check for new symlinks
- Deep mapping validation for removed/retargeted symlinks
- Add `_entry_is_symlink` method to detect symlinks and Windows junctions
- Include first-level symlinks in fingerprint for better cache invalidation
- Re-enable preview path validation for security
- Update tests to verify retargeted symlinks trigger rescan
Adds custom words autocomplete functionality similar to comfyui-custom-scripts,
with the following features:
Backend (Python):
- Create CustomWordsService for CSV parsing and priority-based search
- Add API endpoints: GET/POST /api/lm/custom-words and
GET /api/lm/custom-words/search
- Share storage with pysssss plugin (checks for their user/autocomplete.txt first)
- Fallback to Lora Manager's user directory for storage
Frontend (JavaScript/Vue):
- Add 'custom_words' and 'prompt' model types to autocomplete system
- Prompt node now supports dual-mode autocomplete:
* Type 'emb:' prefix → search embeddings
* Type normally → search custom words (no prefix required)
- Add AUTOCOMPLETE_TEXT_PROMPT widget type
- Update Vue component and composable types
Key Features:
- CSV format: word[,priority] compatible with danbooru-tags.txt
- Priority-based sorting: 20% top priority + prefix + include matches
- Preview tooltip for embeddings (not for custom words)
- Dynamic endpoint switching based on prefix detection
Breaking Changes:
- Prompt (LoraManager) node widget type changed from
AUTOCOMPLETE_TEXT_EMBEDDINGS to AUTOCOMPLETE_TEXT_PROMPT
- Removed standalone web/comfyui/prompt.js (integrated into main widgets)
Fixes comfy_dir path calculation by prioritizing folder_paths.base_path
from ComfyUI when available, with fallback to computed path.
- Change `STRING` input type to `AUTOCOMPLETE_TEXT_LORAS` in LoraManagerLoader, LoraStacker, and WanVideoLoraSelectLM nodes for LoRA syntax input
- Change `STRING` input type to `AUTOCOMPLETE_TEXT_EMBEDDINGS` in PromptLoraManager node for prompt input
- Remove manual multiline, autocomplete, and dynamicPrompts configurations in favor of built-in autocomplete types
- Update placeholder text for consistency across nodes
- Remove unused `setupInputWidgetWithAutocomplete` mock from frontend tests
- Add Vue app cleanup logic to prevent memory leaks in widget management
Introduce a new PersistentRecipeCache service that stores recipe metadata in an SQLite database to significantly reduce application startup time. The cache eliminates the need to walk directories and parse JSON files on each launch by persisting recipe data between sessions.
Key features:
- Thread-safe singleton implementation with library-specific instances
- Automatic schema initialization and migration support
- JSON serialization for complex recipe fields (LoRAs, checkpoints, generation parameters, tags)
- File system monitoring with mtime/size validation for cache invalidation
- Environment variable toggle (LORA_MANAGER_DISABLE_PERSISTENT_CACHE) for debugging
- Comprehensive test suite covering save/load cycles, cache invalidation, and edge cases
The cache improves user experience by enabling near-instantaneous recipe loading after the initial cache population, while maintaining data consistency through file change detection.
- Modified backend to fetch last 5 releases from GitHub API
- Updated frontend to iterate through and display multiple releases
- Added latest badge and publish date styling
- Added update.latestBadge translation key to all locales
- Maintains backward compatibility for single changelog display
- Add execution_index and next_index fields to CyclerConfig interface
- Introduce beforeQueued hook in widget to handle index shifting for batch executions
- Use execution_index when provided, fall back to current_index for single executions
- Track execution state with Symbol to differentiate first vs subsequent executions
- Update state management to handle dual-index logic for proper LoRA cycling in batch queues
Removed the sort by selection UI from the Lora Cycler widget and
hardcoded the sorting to always use filename. This simplifies the
interface while maintaining all sorting functionality.
Changes:
- Removed sort_by prop/emit from LoraCyclerSettingsView
- Removed sort tabs UI and associated styles
- Hardcoded sort_by = "filename" in backend node
- Removed sort by handling logic from LoraCyclerWidget
- Updated widget height to accommodate removal
Add Lora Cycler node that cycles through LoRAs sequentially from a filtered pool. Supports configurable sort order, strength settings, and persists cycle progress across workflow save/load.
Backend:
- New LoraCyclerNode with cycle() method
- New /api/lm/loras/cycler-list endpoint
- LoraService.get_cycler_list() for filtered/sorted list
Frontend:
- LoraCyclerWidget with Vue.js component
- useLoraCyclerState composable
- LoraCyclerSettingsView for UI display
Fixes a critical bug in FTS query building where multi-word searches
with field restrictions incorrectly used OR between all word+field
combinations instead of requiring ALL words to match within at least
one field.
Example: searching "cute cat" in {title, tags} previously produced:
title:cute* OR title:cat* OR tags:cute* OR tags:cat*
Which matched recipes with ANY word in ANY field.
Now produces:
(title:cute* title:cat*) OR (tags:cute* tags:cat*)
Which requires ALL words to match within at least one field.
Also adds fallback to fuzzy search when FTS returns empty results,
improving search reliability.
Co-Authored-By: Claude <noreply@anthropic.com>
Temporary workaround for issues #772 and #774 where valid previews
are rejected. Path validation is disabled until proper fix for
preview root path handling is implemented.
- Log preview root rebuilding with counts of different root types
- Add detailed debug output when preview paths are rejected
- Improve visibility into path mapping and validation processes
Previously, `map_path_to_link` and `map_link_to_path` returned the original input path when no mapping was found, instead of the normalized version. This could cause inconsistencies when paths with different representations (e.g., trailing slashes) were used. Now both methods consistently return the normalized path, ensuring uniform path handling throughout the application.
Use os.path.normcase to ensure case-insensitive path matching on Windows, addressing issues where drive letter case mismatches (e.g., 'a:/folder' vs 'A:/folder') prevented correct detection of paths under preview roots. Replace Path.relative_to() with string-based comparison for consistent behavior across platforms.
- Add execution_seed and next_seed parameters to support deterministic randomization across batch executions
- Separate UI display generation from execution stack generation to maintain consistency in batch queues
- Update LoraService to accept optional seed parameter for reproducible randomization
- Ensure each execution with a different seed produces unique results without affecting global random state
- Normalize string quotes to double quotes across all constants for consistency
- Add trailing commas in dictionaries and lists to improve diff readability
- Expand DIFFUSION_MODEL_BASE_MODELS with additional Wan Video and Qwen models
- Fix comment spacing in NSFW_LEVELS dictionary
- Maintain all existing functionality while improving code style
When force=true is passed via API, models in failed_models set are
re-downloaded instead of being skipped. On successful download, model is
removed from failed_models set.
This provides a manual batch repair mechanism for users when CivitAI
media server is temporarily down and causes empty folders.
Changes:
- Backend: Add force parameter to start_download(), _download_all_example_images(), _process_model()
- Backend: Skip failed_models check when force=true
- Backend: Remove model from failed_models on successful force retry
- Frontend: GlobalContextMenu now calls API with force=true directly
- Tests: Update mock to accept force parameter
Introduce a new RecipeFTSIndex class that provides fast prefix-based search across recipe fields (title, tags, LoRA names/models, prompts) using SQLite's FTS5 extension. The implementation supports sub-100ms search times for large datasets (20k+ recipes) and includes asynchronous indexing, incremental updates, and comprehensive unit tests.
The default_unet_root setting was not being synced from backend to frontend
because it was missing from the _SYNC_KEYS tuple in misc_handlers.py. This
caused the "Default Diffusion Model Root" setting to always display "No Default"
even when a valid path was configured in settings.json.
CivitAI does not distinguish between checkpoint and diffusion model types -
both are labeled as "checkpoint". For certain base model types like
"ZImageTurbo", all models are actually diffusion models and should be
saved to the unet/diffusion model folder instead of the checkpoint folder.
- Add DIFFUSION_MODEL_BASE_MODELS constant for known diffusion model types
- Add default_unet_root setting with auto-set logic
- Route downloads to unet folder when baseModel matches known diffusion types
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add failed_model_timestamps to track when models fail
- Retry failed models after 24-hour cooldown period
- Skip retry if example folder already has files
- Skip retry if failure was less than 24 hours ago
- Log count of failed models with retry message
- Fix unbound snapshot variable in exception path
- Remove duplicate/unreachable directory check code
- Update string quotes to double quotes (PEP 8)
This fixes the issue where failed models were permanently skipped in
auto-download mode, even when their example folders were empty.
- Update DebugMetadata node to return metadata via ComfyUI's UI system instead of websocket
- Add new JsonDisplayWidget Vue component for displaying metadata in the UI
- Remove dependency on PromptServer and websocket communication
- Improve error handling with proper UI feedback
- Maintain backward compatibility with existing metadata collection system
When users paste CivArchive URLs, the system now fetches metadata from
CivArchive API first instead of Civitai. This prevents download failures
when a model has been deleted from Civitai but remains available on
CivArchive with alternative mirrors.
Changes:
- Source-aware metadata fetching: Uses CivArchive API when source='civarchive'
- URL prioritization: Prefers non-Civitai mirrors for CivArchive downloads
- Fallback mechanism: Falls back to default provider if CivArchive fails
Fixes#769
- Add folder_include parameter support in backend API handlers
- Add folder_include to FilterCriteria and implement multi-folder filtering logic
- Update frontend to send all include folders instead of only the first
- Add tests for single/multiple include folders, include with exclude, and non-recursive filtering
Add support for respecting recommended strength values from LoRA usage_tips
when randomizing LoRA selection.
Features:
- New toggle setting to enable/disable recommended strength respect (default off)
- Scale range slider (0-2, default 0.5-1.0) to adjust recommended values
- Uses recommended strength × random(scale) when feature enabled
- Fallbacks to original Model/Clip Strength range when no recommendation exists
- Clip strength recommendations only apply when using Custom Range mode
Backend changes:
- Parse usage_tips JSON string to extract strength/clipStrength
- Apply scale factor to recommended values during randomization
- Pass new parameters through API route and node
Frontend changes:
- Update RandomizerConfig type with new properties
- Add new UI section with toggle and dual-range slider
- Wire up state management and event handlers
- No layout shift (removed description text)
Tests:
- Add tests for enabled/disabled recommended strength in API routes
- Add test verifying config passed to service
- All existing tests pass
Build: Include compiled Vue widgets
