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.
Node code reads cache.raw_data while MetadataSyncService may mutate it
from a background thread; iterate over a list() snapshot to avoid a
possible 'list changed size during iteration' RuntimeError.
- Remove py/nodes/random_checkpoint_loader.py and random_unet_loader.py
- Remove their dedicated test file
- Clean up imports and NODE_CLASS_MAPPINGS in __init__.py
- Update loader-pool comments/docstrings to reference the remaining Checkpoint/Unet Loader nodes' control_after_generate feature
The previous boolean 'control_after_generate': true defaulted the control
widget to 'randomize', silently changing existing workflows into random
model selection on every queue. A string value sets the default mode, so
'fixed' preserves the prior behavior; users opt into randomization
explicitly.
The Checkpoint/Unet Loader (LoraManager) nodes now support ComfyUI's
built-in control_after_generate mechanism on the ckpt_name/unet_name combos,
letting users pick a random model on every queue with the selected model
written back into the widget (visible, and lockable via the 'fixed' mode).
A base_model input narrows the random pool: a front-end extension fetches
the name/base_model mapping from the new /api/lm/checkpoints/loader-pool
endpoint and filters the combo options, wired through the node callback,
the refreshComboInNodes extension hook, and a graph.onConfigure hook
installed from onAdded (onNodeCreated fires before the node is attached to
a graph, so the graph reference is unavailable there).
load_checkpoint returns a 4-tuple (MODEL, CLIP, VAE, model_name) since the
random loader exposes the selected model name; the annotation still claimed
a 3-tuple.
Add dedicated Random Checkpoint/Unet Loader (LoraManager) nodes that pick a random model from the indexed pool on every run, optionally filtered by base_model, and expose the selected model name via a STRING output.
The sampler field now accepts either a manual string or a SAMPLER
connection. When wired, the sampler name is extracted from the
KSAMPLER object's sampler_function __name__ (sample_euler -> euler),
with special-casing for dpm_fast/dpm_adaptive local closures and
uni_pc/uni_pc_bh2 function names.
- sampler input declared as "STRING,SAMPLER" with widgetType STRING,
mirroring the existing model field union pattern
- shared collect_overwrite_params() handles the non-str branch so the
node and the metadata extractor conversion logic stay in sync;
unrecognized sampler functions are logged and skipped
- note: ddim is constructed by ComfyUI as euler with random inpaint,
so the ddim name is unrecoverable and extracts as euler
The model field now accepts either a manual string or a MODEL connection.
When wired, the model name is extracted from the patcher's
cached_patcher_init (registered by core loaders load_checkpoint_guess_config
and load_diffusion_model, preserved through LoRA clones) and converted to a
ComfyUI-style relative name via config model roots.
- model input declared as "STRING,MODEL" with widgetType STRING, so the
text widget and the dual-type connection slot coexist; non-STRING/MODEL
links are rejected by frontend and backend type validation
- UNETLoaderLM GGUF branch now registers a custom cached_patcher_init reload
factory so GGUF models participate in name extraction and ModelPatcher
deepclone/dynamic machinery
- shared collect_overwrite_params() helper keeps the node and the metadata
extractor conversion logic in sync; extraction failures are logged instead
of silently dropping the overwrite
The previous tooltip was misleading: users thought workflow embedding was
automatic. New wording explains this opt-in flag stores the complete
workflow inside images, allowing one-click restoration via drag-and-drop.
PNG and WebP only.
Add two new optional parameters to the Save Image node:
- webp_method (INT, 0-6, default 6): Controls WebP compression level.
0=fastest/largest, 6=slowest/smallest. Previously hardcoded to 0.
- jpeg_subsampling (INT, 0-2, default 0): Controls JPEG chroma
subsampling. 0=4:4:4 (best quality), 1=4:2:2, 2=4:2:0.
Frontend JS extension hides/disables each parameter when the
selected file_format doesn't apply (e.g., webp_method is hidden
when saving as PNG or JPEG). 7 new tests cover parameter plumbing
and default consistency across INPUT_TYPES, save_images(), and
process_image().
- Replace plain-text Lora hashes with Hashes JSON dict matching A1111 convention
- Add Civitai resources JSON array with AIR URNs for direct model version linking
- Add Clip skip, Version: ComfyUI fields to generation params line
- Build AIR strings from local scanner cache (no API calls needed)
- Add complete sampler name mapping (CIVITAI_SAMPLER_MAP) and base model → AIR slug mapping (BASE_MODEL_AIR_SLUG) sourced from civitai ecosystem constants
- Remove lora text prepending from prompt line; LoRA info now in structured JSON sections
Add a pure frontend node that shows filename and editable notes for
a selected LoRA. Connect any output from a LoRA Loader/Stacker/Randomizer/
WanVideoSelect to the lora_source input — selecting a LoRA in the source
widget updates the info display automatically.
- Python node (LoraInfoLM): display-only, no workflow execution
- Vue widget: filename label, auto-sizing notes textarea, save button
with ComfyUI toast feedback on save
- Frontend extension: wire-based selection propagation with stale-response
race guard; clears display on wire disconnect
- Backend: get-notes endpoint now returns file_path alongside notes;
matching supports full-path lora syntax; fix NoneType crash in
trigger words endpoint; document cache file_name invariant
- Wired into all four lora widget nodes (Loader, Stacker, Randomizer,
WanVideoSelect)
Adds lora_syntax_format setting (full/legacy) that controls whether <lora:...> syntax uses relative paths (full) or filename only (legacy). Default is legacy for backward compatibility with A1111 convention. The full path format (<lora:relative/path/filename:strength>) enables lossless model resolution across subfolders.
Ultraworked with Sisyphus (https://github.com/code-yeongyu/oh-my-openagent)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
Autocomplete, copy/send-to-workflow, and recipe syntax now emit
<lora:folder/name:strength> instead of <lora:name:strength>, using
relative paths to disambiguate identically-named loras in different
subfolders without requiring file renames.
Backend: 3-tier hybrid resolution (path → bare → basename fallback)
across get_lora_info, get_lora_info_absolute, get_model_preview_url,
get_model_civitai_url, get_model_info_by_name, get_lora_metadata_by_filename,
and get_hash_by_filename. Also fix get_random_loras and get_cycler_list
to return path-prefixed names for randomizer/cycler consistency.
Frontend: autocomplete, copyLoraSyntax, handleSendToWorkflow emit
folder-prefixed syntax. extract_lora_name preserves relative paths.
Saved image metadata (<lora:...> in EXIF) intentionally keeps basename-only
for compatibility with A1111/Forge ecosystem.
- Use get_lora_info_absolute to obtain correct absolute paths for loras
in LM extra folder paths, instead of folder_paths.get_full_path which
only searches ComfyUI's standard loras directories (returned None)
- Fix name field truncation: str.split('.')[0] stopped at the first dot,
replaced with os.path.splitext to only strip the file extension
- Add _relpath_within_loras helper to preserve subdirectory info in the
name field, matching WanVideoWrapper's os.path.splitext(lora)[0] format
Add name pattern filtering to LoRA Pool node allowing users to filter
LoRAs by filename or model name using either plain text or regex patterns.
Features:
- Include patterns: only show LoRAs matching at least one pattern
- Exclude patterns: exclude LoRAs matching any pattern
- Regex toggle: switch between substring and regex matching
- Case-insensitive matching for both modes
- Invalid regex automatically falls back to substring matching
- Filters apply to both file_name and model_name fields
Backend:
- Update LoraPoolLM._default_config() with namePatterns structure
- Add name pattern filtering to _apply_pool_filters() and _apply_specific_filters()
- Add API parameter parsing for name_pattern_include/exclude/use_regex
- Update LoraPoolConfig type with namePatterns field
Frontend:
- Add NamePatternsSection.vue component with pattern input UI
- Update useLoraPoolState to manage pattern state and API integration
- Update LoraPoolSummaryView to display NamePatternsSection
- Increase LORA_POOL_WIDGET_MIN_HEIGHT to accommodate new UI
Tests:
- Add 7 test cases covering text/regex include, exclude, combined
filtering, model name fallback, and invalid regex handling
Closes#839
- Add `# type: ignore` comments to comfy.sd and folder_paths imports
- Remove unused imports: os, random, and extract_lora_name
- Clean up import statements across checkpoint_loader, lora_randomizer, and unet_loader nodes
- Delay torch import until needed in load_unet and load_unet_gguf methods
- This improves module loading performance by avoiding unnecessary imports
- Maintains functionality while reducing initial import overhead
Move the 'empty/no LoRA' cycling functionality from the LoRA Pool node
to the Lora Cycler widget for cleaner architecture:
Frontend changes:
- Add include_no_lora field to CyclerConfig interface
- Add includeNoLora state and logic to useLoraCyclerState composable
- Add toggle UI in LoraCyclerSettingsView with special styling
- Show 'No LoRA' entry in LoraListModal when enabled
- Update LoraCyclerWidget to integrate new logic
Backend changes:
- lora_cycler.py reads include_no_lora from config
- Calculate effective_total_count (actual count + 1 when enabled)
- Return empty lora_stack when on No LoRA position
- Return actual LoRA count in total_count (not effective count)
Reverted files to pre-PR state:
- lora_loader.py, lora_pool.py, lora_randomizer.py, lora_stacker.py
- lora_routes.py, lora_service.py
- LoraPoolWidget.vue and related files
Related to PR #861
Co-authored-by: dogatech <dogatech@dogatech.home>
Martial Michel