"""Image monitoring system for ComfyUI generated images. This module provides real-time monitoring of ComfyUI output directories to automatically detect newly generated images and associate them with their corresponding prompts. The system uses filesystem watchers to detect image creation events and extract metadata from the images to maintain a gallery system. The main components are: - ImageGenerationHandler: Handles filesystem events for new image creation - ImageMonitor: Main monitoring system that manages directory watching Typical usage: from utils.image_monitor import ImageMonitor monitor = ImageMonitor(db_manager, prompt_tracker) monitor.start_monitoring(['/path/to/comfyui/output']) The system automatically: - Detects new image files in monitored directories - Extracts ComfyUI workflow metadata from PNG chunks - Links images to active prompts using the prompt tracker - Handles fallback linking when no active prompt is available - Provides status information and monitoring control """ import os import queue import time import threading from typing import Optional, Dict, Any, Tuple from watchdog.observers import Observer from watchdog.events import FileSystemEventHandler from .hashing import generate_prompt_hash from .metadata_extractor import ComfyUIMetadataExtractor from .logging_config import get_logger from .prompt_graph import resolve_text, run_prompt_nodes from .prompt_pair_resolver import resolve_prompt_pair try: from ..database.pair_ingestion import ingest_resolved_pair except ImportError: from database.pair_ingestion import ingest_resolved_pair # Upper bound on images waiting to be processed (a burst beyond this is skipped) MAX_PENDING_IMAGES = 1000 # A file counts as fully written once its size is unchanged across two reads SETTLE_INTERVAL_SECONDS = 0.1 SETTLE_ATTEMPTS = 20 class ImageGenerationHandler(FileSystemEventHandler): """Filesystem event handler for detecting new image generation. This handler extends watchdog's FileSystemEventHandler to specifically handle new image file creation events in ComfyUI output directories. When a new image is detected, it attempts to: 1. Extract ComfyUI metadata from the image 2. Associate the image with the currently active prompt 3. Store the relationship in the database The handler implements a small delay before processing to ensure files are completely written before attempting to read them. """ def __init__(self, db_manager, prompt_tracker): """ Initialize the image generation handler. Args: db_manager: Database manager instance for storing image-prompt relationships prompt_tracker: Prompt tracking instance for getting current active prompts """ self.db_manager = db_manager self.prompt_tracker = prompt_tracker self.metadata_extractor = ComfyUIMetadataExtractor() self.logger = get_logger("prompt_manager.image_monitor") # Read from GalleryConfig if available, otherwise use defaults try: from ..py.config import GalleryConfig self.processing_delay = GalleryConfig.PROCESSING_DELAY self.supported_extensions = tuple(GalleryConfig.SUPPORTED_EXTENSIONS) except Exception: self.processing_delay = 2.0 self.supported_extensions = (".png", ".jpg", ".jpeg", ".webp", ".gif") # Settle check: the file size must be unchanged across two reads self.settle_interval = SETTLE_INTERVAL_SECONDS self.settle_attempts = SETTLE_ATTEMPTS # Images scheduled but not yet processed: path -> (prompt snapshot, due time). # One daemon worker processes them in order, so batch images pop the # prompt queue in creation order. self._pending: Dict[str, Tuple[Any, float]] = {} self._pending_lock = threading.Lock() # Each worker owns its queue and stop event (replaced on shutdown), # so a worker that is still draining never swallows a later image. self._work: "queue.Queue[Optional[str]]" = queue.Queue() self._stop = threading.Event() self._worker: Optional[threading.Thread] = None def on_created(self, event): """Handle filesystem creation events. This method is called by watchdog when a new file is created in a monitored directory. It filters for image files and schedules them for processing after a small delay to ensure the file is fully written. The prompt context is captured immediately (before the delay) to handle batch workflows where the prompt tracker advances before images are processed. Args: event: FileSystemEvent object containing event details """ if event.is_directory or not self.is_image_file(event.src_path): return self.logger.info(f"New image detected: {event.src_path}") # Snapshot prompt context NOW before the delay — in batch workflows # the tracker advances to the next prompt before images are processed. prompt_snapshot = self.prompt_tracker.get_current_prompt() if prompt_snapshot: self.logger.debug( f"Snapshot prompt {prompt_snapshot.get('id', '?')} for " f"{os.path.basename(event.src_path)}" ) self.schedule(event.src_path, prompt_snapshot) def schedule(self, image_path: str, prompt_snapshot=None) -> bool: """Queue an image for processing on the worker thread, once per path. Images are processed in the order they were scheduled (batch images must pop the prompt queue in creation order), each after ``processing_delay`` and once its size has stopped changing. Returns: False when the path is already pending or the pending set is full """ with self._pending_lock: if image_path in self._pending: self.logger.debug(f"Already pending, not rescheduled: {image_path}") return False if len(self._pending) >= MAX_PENDING_IMAGES: self.logger.warning( f"Too many images pending ({len(self._pending)}), " f"skipping: {image_path}" ) return False due = time.monotonic() + self.processing_delay self._pending[image_path] = (prompt_snapshot, due) self._ensure_worker() self._work.put(image_path) return True def pending_count(self) -> int: """Number of images scheduled but not yet processed.""" with self._pending_lock: return len(self._pending) def shutdown(self) -> None: """Stop the worker thread and drop every image still waiting. The worker wakes from any delay or settle wait, finishes nothing more and exits; the next ``schedule`` starts a fresh worker with its own queue. Safe to call repeatedly and before anything was scheduled. """ with self._pending_lock: self._pending.clear() worker, self._worker = self._worker, None work, self._work = self._work, queue.Queue() stop, self._stop = self._stop, threading.Event() stop.set() if worker is not None and worker.is_alive(): work.put(None) # sentinel: wake a worker idle on an empty queue def _ensure_worker(self) -> None: """Start the single daemon worker thread if it isn't running. Called with ``_pending_lock`` held. """ if self._worker is None or not self._worker.is_alive(): self._worker = threading.Thread( target=self._run_worker, args=(self._work, self._stop), name="PromptManagerImageWorker", daemon=True, ) self._worker.start() def _run_worker(self, work: "queue.Queue[Optional[str]]", stop) -> None: while not stop.is_set(): image_path = work.get() try: if image_path is None or stop.is_set(): break self._process_pending(image_path, stop) except Exception as e: self.logger.error(f"Image worker failed on {image_path}: {e}") finally: work.task_done() def _process_pending(self, image_path: str, stop=None) -> None: """Process one scheduled image; the pending entry is always cleared.""" stop = stop if stop is not None else self._stop try: with self._pending_lock: prompt_snapshot, due = self._pending.get(image_path, (None, 0.0)) remaining = due - time.monotonic() if remaining > 0 and stop.wait(remaining): return if not self.wait_until_settled(image_path, stop): if not stop.is_set(): self.logger.warning( f"Image never settled or vanished, skipping: {image_path}" ) return self.process_new_image(image_path, prompt_snapshot=prompt_snapshot) finally: with self._pending_lock: self._pending.pop(image_path, None) def wait_until_settled(self, image_path: str, stop=None) -> bool: """True once the file's size is non-zero and unchanged across two reads. Returns False as soon as ``stop`` (an Event) is set. """ stop = stop if stop is not None else self._stop previous = None for _ in range(self.settle_attempts): try: size = os.path.getsize(image_path) except OSError: return False if previous is not None and size == previous and size > 0: return True previous = size if stop.wait(self.settle_interval): return False return False def is_image_file(self, filepath: str) -> bool: """Check if file is a supported image format and not a thumbnail. Args: filepath: Path to the file to check Returns: True if the file has a supported image extension and is not inside a thumbnails directory, False otherwise """ # Skip files in thumbnails directory - those are derivatives, not # generated images if "/thumbnails/" in filepath or "\\thumbnails\\" in filepath: return False return filepath.lower().endswith(self.supported_extensions) def process_new_image(self, image_path: str, prompt_snapshot=None): """Process a newly created image file for gallery integration. This method handles the complete processing pipeline for a new image: 1. Verifies the file still exists 2. Extracts ComfyUI metadata from the image 3. Tries to identify the correct prompt via metadata-based lookup 4. Falls back to the prompt snapshot captured at event time 5. Links the image to the appropriate prompt in the database Args: image_path: Full path to the newly created image file prompt_snapshot: Prompt context captured at file-creation time (optional) """ try: self.logger.info(f"Processing image: {image_path}") if not os.path.exists(image_path): self.logger.warning(f"Image file no longer exists: {image_path}") return # Extract ComfyUI metadata first — needed both for linking and prompt lookup metadata = None try: metadata = self.metadata_extractor.extract_metadata(image_path) self.logger.debug(f"Extracted metadata: {bool(metadata)}") except Exception as meta_error: self.logger.warning(f"Metadata extraction failed: {meta_error}") # Strategy 1: The image's own metadata. It records the positive prompt # that produced this image, independent of node caching and of any # negative-prompt PromptManager nodes in the workflow. state, current_prompt = self._find_prompt_from_metadata(metadata) if state == "foreign": # Another workflow saved this image: it belongs to no prompt of # ours, and must not consume a batch queue entry. self.logger.info( f"No PromptManager prompt in workflow, skipping image: {image_path}" ) return if current_prompt: self.logger.info( f"Metadata match: prompt {current_prompt['id']} for " f"{os.path.basename(image_path)}" ) # Strategy 2: Pop from batch queue. Batch items (text supplied by # PromptSearchList) share one graph, so metadata can't tell them apart; # they are queued in encoding order and images save in the same order. if not current_prompt: current_prompt = self.prompt_tracker.pop_next_prompt() if current_prompt: self.logger.info( f"Queue match: prompt {current_prompt['id']} for " f"{os.path.basename(image_path)}" ) # Strategy 3: Use snapshot captured at file-creation time if not current_prompt and prompt_snapshot: current_prompt = prompt_snapshot self.logger.info( f"Snapshot match: prompt {current_prompt.get('id', 'unknown')}" ) # Strategy 4: Check live prompt tracker if not current_prompt: current_prompt = self.prompt_tracker.get_current_prompt() if current_prompt: self.logger.info( f"Live tracker match: prompt {current_prompt['id']}" ) # Strategy 5: Fallback to most recent prompt in DB if not current_prompt: current_prompt = self._get_fallback_prompt() if current_prompt: self.logger.debug(f"Fallback match: prompt {current_prompt['id']}") else: self.logger.warning( f"No prompt context available, skipping image: {image_path}" ) return # Extend timeout if we have an active execution if current_prompt.get("execution_id"): self.prompt_tracker.extend_prompt_timeout( current_prompt["execution_id"], 300 ) if metadata: self.logger.debug( f"Linking image with full metadata to prompt {current_prompt['id']}" ) self.link_image_to_prompt(image_path, current_prompt, metadata) else: self.logger.debug( f"Linking image with basic info to prompt {current_prompt['id']}" ) basic_metadata = self.get_basic_file_info(image_path) self.link_image_to_prompt( image_path, current_prompt, {"file_info": basic_metadata} ) except Exception as e: self.logger.error(f"Error processing image {image_path}: {e}") import traceback self.logger.error(traceback.format_exc()) def _find_prompt_from_metadata(self, metadata) -> Tuple[str, Optional[Dict]]: """Identify the prompt that produced an image from its embedded metadata. Parses the ComfyUI workflow/prompt data embedded in the image to find the PromptManager node feeding the positive input, then matches its text against the database by hash. Args: metadata: Extracted metadata dict from the image, or None Returns: ``("linked", prompt)`` with a context dict holding 'id' and 'text'; ``("foreign", None)`` when the workflow has no positive PromptManager node, so the image belongs to no prompt of ours; ``("unknown", None)`` when metadata can't decide (missing, a batch item whose text came from another node, or a lookup failure) and the caller's queue and tracker fallbacks apply. """ if not metadata: return "unknown", None prompt_text = None resolution = resolve_prompt_pair(metadata) graph = metadata.get("prompt") if resolution["state"] == "resolved" and isinstance(graph, dict): positive = resolution["positive"] path = positive["provenance"].get("path", []) paired_source = any( isinstance(graph.get(item.get("node_id")), dict) and graph[item["node_id"]].get("class_type") == "PromptManagerPair" for item in path ) if paired_source: try: existing = self.db_manager.get_prompt_by_hash( generate_prompt_hash(positive["storage_text"]) ) if existing: return "linked", { "id": existing["id"], "text": existing["text"], "from_metadata": True, } except Exception as error: self.logger.warning(f"Paired metadata lookup failed: {error}") # The executed graph: use the PromptManager node feeding the positive input prompt_data = metadata.get("prompt") if isinstance(prompt_data, dict): run_nodes = run_prompt_nodes(prompt_data) if not run_nodes: return "foreign", None for node_id in run_nodes: prompt_text = resolve_text(prompt_data, node_id) if prompt_text: break if not prompt_text: # Positive text can't be known from the graph (batch item or unknown # node): leave it to the queue rather than guess from workflow widgets return "unknown", None # Fallback (no executed graph): check text_encoder_nodes from workflow, but # only if the text input is NOT connected (connected inputs override widget # values, so the widget value would be stale in batch workflows). if not prompt_text: text_nodes = metadata.get("text_encoder_nodes", []) if text_nodes and not any( "PromptManager" in (n.get("type") or n.get("class_type") or "") for n in text_nodes if isinstance(n, dict) ): return "foreign", None for node in text_nodes: node_type = node.get("type") or node.get("class_type") or "" if "PromptManager" in node_type: # Check if text input is connected — if so, widget value is stale text_connected = False for inp in node.get("inputs", []): if isinstance(inp, dict) and inp.get("name") == "text": if inp.get("link") is not None: text_connected = True break if text_connected: self.logger.debug( "PromptManager text input is connected — " "skipping stale widget value, deferring to snapshot" ) continue widgets = node.get("widgets_values", []) if widgets and isinstance(widgets[0], str) and widgets[0].strip(): prompt_text = widgets[0].strip() break if not prompt_text: return "unknown", None # Look up by hash in database try: existing = self.db_manager.get_prompt_by_hash( generate_prompt_hash(prompt_text) ) if existing: self.logger.debug( f"Found DB prompt {existing['id']} from metadata text" ) return "linked", { "id": existing["id"], "text": existing["text"], "from_metadata": True, } except Exception as e: self.logger.warning(f"Metadata-based prompt lookup failed: {e}") return "unknown", None def get_basic_file_info(self, image_path: str) -> Dict[str, Any]: """Get basic file information when metadata extraction fails. Provides fallback file information when ComfyUI metadata cannot be extracted from the image. Includes file size, format, and dimensions when possible. Args: image_path: Path to the image file Returns: Dictionary containing basic file information: - size: File size in bytes - format: Image format (PNG, JPEG, etc.) - dimensions: Image width and height as list [width, height] """ try: from PIL import Image stat = os.stat(image_path) file_info = {"size": stat.st_size, "format": None, "dimensions": None} # Try to get image dimensions try: with Image.open(image_path) as img: file_info["dimensions"] = list(img.size) file_info["format"] = img.format except Exception: pass return file_info except Exception as e: self.logger.error(f"Error getting file info: {e}") return {} def _get_fallback_prompt(self) -> Optional[Dict[str, Any]]: """Get the most recent prompt from database as fallback. When no active prompt is available from the tracker, this method attempts to find the most recently created prompt in the database to use as a fallback for image linking. Returns: Dictionary containing prompt information with 'fallback' flag set to True, or None if no recent prompt is available """ try: recent = self.db_manager.get_recent_prompts(limit=1) # Paginated dict ({"prompts": [...]}) or a plain list recent_prompts = ( recent.get("prompts") if isinstance(recent, dict) else recent ) if recent_prompts: prompt = recent_prompts[0] return { "id": prompt["id"], "text": prompt["text"], "timestamp": prompt.get("created_at"), "fallback": True, } except Exception as e: self.logger.error(f"Error getting fallback prompt: {e}") return None def link_image_to_prompt( self, image_path: str, prompt_context: Dict, metadata: Dict ): """Link an image to a prompt in the database. Creates a database record associating the generated image with its source prompt, including any extracted metadata from the image file. Args: image_path: Full path to the image file prompt_context: Dictionary containing prompt information including ID and text metadata: Extracted metadata from the image file (workflow, parameters, etc.) """ try: resolution = resolve_prompt_pair(metadata) if resolution["state"] == "resolved": positive = resolution["positive"] expected = self.db_manager.get_prompt_by_hash( generate_prompt_hash(positive.get("storage_text", positive["text"])) ) if expected and expected["id"] == prompt_context["id"]: ingest_resolved_pair( self.db_manager, image_path, metadata, resolution ) return image_id = self.db_manager.link_image_to_prompt( prompt_id=prompt_context["id"], image_path=image_path, metadata=metadata ) fallback_note = " (fallback)" if prompt_context.get("fallback") else "" self.logger.debug( f"Successfully linked image {image_id} to prompt " f"{prompt_context['id']}{fallback_note}" ) except Exception as e: self.logger.error(f"Failed to link image to prompt: {e}") class ImageMonitor: """Main image monitoring system for ComfyUI gallery integration. This class manages the overall image monitoring system, including: - Setting up filesystem watchers for output directories - Auto-detecting ComfyUI output locations - Managing the lifecycle of monitoring operations - Providing status information The monitor uses watchdog to efficiently watch filesystem changes and can monitor multiple directories simultaneously with recursive subdirectory support. """ def __init__(self, db_manager, prompt_tracker): """ Initialize the image monitor. Args: db_manager: Database manager instance for storing image relationships prompt_tracker: Prompt tracking instance for getting active prompt context """ self.db_manager = db_manager self.prompt_tracker = prompt_tracker self.observer = None self.handler = None self.monitored_directories = [] self.logger = get_logger("prompt_manager.image_monitor") def start_monitoring(self, output_directories: Optional[list] = None): """ Start monitoring ComfyUI output directories for new images. Begins filesystem watching on the specified directories. If no directories are provided, the system will first check GalleryConfig.MONITORING_DIRECTORIES, then fall back to auto-detecting ComfyUI output locations. All monitoring is done recursively to catch images in subdirectories. Args: output_directories: List of directory paths to monitor. If None, uses config or auto-detection. """ if self.observer: self.logger.warning("Image monitoring already running") return # Check config for monitoring settings config = self._gallery_config() if config is not None: if not config.MONITORING_ENABLED: self.logger.info("Image monitoring disabled in config") return # Use configured directories if set if not output_directories and config.MONITORING_DIRECTORIES: output_directories = config.MONITORING_DIRECTORIES self.logger.info( f"Using configured monitoring directories: {output_directories}" ) # Auto-detect ComfyUI output directory if still none if not output_directories: output_directories = self.detect_comfyui_output_dirs() valid_directories = [] for output_dir in output_directories or []: if not os.path.isdir(output_dir): self.logger.warning(f"Directory does not exist: {output_dir}") continue # Same containment rule as the settings endpoint, so a root # hand-edited into config.json is never watched (and its files # never linked) when it lies outside ComfyUI's directories. validate = getattr(config, "validate_gallery_root", None) if callable(validate): ok, reason = validate(output_dir) if not ok: self.logger.warning(f"Not watching {output_dir}: {reason}") continue valid_directories.append(output_dir) if not valid_directories: self.logger.warning("No valid directories to monitor") return # Only a started observer becomes self.observer, so a failed start never # leaves the monitor stuck "already running" with nothing to stop. handler = ImageGenerationHandler(self.db_manager, self.prompt_tracker) observer = Observer() scheduled = [] for output_dir in valid_directories: try: observer.schedule(handler, output_dir, recursive=True) scheduled.append(output_dir) self.logger.info(f"Monitoring directory (recursive): {output_dir}") except Exception as e: self.logger.warning(f"Cannot watch {output_dir}: {e}") if not scheduled: self.logger.warning("No directory could be watched") return try: observer.start() except Exception as e: self.logger.error(f"Image monitoring could not start: {e}") return self.handler = handler self.observer = observer self.monitored_directories = scheduled self.logger.info(f"Image monitoring started for {len(scheduled)} directories") @staticmethod def _gallery_config(): """GalleryConfig when the package's config is importable, else None.""" try: from ..py.config import GalleryConfig return GalleryConfig except Exception: return None @property def is_monitoring(self) -> bool: """Whether a filesystem observer is running.""" return self.observer is not None def stop_monitoring(self): """Stop the image monitoring system. Cleanly shuts down the filesystem watcher and clears all monitoring state. This method should be called before program exit to ensure proper cleanup. Safe to call when monitoring never started. """ observer, self.observer = self.observer, None handler, self.handler = self.handler, None self.monitored_directories = [] if observer is None: return try: observer.stop() observer.join() except Exception as e: self.logger.warning(f"Error while stopping image monitoring: {e}") if handler is not None: try: handler.shutdown() except Exception as e: self.logger.warning(f"Error while stopping the image worker: {e}") self.logger.debug("Image monitoring stopped") def detect_comfyui_output_dirs(self) -> list: """Auto-detect ComfyUI output directories. Attempts to locate ComfyUI output directories using multiple strategies: 1. Import ComfyUI's folder_paths module to get the configured output directory 2. Search common relative paths where ComfyUI output directories are typically located 3. Verify that detected directories actually exist Returns: List of absolute paths to detected output directories """ potential_dirs = [] try: # Try to import ComfyUI's folder_paths import folder_paths output_dir = folder_paths.get_output_directory() if output_dir and os.path.exists(output_dir): potential_dirs.append(output_dir) self.logger.debug(f"Detected ComfyUI output directory: {output_dir}") except ImportError: self.logger.debug( "ComfyUI folder_paths not available, using fallback detection" ) # Fallback: Look for common ComfyUI directory structures fallback_paths = [ "output", "../output", "../../output", "ComfyUI/output", "../ComfyUI/output", ] for path in fallback_paths: abs_path = os.path.abspath(path) if os.path.exists(abs_path) and abs_path not in potential_dirs: potential_dirs.append(abs_path) self.logger.debug(f"Found output directory: {abs_path}") return potential_dirs def get_status(self) -> Dict[str, Any]: """Get monitoring status information. Returns: Dictionary containing: - running: Boolean indicating if monitoring is active - monitored_directories: List of currently monitored directory paths - handler_active: Boolean indicating if the event handler is active - observer_alive: Boolean indicating if observer thread is alive """ observer_alive = False if self.observer is not None: try: observer_alive = self.observer.is_alive() except Exception: pass return { "running": self.observer is not None, "monitored_directories": self.monitored_directories, "handler_active": self.handler is not None, "observer_alive": observer_alive, } # Singleton instance management _monitor_instance: Optional[ImageMonitor] = None _monitor_lock = threading.Lock() def get_image_monitor(db_manager, prompt_tracker) -> ImageMonitor: """Get or create the singleton ImageMonitor instance. This ensures only one ImageMonitor exists across all PromptManager nodes, preventing duplicate image detection when multiple nodes are used. Args: db_manager: Database manager instance for storing image relationships prompt_tracker: Prompt tracking instance for getting active prompt context Returns: The singleton ImageMonitor instance """ global _monitor_instance if _monitor_instance is None: with _monitor_lock: if _monitor_instance is None: _monitor_instance = ImageMonitor(db_manager, prompt_tracker) return _monitor_instance