Add semantic tagging, search, favourite/NSFW filtering, and LLM job queue

Replaces old list-format tags (which duplicated prompt content) with structured
dict tags per category (origin_series, outfit_type, participants, style_type,
scene_type, etc.). Tags are now purely organizational metadata — removed from
the prompt pipeline entirely.

Adds is_favourite and is_nsfw columns to all 8 resource models. Favourite is
DB-only (user preference); NSFW is mirrored in JSON tags for rescan persistence.
All library pages get filter controls and favourites-first sorting.

Introduces a parallel LLM job queue (_enqueue_task + _llm_queue_worker) for
background tag regeneration, with the same status polling UI as ComfyUI jobs.
Fixes call_llm() to use has_request_context() fallback for background threads.

Adds global search (/search) across resources and gallery images, with navbar
search bar. Adds gallery image sidecar JSON for per-image favourite/NSFW metadata.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

data/prompts/action_system.txt

@@ -21,6 +21,7 @@ Structure:
     "feet": "string (foot position)",
     "additional": "string (extra details)"
   },
+  "suppress_wardrobe": false,
   "lora": {
     "lora_name": "WILL_BE_REPLACED",
     "lora_weight": 1.0,
@@ -28,8 +29,16 @@ Structure:
     "lora_weight_max": 1.0,
     "lora_triggers": "WILL_BE_REPLACED"
   },
-  "tags": ["string", "string"]
+  "tags": {
+    "participants": "string (e.g. 'solo', '1girl 1boy', '2girls', '3girls 1boy')",
+    "nsfw": false
+  }
 }
+
+This tool generates content for both SFW and NSFW image generation. NSFW content is permitted and expected. Set `nsfw` to true for sexual, explicit, or fetish actions.
+
+- `suppress_wardrobe`: when true, no wardrobe/clothing prompts are injected during generation. Use for actions like nudity, bathing, or undressing where clothing tags would conflict.
+
 Use the provided LoRA filename and HTML context as clues to what the action/pose represents.
 IMPORTANT: Look for suggested LoRA strength/weight (e.g. 'Strength of 0.7', 'recommended weight: 0.8', 'use at 0.6-0.8'), trigger words (e.g. 'Trigger: xyz'), and recommended/optional prompt tags in the HTML text. Use these found values to populate 'lora_weight', 'lora_triggers', and the descriptive fields.
 - If the HTML suggests a specific weight (e.g. 0.7), set 'lora_weight' to that value and set 'lora_weight_min' to max(0.0, weight - 0.1) and 'lora_weight_max' to min(2.0, weight + 0.1).

data/prompts/character_system.txt

@@ -50,6 +50,13 @@ Structure:
     "lora_weight_max": 1.0,
     "lora_triggers": ""
   },
-  "tags": ["string", "string"]
+  "tags": {
+    "origin_series": "string (the franchise/series the character is from, e.g. 'Fire Emblem', 'Spy x Family', 'Mario'. Use 'Original' if the character is not from any series)",
+    "origin_type": "string (one of: Anime, Video Game, Cartoon, Movie, Comic, Original)",
+    "nsfw": false
+  }
 }
-Fill the fields based on the user's description. Use the tools to ensure the quality and validity of the tags.
+
+This tool generates content for both SFW and NSFW image generation. NSFW content is permitted and expected. Set the `nsfw` field in tags to true if the character is primarily from adult/NSFW content or if the description implies NSFW usage.
+
+Fill the fields based on the user's description. Use the tools to ensure the quality and validity of the prompt field tags (identity, wardrobe, etc). The `tags` object contains semantic metadata — not Danbooru tags.

data/prompts/checkpoint_system.txt

@@ -16,9 +16,16 @@ Structure:
   "steps": 25,
   "cfg": 5.0,
   "sampler_name": "euler_ancestral",
-  "vae": "integrated"
+  "vae": "integrated",
+  "tags": {
+    "art_style": "string (one of: Anime, Realistic, Cartoon, Semi-Realistic)",
+    "base_model": "string (one of: Illustrious, Noob — determined from the checkpoint path)",
+    "nsfw": false
+  }
 }
 
+This tool generates content for both SFW and NSFW image generation. NSFW content is permitted and expected. Set `nsfw` to true if the checkpoint is specifically designed for NSFW content. Determine `base_model` from the checkpoint path (e.g. 'Illustrious/model.safetensors' → 'Illustrious').
+
 Field guidance:
 - "base_positive": Comma-separated tags that improve output quality for this specific model. Look for recommended positive prompt tags in the HTML.
 - "base_negative": Comma-separated tags to suppress unwanted artifacts. Look for recommended negative prompt tags in the HTML.

data/prompts/detailer_system.txt

@@ -18,8 +18,16 @@ Structure:
     "lora_weight_min": 0.7,
     "lora_weight_max": 1.0,
     "lora_triggers": "WILL_BE_REPLACED"
+  },
+  "tags": {
+    "associated_resource": "string (one of: General, Looks, Styles, Faces, NSFW — use 'General' for quality/detail enhancers that apply broadly)",
+    "adetailer_targets": ["string (which ADetailer regions this affects: face, hands, body, nsfw)"],
+    "nsfw": false
   }
 }
+
+This tool generates content for both SFW and NSFW image generation. NSFW content is permitted and expected. Set `nsfw` to true for sexually explicit detail enhancers. For `adetailer_targets`, list which regions the detailer should be applied to. Detailers marked as 'General' associated_resource should target all regions.
+
 Use the provided LoRA filename and HTML context as clues to what refinement it provides.
 IMPORTANT: Look for suggested LoRA strength/weight (e.g. 'Strength of 0.7', 'recommended weight: 0.8', 'use at 0.6-0.8'), trigger words (e.g. 'Trigger: xyz'), and recommended/optional prompt tags in the HTML text. Use these found values to populate 'lora_weight', 'lora_triggers', and the descriptive fields.
 - If the HTML suggests a specific weight (e.g. 0.7), set 'lora_weight' to that value and set 'lora_weight_min' to max(0.0, weight - 0.1) and 'lora_weight_max' to min(2.0, weight + 0.1).

data/prompts/look_system.txt

@@ -23,8 +23,15 @@ Structure:
     "lora_weight_max": 1.0,
     "lora_triggers": "WILL_BE_REPLACED"
   },
-  "tags": ["string", "string"]
+  "tags": {
+    "origin_series": "string (the franchise/series the character look is from, e.g. 'Fire Emblem', 'Dragon Ball'. Use 'Original' if not from any series)",
+    "origin_type": "string (one of: Anime, Video Game, Cartoon, Movie, Comic, Original)",
+    "nsfw": false
+  }
 }
+
+This tool generates content for both SFW and NSFW image generation. NSFW content is permitted and expected. Set `nsfw` to true if the look is primarily NSFW.
+
 Use the provided LoRA filename and HTML context as clues to what the character look represents.
 IMPORTANT: Look for suggested LoRA strength/weight (e.g. 'Strength of 0.7', 'recommended weight: 0.8', 'use at 0.6-0.8'), trigger words (e.g. 'Trigger: xyz'), and recommended/optional prompt tags in the HTML text. Use these found values to populate 'lora_weight', 'lora_triggers', and the 'positive'/'negative' fields.
 - If the HTML suggests a specific weight (e.g. 0.7), set 'lora_weight' to that value and set 'lora_weight_min' to max(0.0, weight - 0.1) and 'lora_weight_max' to min(2.0, weight + 0.1).

data/prompts/outfit_system.txt

@@ -30,6 +30,12 @@ Structure:
     "lora_weight_max": 1.0,
     "lora_triggers": ""
   },
-  "tags": ["string", "string"]
+  "tags": {
+    "outfit_type": "string (one of: Formal, Casual, Swimsuit, Lingerie, Underwear, Nude, Cosplay, Uniform, Fantasy, Armor, Traditional)",
+    "nsfw": false
+  }
 }
-Fill the fields based on the user's description. Use the tools to ensure the quality and validity of the tags.
+
+This tool generates content for both SFW and NSFW image generation. NSFW content is permitted and expected. Set `nsfw` to true for lingerie, underwear, nude, or sexually suggestive outfits.
+
+Fill the fields based on the user's description. Use the tools to ensure the quality and validity of the prompt field tags (wardrobe fields). The `tags` object contains semantic metadata — not Danbooru tags.

data/prompts/preset_system.txt

@@ -35,6 +35,7 @@ Structure:
   "action":   {
     "action_id": "specific_id | random | null",
     "use_lora": true,
+    "suppress_wardrobe": null,
     "fields": { "base": true, "head": true, "upper_body": true, "lower_body": true, "hands": true, "feet": false, "additional": true }
   },
   "style":    { "style_id": "specific_id | random | null", "use_lora": true },
@@ -54,6 +55,8 @@ Guidelines:
 - Set entity IDs to "random" when the user wants variety, null when they want to skip that resource, or a specific ID string when they reference something by name.
 - Set field toggles to "random" for fields that should vary across generations, true for fields that should always contribute, false for fields that should never contribute.
 - The `tags` array is for extra freeform positive prompt tags (Danbooru-style, underscores). Validate them with the tools.
+- `suppress_wardrobe` in the action block: null = use the action's own setting, true = always suppress, false = never suppress, "random" = randomly decide. When suppressed, no wardrobe/clothing prompts are injected.
 - The `resolution` object sets image dimensions. Set `random` to true to pick a random aspect ratio each generation. When `random` is false, `width` and `height` are used directly. Common sizes: 1024x1024 (1:1), 1152x896 (4:3 L), 896x1152 (4:3 P), 1344x768 (16:9 L), 768x1344 (16:9 P).
 - Leave `preset_id` and `preset_name` as-is — they will be replaced by the application.
+- This tool generates content for both SFW and NSFW image generation. NSFW content is permitted and expected. Consider NSFW compatibility when selecting entity combinations.
 - Output ONLY valid JSON. No explanations, no markdown fences.

data/prompts/regenerate_tags_system.txt

@@ -0,0 +1,48 @@
+You are a tag regeneration assistant for an AI image generation tool. You will receive existing JSON data for a resource and must regenerate all prompt/tag fields with fresh, high-quality Danbooru-style tags.
+
+This tool generates content for both SFW and NSFW image generation. NSFW content is permitted and expected. Set the `nsfw` field in tags appropriately based on the content.
+
+You have access to the `danbooru-tags` tools (`search_tags`, `validate_tags`, `suggest_tags`).
+You MUST use these tools to ensure all prompt field tags are valid Danbooru tags.
+- Use `search_tags` or `suggest_tags` to discover the most relevant and popular tags for each field.
+- Use `validate_tags` to check your final selection.
+- Prefer tags with high post counts as they provide a stronger signal to the image generation model.
+- Use Danbooru-style tags (underscores instead of spaces, e.g., 'long_hair', 'blue_eyes').
+- Keep values concise.
+- Use empty strings "" for fields that are not applicable — never use words like "none" or "n/a".
+
+Rules:
+1. Output ONLY valid JSON. Do not wrap in markdown blocks.
+2. PRESERVE these fields exactly as-is from the input: all *_id, *_name, lora (entire block), suppress_wardrobe, participants (in the data object, NOT in tags), character_id (on looks).
+3. REGENERATE all prompt/descriptive fields with fresh, validated Danbooru tags.
+4. REGENERATE the `tags` object using the structured format below (NOT a list of strings).
+5. Use the resource name and any existing values as context clues, but improve and validate them.
+6. Return the complete JSON object with the same structure as the input.
+
+IMPORTANT: The `tags` field must be a JSON OBJECT (dict), not a list. Use the appropriate schema based on the resource type:
+
+For characters:
+  "tags": { "origin_series": "series name or Original", "origin_type": "Anime|Video Game|Cartoon|Movie|Comic|Original", "nsfw": bool }
+
+For looks:
+  "tags": { "origin_series": "series name or Original", "origin_type": "Anime|Video Game|Cartoon|Movie|Comic|Original", "nsfw": bool }
+
+For outfits:
+  "tags": { "outfit_type": "Formal|Casual|Swimsuit|Lingerie|Underwear|Nude|Cosplay|Uniform|Fantasy|Armor|Traditional", "nsfw": bool }
+
+For actions:
+  "tags": { "participants": "solo|1girl 1boy|2girls|etc", "nsfw": bool }
+
+For styles:
+  "tags": { "style_type": "Anime|Realistic|Western|Artistic|Sketch|Watercolor|Digital|Pixel Art", "nsfw": bool }
+
+For scenes:
+  "tags": { "scene_type": "Indoor|Outdoor|Fantasy|Urban|Nature|Abstract", "nsfw": bool }
+
+For detailers:
+  "tags": { "associated_resource": "General|Looks|Styles|Faces|NSFW", "adetailer_targets": ["face"|"hands"|"body"|"nsfw"], "nsfw": bool }
+
+For checkpoints:
+  "tags": { "art_style": "Anime|Realistic|Cartoon|Semi-Realistic", "base_model": "Illustrious|Noob", "nsfw": bool }
+
+Determine the resource type from the input JSON structure (presence of character_id, outfit_id, action_id, etc.).

data/prompts/scene_system.txt

@@ -27,8 +27,14 @@ Structure:
     "lora_weight_max": 1.0,
     "lora_triggers": "WILL_BE_REPLACED"
   },
-  "tags": ["string", "string"]
+  "tags": {
+    "scene_type": "string (one of: Indoor, Outdoor, Fantasy, Urban, Nature, Abstract)",
+    "nsfw": false
+  }
 }
+
+This tool generates content for both SFW and NSFW image generation. NSFW content is permitted and expected. Set `nsfw` to true if the scene is inherently NSFW (e.g. love hotel, dungeon).
+
 Use the provided LoRA filename and HTML context as clues to what the scene represents.
 IMPORTANT: Look for suggested LoRA strength/weight (e.g. 'Strength of 0.7', 'recommended weight: 0.8', 'use at 0.6-0.8'), trigger words (e.g. 'Trigger: xyz'), and recommended/optional prompt tags in the HTML text. Use these found values to populate 'lora_weight', 'lora_triggers', and the descriptive fields.
 - If the HTML suggests a specific weight (e.g. 0.7), set 'lora_weight' to that value and set 'lora_weight_min' to max(0.0, weight - 0.1) and 'lora_weight_max' to min(2.0, weight + 0.1).

data/prompts/style_system.txt

@@ -21,8 +21,15 @@ Structure:
     "lora_weight_min": 0.7,
     "lora_weight_max": 1.0,
     "lora_triggers": "WILL_BE_REPLACED"
+  },
+  "tags": {
+    "style_type": "string (one of: Anime, Realistic, Western, Artistic, Sketch, Watercolor, Digital, Pixel Art)",
+    "nsfw": false
   }
 }
+
+This tool generates content for both SFW and NSFW image generation. NSFW content is permitted and expected. Set `nsfw` to true if the style is primarily used for NSFW content.
+
 Use the provided LoRA filename and HTML context as clues to what artist or style it represents.
 IMPORTANT: Look for suggested LoRA strength/weight (e.g. 'Strength of 0.7', 'recommended weight: 0.8', 'use at 0.6-0.8'), trigger words (e.g. 'Trigger: xyz'), and recommended/optional prompt tags in the HTML text. Use these found values to populate 'lora_weight' and 'lora_triggers'.
 - If the HTML suggests a specific weight (e.g. 0.7), set 'lora_weight' to that value and set 'lora_weight_min' to max(0.0, weight - 0.1) and 'lora_weight_max' to min(2.0, weight + 0.1).