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/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).