import filters for pixiv sidecars, tag filtering, and favicon generation

app/utils/metadata_enrichment.py

@@ -31,6 +31,7 @@ def find_sidecar_json(image_path: str) -> Optional[str]:
     Gallery-DL creates JSON files with various naming patterns:
     - Patreon: {filename}.json in same directory
     - HentaiFoundry: {artist}-{index}-{title}.json (may differ from image filename)
+    - Pixiv: {id}_p{num}.json for image named {id}_{title}_{num}.png
     - UUID-based: {uuid}.json where UUID is extracted from filename
 
     Returns the path to the sidecar JSON if found, None otherwise.
@@ -50,7 +51,19 @@ def find_sidecar_json(image_path: str) -> Optional[str]:
     if direct_json.exists():
         return str(direct_json)
 
-    # Pattern 3: Extract UUID from filename and look for {uuid}.json
+    # Pattern 3: Pixiv naming pattern
+    # Image: {id}_{title}_{num}.png (e.g., "115358744_Koseki Bijou (Bunny)_00.png")
+    # JSON: {id}_p{num}.json (e.g., "115358744_p0.json")
+    # Extract the leading numeric ID and trailing index
+    pixiv_match = re.match(r'^(\d{6,12})_.*_(\d{2})$', stem)
+    if pixiv_match:
+        pixiv_id = pixiv_match.group(1)
+        pixiv_num = int(pixiv_match.group(2))  # Convert "00" to 0
+        pixiv_json = parent_dir / f"{pixiv_id}_p{pixiv_num}.json"
+        if pixiv_json.exists():
+            return str(pixiv_json)
+
+    # Pattern 4: Extract UUID from filename and look for {uuid}.json
     # Common patterns: "01_UUID", "UUID", "{num}_{UUID}"
     uuid_pattern = re.compile(r'([A-F0-9]{8}-[A-F0-9]{4}-[A-F0-9]{4}-[A-F0-9]{4}-[A-F0-9]{12})', re.IGNORECASE)
     match = uuid_pattern.search(stem)
@@ -60,7 +73,7 @@ def find_sidecar_json(image_path: str) -> Optional[str]:
         if uuid_json.exists():
             return str(uuid_json)
 
-    # Pattern 4: Just the stem without prefix numbers
+    # Pattern 5: Just the stem without prefix numbers
     # e.g., "01_filename" -> look for "filename.json"
     stem_no_prefix = re.sub(r'^\d+_', '', stem)
     if stem_no_prefix != stem:
@@ -68,7 +81,7 @@ def find_sidecar_json(image_path: str) -> Optional[str]:
         if alt_json.exists():
             return str(alt_json)
 
-    # Pattern 5: Extract numeric ID from filename and find JSON with same ID
+    # Pattern 6: Extract numeric ID from filename and find JSON with same ID
     # Handles HentaiFoundry where image is "hentaifoundry_1000220_title.jpg"
     # but JSON is "Artist-1000220-title.json"
     # Look for a 5-7 digit number that could be a post ID
@@ -165,7 +178,7 @@ def load_sidecar_metadata(json_path: str) -> Optional[dict]:
 
 def extract_platform(metadata: dict) -> Optional[str]:
     """Extract the platform name from metadata."""
-    # Check common fields
+    # Check common fields (covers Pixiv which uses "category": "pixiv")
     if "category" in metadata:
         return metadata["category"].lower()
 
@@ -179,6 +192,8 @@ def extract_platform(metadata: dict) -> Optional[str]:
         return "subscribestar"
     if "discord.com" in url:
         return "discord"
+    if "pixiv.net" in url or "pximg.net" in url:
+        return "pixiv"
 
     return None
 
@@ -195,8 +210,13 @@ def extract_artist(metadata: dict) -> Optional[str]:
     if "artist" in metadata:
         return metadata["artist"]
 
-    # For hentaifoundry
-    if "user" in metadata:
+    # For Pixiv - user is an object with name and account fields
+    if "user" in metadata and isinstance(metadata["user"], dict):
+        # Prefer account (username) over display name for consistency
+        return metadata["user"].get("account") or metadata["user"].get("name")
+
+    # For hentaifoundry - user is a string
+    if "user" in metadata and isinstance(metadata["user"], str):
         return metadata["user"]
 
     return None
@@ -238,6 +258,7 @@ def extract_description(metadata: dict) -> Optional[str]:
     - Patreon: 'content' (may be HTML)
     - SubscribeStar: 'content' (may be HTML)
     - HentaiFoundry: 'description' (plain text)
+    - Pixiv: 'caption' (plain text or HTML)
     """
     # Try description first (HentaiFoundry uses this)
     if "description" in metadata:
@@ -245,6 +266,12 @@ def extract_description(metadata: dict) -> Optional[str]:
         if desc and isinstance(desc, str) and desc.strip():
             return desc.strip()
 
+    # Try caption (Pixiv uses this)
+    if "caption" in metadata:
+        caption = metadata["caption"]
+        if caption and isinstance(caption, str) and caption.strip():
+            return caption.strip()
+
     # Try content (Patreon/SubscribeStar use this)
     if "content" in metadata:
         content = metadata["content"]
@@ -268,14 +295,23 @@ def extract_source_url(metadata: dict) -> Optional[str]:
     - Patreon: 'url' (direct post link)
     - SubscribeStar: construct from author_name and post_id
     - HentaiFoundry: construct from user and index
+    - Pixiv: construct from id
     """
     platform = extract_platform(metadata)
 
-    # Direct URL field (Patreon)
+    # Construct URL for Pixiv (do this first to avoid using the pximg.net image URL)
+    if platform == "pixiv":
+        post_id = metadata.get("id")
+        if post_id:
+            return f"https://www.pixiv.net/artworks/{post_id}"
+
+    # Direct URL field (Patreon) - but not for Pixiv where 'url' is the image URL
     if "url" in metadata:
         url = metadata["url"]
         if url and isinstance(url, str) and url.startswith("http"):
-            return url
+            # Skip pximg.net URLs (these are image URLs, not post URLs)
+            if "pximg.net" not in url:
+                return url
 
     # Construct URL for SubscribeStar
     if platform == "subscribestar":
@@ -298,6 +334,10 @@ def extract_attachment_count(metadata: dict) -> int:
     """
     Extract the number of attachments/images in the post.
     """
+    # Pixiv: use page_count field
+    if "page_count" in metadata:
+        return int(metadata["page_count"])
+
     # Patreon: check images array or post_metadata.image_order
     if "images" in metadata and isinstance(metadata["images"], list):
         return len(metadata["images"])
@@ -319,7 +359,8 @@ def extract_date(metadata: dict) -> Optional[datetime]:
     Extract the publication/post date from metadata.
     Tries multiple common date fields and formats.
     """
-    date_fields = ["published_at", "date", "created_at", "timestamp"]
+    # Pixiv uses create_date, others use published_at, date, etc.
+    date_fields = ["published_at", "create_date", "date", "created_at", "timestamp"]
 
     for field in date_fields:
         if field not in metadata: