Match Jellyfin punctuation variants using provider metadata and strict fallback

backend/app/clients/jellyfin.py

@@ -1,3 +1,4 @@
+import re
 from typing import Any, Dict, Optional
 import httpx
 import time
@@ -192,15 +193,27 @@ class JellyfinClient(ApiClient):
             "SearchTerm": term,
             "IncludeItemTypes": ",".join(item_types or []),
             "Recursive": "true",
-            "Fields": "Path,MediaSources",
+            "Fields": "Path,MediaSources,ProviderIds,OriginalTitle,SortName",
             "Limit": limit,
         }
         headers = self._emby_headers()
         try:
             async with httpx.AsyncClient(timeout=10.0) as client:
-                response = await client.get(url, headers=headers, params=params)
-                response.raise_for_status()
-                result = response.json()
+                normalized = ' '.join(re.sub(r"[^\w\s]", ' ', term, flags=re.UNICODE).split())
+                terms = list(dict.fromkeys([term, normalized]))
+                if normalized != term and normalized.split():
+                    terms.append(max(normalized.split(), key=len))
+                items = {}
+                for search_term in dict.fromkeys(terms):
+                    if not search_term:
+                        continue
+                    response = await client.get(url, headers=headers, params={**params, "SearchTerm": search_term})
+                    response.raise_for_status()
+                    payload = response.json()
+                    for item in payload.get('Items', []):
+                        if isinstance(item, dict) and item.get('Id'):
+                            items[item['Id']] = item
+                result = {'Items': list(items.values()), 'TotalRecordCount': len(items)}
             duration_ms = round((time.perf_counter() - started_at) * 1000, 2)
             finish_remote_call(
                 operation_event_id,

backend/app/services/snapshot.py

@@ -137,12 +137,10 @@ def jellyfin_item_matches_request(
     request_provider_ids = extract_request_provider_ids(request_payload or {})
     item_provider_ids = extract_request_provider_ids(item)
 
-    provider_priority = ("tmdb", "tvdb", "imdb")
-    for key in provider_priority:
-        request_id = request_provider_ids.get(key)
-        item_id = item_provider_ids.get(key)
-        if request_id and item_id and request_id == item_id:
-            return True
+    shared = set(request_provider_ids) & set(item_provider_ids)
+    if shared:
+        # Conflicting metadata must never fall through to title matching.
+        return all(request_provider_ids[key] == item_provider_ids[key] for key in shared)
 
     request_title = _normalize_media_title(title)
     if not request_title:
@@ -169,11 +167,6 @@ def jellyfin_item_matches_request(
     if request_title in item_titles:
         return True
 
-    if request_type == RequestType.tv:
-        for candidate in item_titles:
-            if candidate and (candidate.startswith(request_title) or request_title.startswith(candidate)):
-                return True
-
     return False