merged ln metadata into manga mover

src/AniListResolver.py

@@ -32,27 +32,35 @@ Dependencies
 from __future__ import annotations
 
 import datetime
-import difflib
 import time
 
 import requests
 
 from MediaResolver import MediaResolver
+from TextUtils import best_similarity
 
 
 # --------------------------------------------------------------------------
 # GraphQL query strings
 # --------------------------------------------------------------------------
-_SEARCH_MANGA = """
+# AniList models both manga and light novels as type MANGA; the format
+# clause decides which of the two a search returns.  The placeholder is
+# substituted at construction time (see `media_format`).
+_SEARCH_MANGA_TEMPLATE = """
 query ($search: String) {
   Page(page: 1, perPage: 5) {
-    media(search: $search, type: MANGA, format_not_in: [NOVEL]) {
+    media(search: $search, type: MANGA, __FORMAT_CLAUSE__) {
       id title { romaji english native } siteUrl
     }
   }
 }
 """
 
+_FORMAT_CLAUSES = {
+    "manga": "format_not_in: [NOVEL]",
+    "novel": "format_in: [NOVEL]",
+}
+
 _MANGA_STATS = """
 query ($id: Int) {
   Media(id: $id, type: MANGA) {
@@ -131,10 +139,24 @@ class AniListResolver(MediaResolver):
             cls._instance._initialized = False
         return cls._instance
 
-    def __init__(self, *, request_timeout: int = 30):
+    def __init__(self, *, request_timeout: int = 30,
+                 media_format: str = "manga"):
+        """
+        media_format : "manga" (excludes novels) or "novel" (novels only).
+                       Only the FIRST construction in the process sets it
+                       (singleton); construct the resolver with the correct
+                       format in the entry point / orchestrator.
+        """
         if self._initialized:
             return
 
+        if media_format not in _FORMAT_CLAUSES:
+            raise ValueError(f"media_format must be one of "
+                             f"{sorted(_FORMAT_CLAUSES)}, got {media_format!r}")
+        self.media_format = media_format
+        self._search_query = _SEARCH_MANGA_TEMPLATE.replace(
+            "__FORMAT_CLAUSE__", _FORMAT_CLAUSES[media_format])
+
         self.request_timeout = request_timeout
 
         self._session = requests.Session()
@@ -178,7 +200,7 @@ class AniListResolver(MediaResolver):
             return self._id_cache[key]
 
         try:
-            data = self._gql(_SEARCH_MANGA, {"search": title})
+            data = self._gql(self._search_query, {"search": title})
             results = ((data.get("data") or {})
                        .get("Page", {})
                        .get("media") or [])
@@ -469,18 +491,11 @@ class AniListResolver(MediaResolver):
 def _score_title(query: str, entry: dict) -> float:
     """Returns the best title-similarity score for an AniList media entry."""
     title_obj = entry.get("title") or {}
-    candidates = [
-        title_obj.get("romaji") or "",
-        title_obj.get("english") or "",
-        title_obj.get("native") or "",
-    ]
-    best = 0.0
-    q = query.lower()
-    for t in candidates:
-        if t:
-            ratio = difflib.SequenceMatcher(None, q, t.lower()).ratio()
-            best = max(best, ratio)
-    return best
+    return best_similarity(query, (
+        title_obj.get("romaji"),
+        title_obj.get("english"),
+        title_obj.get("native"),
+    ))
 
 
 # --------------------------------------------------------------------------