Remove non letters from titles for xem lookup

scanner/providers/implementations/thexem.py

@@ -1,3 +1,4 @@
+import re
 import logging
 from typing import Dict, List, Literal
 from aiohttp import ClientSession
@@ -133,7 +134,9 @@ class TheXem:
 		titles = []
 
 		def clean(s: str):
-			return s.lower().replace(" ", "")
+			s = s.lower()
+			s = re.sub(r"\([^)]*\)", "", s) # remove content of () (guessit does not allow them as part of a name)
+			return re.sub(r"[\W_]+", "", s) # remove non alphanum content (it does keep non us chars like kanjis or accents)
 
 		for x in map.values():
 			# Only the first element is a string (the show name) so we need to ignore the type hint

scanner/scanner/parser/rules.py

@@ -3,6 +3,7 @@
 from typing import Any, List, Optional, cast
 from rebulk import Rule, RemoveMatch, AppendMatch, POST_PROCESS
 from rebulk.match import Matches, Match
+import re
 from copy import copy
 
 
@@ -249,5 +250,8 @@ class XemFixup(Rule):
 		new_title.end = nmatch[0].end
 		new_title.value = f"{title.value}{hole}{nmatch[0].value}"
 
-		if new_title.value.lower().replace(" ", "") in context["xem_titles"]:
+		def clean(s: str):
+			return re.sub(r"[\W_]+", "", s.lower())
+
+		if clean(new_title.value) in context["xem_titles"]:
 			return [[title, nmatch[0]], [new_title]]