Better handle titles list (#445)

scanner/matcher/matcher.py

@@ -51,15 +51,10 @@ class Matcher:
 
 		if "mimetype" not in raw or not raw["mimetype"].startswith("video"):
 			return
-		# Remove seasons in "One Piece (1999) 152.mkv" for example
-		if raw.get("season") == raw.get("year") and "season" in raw:
-			del raw["season"]
 
 		logger.info("Identied %s: %s", path, raw)
 
 		title = raw.get("title")
-		if isinstance(title, list):
-			title = title[0]
 		if not isinstance(title, str):
 			raise ProviderError(f"Could not guess title, found: {title}")
 

scanner/matcher/parser/guess.py

@@ -7,7 +7,7 @@ if __name__ == "__main__":
 	sys.path.append(str(Path(f"{__file__}/../../..").resolve()))
 
 from guessit.api import default_api
-from typing import cast, List
+from typing import cast, List, Any
 from rebulk import Rebulk
 
 try:
@@ -20,14 +20,15 @@ rblk = cast(Rebulk, default_api.rebulk)
 rblk.rules(rules)
 
 
-def guessit(name: str, *, xem_titles: List[str] = []):
+def guessit(name: str, *, xem_titles: List[str] = [], extra_flags: dict[str, Any] = {}):
 	return default_api.guessit(
 		name,
 		{
 			"episode_prefer_number": True,
 			"excludes": "language",
 			"xem_titles": xem_titles,
-		},
+		}
+		| extra_flags,
 	)
 
 
@@ -44,7 +45,11 @@ if __name__ == "__main__":
 		async with ClientSession() as client:
 			xem = TheXemClient(client)
 
-			ret = guessit(sys.argv[1], xem_titles=await xem.get_expected_titles())
+			ret = guessit(
+				sys.argv[1],
+				xem_titles=await xem.get_expected_titles(),
+				# extra_flags={"advanced": True},
+			)
 			print(json.dumps(ret, cls=GuessitEncoder, indent=4))
 
 	asyncio.run(main())

scanner/matcher/parser/rules.py

@@ -1,5 +1,6 @@
 # Read that for examples/rules: https://github.com/pymedusa/Medusa/blob/master/medusa/name_parser/rules/rules.py
 
+from logging import getLogger
 from typing import Any, List, Optional, cast
 from rebulk import Rule, RemoveMatch, AppendMatch, POST_PROCESS
 from rebulk.match import Matches, Match
@@ -7,6 +8,68 @@ from copy import copy
 
 from providers.implementations.thexem import clean
 
+logger = getLogger(__name__)
+
+
+class UnlistTitles(Rule):
+	"""Join titles to a single string instead of a list
+
+	Example: '/media/series/Demon Slayer - Kimetsu no Yaiba/Season 4/Demon Slayer - Kimetsu no Yaiba - S04E10 - Love Hashira Mitsuri Kanroji WEBDL-1080p.mkv'
+	Default:
+	```json
+	 {
+		"title": [
+			"Demon Slayer",
+			"Kimetsu no Yaiba"
+		],
+		"season": 4,
+		"episode_title": "Demon Slayer",
+		"alternative_title": "Kimetsu no Yaiba",
+		"episode": 10,
+		"source": "Web",
+		"screen_size": "1080p",
+		"container": "mkv",
+		"mimetype": "video/x-matroska",
+		"type": "episode"
+	}
+	```
+	Expected:
+	```json
+	{
+		"title": "Demon Slayer - Kimetsu no Yaiba",
+		"season": 4,
+		"episode_title": "Demon Slayer",
+		"alternative_title": "Kimetsu no Yaiba",
+		"episode": 10,
+		"source": "Web",
+		"screen_size": "1080p",
+		"container": "mkv",
+		"mimetype": "video/x-matroska",
+		"type": "episode"
+	}
+	```
+	"""
+
+	priority = POST_PROCESS
+	consequence = [RemoveMatch, AppendMatch]
+
+	def when(self, matches: Matches, context) -> Any:
+		titles: List[Match] = matches.named("title")  # type: ignore
+
+		if not titles or len(titles) <= 1:
+			return
+
+		title = copy(titles[0])
+		for nmatch in titles[1:]:
+			# Check if titles are next to each other, if they are not ignore it.
+			next: List[Match] = matches.next(title)  # type: ignore
+			if not next or next[0] != nmatch:
+				logger.warn(f"Ignoring potential part of title: {nmatch.value}")
+				continue
+			title.end = nmatch.end
+
+		return [titles, [title]]
+
 
 class EpisodeTitlePromotion(Rule):
 	"""Promote "episode_title" to "episode" when the title is in fact the episode number
@@ -253,3 +316,42 @@ class XemFixup(Rule):
 
 		if clean(new_title.value) in context["xem_titles"]:
 			return [[title, nmatch[0]], [new_title]]
+
+
+class SeasonYearDedup(Rule):
+	"""Remove "season" when it's the same as "year"
+
+	Example: "One Piece (1999) 152.mkv"
+	Default:
+	```json
+	{
+		"title": "One Piece",
+		"year": 1999,
+		"season": 1999,
+		"episode": 152,
+		"container": "mkv",
+		"mimetype": "video/x-matroska",
+		"type": "episode"
+	}
+	```
+	Expected:
+	```json
+	{
+		"title": "One Piece",
+		"year": 1999,
+		"episode": 152,
+		"container": "mkv",
+		"mimetype": "video/x-matroska",
+		"type": "episode"
+	}
+	```
+	"""
+
+	priority = POST_PROCESS
+	consequence = [RemoveMatch]
+
+	def when(self, matches: Matches, context) -> Any:
+		season: List[Match] = matches.named("season")  # type: ignore
+		year: List[Match] = matches.named("year")  # type: ignore
+		if len(season) == 1 and len(year) == 1 and season[0].value == year[0].value:
+			return [season]

shell.nix

@@ -1,5 +1,5 @@
 {pkgs ? import <nixpkgs> {}}: let
-  python = pkgs.python311.withPackages (ps:
+  python = pkgs.python312.withPackages (ps:
     with ps; [
       guessit
       aiohttp