Proper rendering computation (god that was a pain)

scanner/scanner/guess/guess.py

@@ -1,13 +1,6 @@
-#!/usr/bin/env python3
-
-if __name__ == "__main__":
-	import sys
-	from pathlib import Path
-
-	sys.path.append(str(Path(f"{__file__}/../../..").resolve()))
+from typing import Any, List, cast
 
 from guessit.api import default_api
-from typing import cast, List, Any
 from rebulk import Rebulk
 from rebulk.match import Match
 
@@ -27,8 +20,7 @@ def guessit(
 	expected_titles: List[str] = [],
 	extra_flags: dict[str, Any] = {},
 ) -> dict[str, list[Match]]:
-	rendering = []
-	ret = default_api.guessit(
+	return default_api.guessit(
 		name,
 		{
 			"episode_prefer_number": True,
@@ -36,34 +28,6 @@ def guessit(
 			"expected_title": expected_titles,
 			"enforce_list": True,
 			"advanced": True,
-			"rendering": rendering,
 		}
 		| extra_flags,
 	)
-	print(rendering)
-	return ret
-
-
-# Only used to test locally
-if __name__ == "__main__":
-	import sys
-	import json
-
-	# from providers.implementations.thexem import TheXemClient
-	from guessit.jsonutils import GuessitEncoder
-	from aiohttp import ClientSession
-	import asyncio
-
-	async def main():
-		async with ClientSession() as client:
-			# xem = TheXemClient(client)
-
-			advanced = any(x == "-a" for x in sys.argv)
-			ret = guessit(
-				sys.argv[1],
-				expected_titles=[],
-				extra_flags={"advanced": advanced},
-			)
-			print(json.dumps(ret, cls=GuessitEncoder, indent=4))
-
-	asyncio.run(main())

scanner/scanner/guess/rules.py

@@ -1,6 +1,5 @@
 # Read that for examples/rules: https://github.com/pymedusa/Medusa/blob/master/medusa/name_parser/rules/rules.py
 
-import hashlib
 from copy import copy
 from logging import getLogger
 from typing import Any, List, Optional, cast
@@ -187,29 +186,3 @@ class SeasonYearDedup(Rule):
 		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
-
-
-# class RenderingDedup(Rule):
-# 	"""Compute rendering (sha of path - version/part)
-#
-# 	Example: "One Piece (1999) v2 152 part2.mkv"
-# 	Computes: sha("One Piece (1999) 152.mkv")
-# 	```
-# 	"""
-#
-# 	priority = POST_PROCESS + 100000
-# 	consequence = AppendMatch
-#
-# 	def when(self, matches: Matches, context: dict[str, list[str]]) -> Any:
-# 		ret = hashlib.new("sha256")
-#
-# 		value: list[Match] = sorted(
-# 			list(matches) + matches.holes(),  # type: ignore
-# 			key=lambda m: m.start,
-# 		)
-# 		for m in value:
-# 			if m.name == "part" or m.name == "version":
-# 				continue
-# 			ret.update(cast(str, m.raw).encode("utf-8"))
-# 		context["rendering"] = [ret.hexdigest()]
-# 		return [Match(start=0, end=1, value=ret.hexdigest(), raw="", name="rendering")]

scanner/scanner/identify.py

@@ -16,11 +16,7 @@ pipeline: list[Callable[[str, Guess], Awaitable[Guess]]] = [
 
 
 async def identify(path: str) -> Video:
-	raw = guessit(
-		path,
-		expected_titles=[],
-		extra_flags={"advanced": True},
-	)
+	raw = guessit(path, expected_titles=[])
 
 	# guessit should only return one (according to the doc)
 	title = raw.get("title", [])[0]
@@ -34,9 +30,14 @@ async def identify(path: str) -> Video:
 	seasons = raw.get("season", [])
 	episodes = raw.get("episode", [])
 
-	rendering = path[:version.start] + path[version.end:]
+	# just strip the version & part number from the path
+	rendering_path = "".join(
+		c
+		for i, c in enumerate(path)
+		if not (version and version.start <= i < version.end)
+		and not (part and part.start <= i < part.end)
+	)
 
-	print(raw)
 	guess = Guess(
 		title=cast(str, title.value),
 		kind=cast(Literal["episode"] | Literal["movie"], kind.value),
@@ -63,7 +64,7 @@ async def identify(path: str) -> Video:
 
 	return Video(
 		path=path,
-		rendering=sha256(path.encode()).hexdigest(),
+		rendering=sha256(rendering_path.encode()).hexdigest(),
 		part=cast(int, part.value) if part else None,
 		version=cast(int, version.value) if version else 1,
 		guess=guess,