From 4cb2b369ab982faec305fa64fa97b58d73fba1a6 Mon Sep 17 00:00:00 2001
From: Estrella Pan <estrellaxd05@gmail.com>
Date: Sun, 12 Jul 2026 21:07:12 +0200
Subject: [PATCH] fix(parser): preserve CJK movie words in titles

---
 .../parser/analyser/tokenizer/classic.py      |  5 +++-
 .../parser/analyser/tokenizer/parser.py       |  5 +++-
 backend/src/test/test_parser_selector.py      | 28 +++++++++++++++++++
 3 files changed, 36 insertions(+), 2 deletions(-)

diff --git a/backend/src/module/parser/analyser/tokenizer/classic.py b/backend/src/module/parser/analyser/tokenizer/classic.py
index 820ea387..ffb650c6 100644
--- a/backend/src/module/parser/analyser/tokenizer/classic.py
+++ b/backend/src/module/parser/analyser/tokenizer/classic.py
@@ -71,7 +71,10 @@ _PV = re.compile(
 )
 _NCOP = re.compile(r"(?<![A-Za-z0-9])NCOP(?:v?(\d+))?(?![A-Za-z0-9])", re.I)
 _NCED = re.compile(r"(?<![A-Za-z0-9])NCED(?:v?(\d+))?(?![A-Za-z0-9])", re.I)
-_MOVIE_CJK = re.compile(r"劇場版?|剧场版?|電影版|电影版")
+_MOVIE_CJK = re.compile(
+    r"劇場版|剧场版|電影版|电影版|"
+    r"(?<![\u3400-\u9fff])(?:劇場|剧场)(?![\u3400-\u9fff])"
+)
 _MOVIE_ROMAJI = re.compile(r"(?<!\w)Gekijou-?ban(?!\w)", re.I)
 _MOVIE_EN = re.compile(r"(?<!\w)(?:The\s+)?Movie(?!\w)", re.I)
 
diff --git a/backend/src/module/parser/analyser/tokenizer/parser.py b/backend/src/module/parser/analyser/tokenizer/parser.py
index 4bbc097e..8dd1c330 100644
--- a/backend/src/module/parser/analyser/tokenizer/parser.py
+++ b/backend/src/module/parser/analyser/tokenizer/parser.py
@@ -103,7 +103,10 @@ _PV = re.compile(
 )
 _NCOP = re.compile(r"(?<![A-Za-z0-9])NCOP(?:v?(\d+))?(?![A-Za-z0-9])", re.I)
 _NCED = re.compile(r"(?<![A-Za-z0-9])NCED(?:v?(\d+))?(?![A-Za-z0-9])", re.I)
-_MOVIE_CJK = re.compile(r"劇場版?|剧场版?|電影版|电影版")
+_MOVIE_CJK = re.compile(
+    r"劇場版|剧场版|電影版|电影版|"
+    r"(?<![\u3400-\u9fff])(?:劇場|剧场)(?![\u3400-\u9fff])"
+)
 _MOVIE_ROMAJI = re.compile(r"(?<!\w)Gekijou-?ban(?!\w)", re.I)
 _MOVIE_EN = re.compile(r"(?<!\w)(?:The\s+)?Movie(?!\w)", re.I)
 
diff --git a/backend/src/test/test_parser_selector.py b/backend/src/test/test_parser_selector.py
index 0d234300..729212c2 100644
--- a/backend/src/test/test_parser_selector.py
+++ b/backend/src/test/test_parser_selector.py
@@ -95,6 +95,34 @@ def test_classic_and_preview_are_distinct_implementations() -> None:
     )
 
 
+@pytest.mark.parametrize(
+    "parse",
+    [parse_classic_release_title, selector.preview.parse_release_title],
+    ids=["classic", "preview"],
+)
+@pytest.mark.parametrize("title", ["某剧场作品", "小剧场", "某劇場作品", "小劇場"])
+def test_cjk_movie_word_inside_title_is_not_consumed(parse, title: str) -> None:
+    parsed = parse(f"[SubGroup] {title} Movie [1080p][MP4]")
+
+    assert parsed is not None
+    assert parsed.title_zh == title
+    assert parsed.media_type is MediaType.MOVIE
+
+
+@pytest.mark.parametrize(
+    "parse",
+    [parse_classic_release_title, selector.preview.parse_release_title],
+    ids=["classic", "preview"],
+)
+def test_standalone_cjk_movie_marker_is_still_consumed(parse) -> None:
+    parsed = parse("[SubGroup] 某动画 / Some Anime 剧场 [1080p][MP4]")
+
+    assert parsed is not None
+    assert parsed.title_zh == "某动画"
+    assert parsed.title_en == "Some Anime"
+    assert parsed.media_type is MediaType.MOVIE
+
+
 def test_selector_uses_classic_without_preview_fallback(
     monkeypatch: pytest.MonkeyPatch,
 ) -> None: