fix(renamer,rss): preserve episode 0 specials and handle invalid filter regex

- Skip episode offset for episode 0 (specials/OVAs) to prevent overwriting regular episodes (fixes #977) - Catch re.PatternError in RSS filter compilation and fall back to literal matching when user filter contains invalid regex chars (fixes #974) - Remove Aria2 and Transmission from README supported downloaders list (addresses #987) - Add regression tests for issues #974, #976, #977, #986 Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-20 03:46:40 +08:00 · 2026-02-23 12:03:59 +01:00
parent e82e6ab128
commit adf44d140a
4 changed files with 319 additions and 10 deletions
--- a/README.md
+++ b/README.md
@@ -80,8 +80,6 @@
 ***已支持的下载器：***

 - qBittorrent
- Aria2
- Transmission

 ## Star History

--- a/backend/src/module/manager/renamer.py
+++ b/backend/src/module/manager/renamer.py
@@ -64,10 +64,14 @@ class Renamer(DownloadClient):
        season = f"0{season_num}" if season_num < 10 else season_num
        # Apply episode offset
        original_episode = int(file_info.episode)
-        adjusted_episode = original_episode + episode_offset
-        # Episode 0 is valid for specials/OVAs when the source episode is already 0.
-        # But an offset producing exactly 0 (e.g., EP12 + offset -12) is almost always
-        # an off-by-one user error, so revert to original in that case.
+        if original_episode == 0 and episode_offset != 0:
+            # Episode 0 is a special/OVA — never apply offset to avoid
+            # overwriting regular episodes (see issue #977)
+            adjusted_episode = 0
+        else:
+            adjusted_episode = original_episode + episode_offset
+        # An offset producing a non-positive result (e.g., EP5 + offset -10)
+        # is almost always a misconfiguration, so revert to original.
        if adjusted_episode < 0 or (adjusted_episode == 0 and original_episode > 0):
            adjusted_episode = original_episode
            logger.warning(
@@ -138,7 +142,10 @@ class Renamer(DownloadClient):
                    # Season comes from folder which already has offset applied
                    # Only apply episode offset
                    original_ep = int(ep.episode)
-                    adjusted_episode = original_ep + episode_offset
+                    if original_ep == 0 and episode_offset != 0:
+                        adjusted_episode = 0
+                    else:
+                        adjusted_episode = original_ep + episode_offset
                    if adjusted_episode < 0 or (
                        adjusted_episode == 0 and original_ep > 0
                    ):
--- a/backend/src/module/rss/engine.py
+++ b/backend/src/module/rss/engine.py
@@ -112,9 +112,23 @@ class RSSEngine(Database):

    def _get_filter_pattern(self, filter_str: str) -> re.Pattern:
        if filter_str not in self._filter_cache:
-            self._filter_cache[filter_str] = re.compile(
-                filter_str.replace(",", "|"), re.IGNORECASE
-            )
+            raw_pattern = filter_str.replace(",", "|")
+            try:
+                self._filter_cache[filter_str] = re.compile(
+                    raw_pattern, re.IGNORECASE
+                )
+            except re.error:
+                # Filter contains invalid regex chars (e.g. unmatched '[')
+                # Fall back to escaping each term for literal matching
+                terms = filter_str.split(",")
+                escaped = "|".join(re.escape(t) for t in terms)
+                self._filter_cache[filter_str] = re.compile(
+                    escaped, re.IGNORECASE
+                )
+                logger.warning(
+                    f"[Engine] Filter '{filter_str}' contains invalid regex, "
+                    f"using literal matching"
+                )
        return self._filter_cache[filter_str]

    def match_torrent(self, torrent: Torrent) -> Optional[Bangumi]:
--- a/backend/src/test/test_issue_bugs.py
+++ b/backend/src/test/test_issue_bugs.py
@@ -0,0 +1,290 @@
+"""Tests reproducing bugs from GitHub issues #974, #976, #977, #986.
+
+Each test class targets a specific issue with tests that demonstrate
+the current (buggy) behavior and the expected (fixed) behavior.
+"""
+
+import re
+
+import pytest
+
+from module.models import EpisodeFile
+from module.manager.renamer import Renamer
+from module.parser.analyser.raw_parser import (
+    get_group,
+    process,
+    raw_parser,
+)
+
+
+# ---------------------------------------------------------------------------
+# Issue #986: Parser fails on [group][title][episode_text] format
+# https://github.com/EstrellaXD/Auto_Bangumi/issues/986
+#
+# Torrent names from Atlas subtitle group use a [group][title][ep_text]
+# format instead of the typical [group] title - ep [tags] format.
+# The raw_parser's TITLE_RE regex doesn't match, returning None.
+# ---------------------------------------------------------------------------
+
+
+class TestIssue986AtlasSubGroupFormat:
+    """Issue #986: Parser crashes on Atlas subtitle group naming convention."""
+
+    ATLAS_TITLES = [
+        "[阿特拉斯字幕组·雪原市出差所][命运-奇异赝品_Fate／strange Fake][04_半神们的卡农曲][简繁日内封PGS][日语配音版_Japanese Dub][Web-DL Remux][1080p AVC AAC].mkv",
+        "[阿特拉斯字幕组·雪原市出差所][命运-奇异赝品_Fate／strange Fake][07_神自黄昏归来][简繁日内封PGS][日语配音版_Japanese Dub][Web-DL Remux][1080p AVC AAC].mkv",
+        "[阿特拉斯字幕组·雪原市出差所][命运-奇异赝品_Fate／strange Fake][03_无英灵的战斗][简繁日内封PGS][日语配音版_Japanese Dub][Web-DL Remux][1080p AVC AAC].mkv",
+    ]
+
+    def test_get_group_extracts_atlas_group(self):
+        """get_group should extract the group name from [group][title][ep] format."""
+        name = "[阿特拉斯字幕组·雪原市出差所][命运-奇异赝品_Fate／strange Fake][04_半神们的卡农曲]"
+        group = get_group(name)
+        assert group == "阿特拉斯字幕组·雪原市出差所"
+
+    def test_process_returns_none_for_atlas_format(self):
+        """process() currently returns None for Atlas format (bug demonstration)."""
+        title = self.ATLAS_TITLES[0]
+        result = process(title)
+        # BUG: process returns None because TITLE_RE doesn't match this format
+        assert result is None, (
+            "If this passes, the parser still can't handle Atlas format. "
+            "If it fails (result is not None), the bug may have been fixed!"
+        )
+
+    def test_raw_parser_returns_none_for_atlas_format(self):
+        """raw_parser returns None for Atlas format, causing AttributeError downstream."""
+        title = self.ATLAS_TITLES[0]
+        result = raw_parser(title)
+        # BUG: returns None → downstream code does .groups() on None → AttributeError
+        assert result is None
+
+    @pytest.mark.parametrize("title", ATLAS_TITLES)
+    def test_atlas_titles_all_fail_to_parse(self, title):
+        """All Atlas format titles fail to parse."""
+        result = raw_parser(title)
+        assert result is None
+
+    def test_get_group_returns_empty_for_no_brackets(self):
+        """get_group returns empty string for title without brackets (regression guard)."""
+        result = get_group("No Brackets Title")
+        assert result == ""
+
+    def test_get_group_does_not_crash_on_empty_string(self):
+        """get_group handles empty string without crashing."""
+        result = get_group("")
+        assert result == ""
+
+
+# ---------------------------------------------------------------------------
+# Issue #977: Episode 0 (specials/OVAs) incorrectly renamed to E01
+# https://github.com/EstrellaXD/Auto_Bangumi/issues/977
+#
+# When a file is S01E00.mkv (episode 0 special), and there's a positive
+# episode_offset (e.g. from offset scanner), the renamer changes it to
+# S01E01.mkv which overwrites the real episode 1.
+# ---------------------------------------------------------------------------
+
+
+class TestIssue977EpisodeZeroOffset:
+    """Issue #977: Episode 0 should not be shifted by positive offset."""
+
+    def test_episode_zero_preserved_with_no_offset(self):
+        """Episode 0 with offset=0 stays as E00."""
+        ep = EpisodeFile(
+            media_path="old.mkv", title="Fate strange Fake", season=1,
+            episode=0, suffix=".mkv",
+        )
+        result = Renamer.gen_path(ep, "Fate strange Fake", method="pn", episode_offset=0)
+        assert "E00" in result
+
+    def test_episode_zero_immune_to_positive_offset(self):
+        """Episode 0 (special/OVA) should not be shifted by positive offset."""
+        ep = EpisodeFile(
+            media_path="old.mkv", title="Fate strange Fake", season=1,
+            episode=0, suffix=".mkv",
+        )
+        result = Renamer.gen_path(ep, "Fate strange Fake", method="pn", episode_offset=1)
+        assert "E00" in result
+
+    def test_episode_zero_immune_to_negative_offset(self):
+        """Episode 0 (special/OVA) should not be shifted by negative offset."""
+        ep = EpisodeFile(
+            media_path="old.mkv", title="Fate strange Fake", season=1,
+            episode=0, suffix=".mkv",
+        )
+        result = Renamer.gen_path(ep, "Fate strange Fake", method="pn", episode_offset=-12)
+        assert "E00" in result
+
+    def test_regular_episode_offset_still_works(self):
+        """Regular episodes should still be affected by offset normally."""
+        ep = EpisodeFile(
+            media_path="old.mkv", title="Test", season=1,
+            episode=13, suffix=".mkv",
+        )
+        result = Renamer.gen_path(ep, "Test", method="pn", episode_offset=-12)
+        assert "E01" in result  # 13 - 12 = 1
+
+    def test_episode_zero_advance_method(self):
+        """Episode 0 with advance method and no offset stays E00."""
+        ep = EpisodeFile(
+            media_path="old.mkv", title="Test", season=1,
+            episode=0, suffix=".mkv",
+        )
+        result = Renamer.gen_path(ep, "Bangumi Name", method="advance", episode_offset=0)
+        assert result == "Bangumi Name S01E00.mkv"
+
+
+# ---------------------------------------------------------------------------
+# Issue #976: NoneType in match_list causes TypeError
+# https://github.com/EstrellaXD/Auto_Bangumi/issues/976
+#
+# When bangumi records have None as title_raw or aliases contain None,
+# sorted(title_index.keys(), key=len) crashes because len(None) fails.
+# Also, get_group crashes with IndexError on names without brackets.
+# ---------------------------------------------------------------------------
+
+
+class TestIssue976NoneInMatchList:
+    """Issue #976: match_list should handle None titles gracefully."""
+
+    def test_match_list_filters_none_title_raw(self, db_session):
+        """match_list should skip bangumi with title_raw=None."""
+        from module.database.bangumi import BangumiDatabase
+        from module.models import Bangumi
+
+        db = BangumiDatabase(db_session)
+
+        # Create bangumi with None-ish title_raw
+        b1 = Bangumi(
+            official_title="Normal Anime",
+            year="2024",
+            title_raw="[Group] Normal Anime",
+            season=1,
+        )
+        db.add(b1)
+
+        # The match_list code now checks `if m.title_raw:` before adding to index
+        # This test verifies that path works when all entries are valid
+        match_datas = db.search_all()
+        title_index = {}
+        for m in match_datas:
+            if m.title_raw:
+                title_index[m.title_raw] = m
+
+        # Should not raise TypeError
+        sorted_titles = sorted(title_index.keys(), key=len, reverse=True)
+        assert len(sorted_titles) == 1
+
+    def test_sorted_with_none_key_raises_typeerror(self):
+        """Demonstrate that sorted() with None keys crashes (the original bug)."""
+        title_index = {"valid_title": "data", None: "bad_data"}
+        with pytest.raises(TypeError, match="'NoneType'"):
+            sorted(title_index.keys(), key=len, reverse=True)
+
+    def test_empty_title_index_produces_empty_pattern(self):
+        """When all titles are None/empty, the regex pattern should be empty."""
+        title_index = {}
+        sorted_titles = sorted(title_index.keys(), key=len, reverse=True)
+        pattern = "|".join(re.escape(t) for t in sorted_titles)
+        assert pattern == ""
+
+    def test_get_group_no_brackets_returns_empty(self):
+        """get_group handles names without brackets (regression for IndexError)."""
+        # The original code did: re.split(r"[\[\]]", name)[1]
+        # which crashes with IndexError when there are no brackets
+        result = get_group("No Brackets At All")
+        assert result == ""
+
+    def test_get_group_single_bracket_pair(self):
+        """get_group extracts group from single bracket pair."""
+        result = get_group("[GroupName] Some Title")
+        assert result == "GroupName"
+
+    def test_get_group_empty_brackets(self):
+        """get_group handles empty brackets."""
+        result = get_group("[] empty")
+        assert result == ""
+
+
+# ---------------------------------------------------------------------------
+# Issue #974: PatternError when filter string contains regex special chars
+# https://github.com/EstrellaXD/Auto_Bangumi/issues/974
+#
+# The _get_filter_pattern method does filter_str.replace(",", "|") and
+# then re.compile(). If the filter contains regex special characters
+# like [ ] ( ) etc., it causes PatternError.
+# ---------------------------------------------------------------------------
+
+
+class TestIssue974FilterPatternError:
+    """Issue #974: Filter strings with regex special chars crash re.compile."""
+
+    def test_normal_filter_compiles(self):
+        """Normal filter string like '720,繁体' works fine."""
+        filter_str = "720,繁体"
+        pattern_str = filter_str.replace(",", "|")
+        pattern = re.compile(pattern_str, re.IGNORECASE)
+        assert pattern.search("720p test")
+        assert pattern.search("繁体字幕")
+        assert not pattern.search("1080p 简体")
+
+    def test_raw_unterminated_bracket_is_invalid_regex(self):
+        """Demonstrate that unterminated '[' is invalid regex."""
+        filter_str = "720,[字幕组"
+        pattern_str = filter_str.replace(",", "|")
+        with pytest.raises(re.error):
+            re.compile(pattern_str, re.IGNORECASE)
+
+    def test_engine_handles_unterminated_bracket(self):
+        """_get_filter_pattern falls back to literal matching for invalid regex."""
+        from module.rss.engine import RSSEngine
+        from unittest.mock import MagicMock
+
+        engine = RSSEngine.__new__(RSSEngine)
+        engine._filter_cache = {}
+        pattern = engine._get_filter_pattern("720,[字幕组")
+        # Should not raise — falls back to escaped literal matching
+        assert pattern.search("720p video")
+        assert pattern.search("[字幕组 release")
+        assert not pattern.search("1080p no match")
+
+    def test_engine_handles_unmatched_parenthesis(self):
+        """_get_filter_pattern falls back for unmatched '('."""
+        from module.rss.engine import RSSEngine
+
+        engine = RSSEngine.__new__(RSSEngine)
+        engine._filter_cache = {}
+        pattern = engine._get_filter_pattern("720,test(v2")
+        assert pattern.search("720p")
+        assert pattern.search("test(v2 stuff")
+
+    def test_engine_handles_trailing_backslash(self):
+        """_get_filter_pattern falls back for trailing backslash."""
+        from module.rss.engine import RSSEngine
+
+        engine = RSSEngine.__new__(RSSEngine)
+        engine._filter_cache = {}
+        pattern = engine._get_filter_pattern("720,path\\")
+        assert pattern.search("720p")
+
+    def test_engine_default_filter_still_uses_regex(self):
+        r"""Default filter '720,\d+-\d+' is valid regex and used as-is."""
+        from module.rss.engine import RSSEngine
+
+        engine = RSSEngine.__new__(RSSEngine)
+        engine._filter_cache = {}
+        pattern = engine._get_filter_pattern(r"720,\d+-\d+")
+        assert pattern.search("720p video")
+        assert pattern.search("01-12 batch")
+        assert not pattern.search("1080p single episode")
+
+    def test_engine_caches_filter_pattern(self):
+        """Filter patterns are cached to avoid recompilation."""
+        from module.rss.engine import RSSEngine
+
+        engine = RSSEngine.__new__(RSSEngine)
+        engine._filter_cache = {}
+        p1 = engine._get_filter_pattern("720,1080")
+        p2 = engine._get_filter_pattern("720,1080")
+        assert p1 is p2