fix(indexer): keep music site category on rust parsing

app/modules/indexer/spider/__init__.py

@@ -1,24 +1,22 @@
 import datetime
 import re
 import traceback
-from typing import Any, Optional
-from typing import List
-from urllib.parse import quote, urlparse, parse_qs
+from typing import Any, List, Optional
+from urllib.parse import parse_qs, quote, urlparse
 
 from jinja2 import Template
 from pyquery import PyQuery
 
-from app.runtime.execution import run_in_threadpool
-from app.runtime.settings import get_runtime_setting
-
-from app.runtime.log import logger
-from app.schemas.types import MediaType
+from app.adapters.network.http import AsyncRequestUtils, RequestUtils
 from app.adapters.system import rust as rust_accel
-from app.adapters.network.http import RequestUtils, AsyncRequestUtils
 from app.foundation import size as size_tools
 from app.foundation import temporal as time_tools
 from app.foundation import url as url_tools
 from app.foundation.url import UrlUtils
+from app.runtime.execution import run_in_threadpool
+from app.runtime.log import logger
+from app.runtime.settings import get_runtime_setting
+from app.schemas.types import MediaType
 
 
 def select_media_categories(category: Optional[dict], mtype: Optional[MediaType]) -> list[dict]:
@@ -79,7 +77,7 @@ class SiteSpider:
     @property
     def __dir__(self):
         """拒绝外部列举 Spider 的受保护属性。"""
-        raise AttributeError(f"Cannot read protected attribute!")
+        raise AttributeError("Cannot read protected attribute!")
 
     def __init__(self,
                  indexer: dict,
@@ -810,17 +808,28 @@ class SiteSpider:
             resolved_type = self.site_media_type
         self.torrents_info['category'] = resolved_type.value
 
-    def __apply_requested_result_media_type(self, torrents: Optional[List[dict]]) -> List[dict]:
+    def __apply_result_media_type(self, torrents: Optional[List[dict]]) -> List[dict]:
         """
-        为已由站点查询条件约束类型的结果补充统一媒体类型。
+        为解析结果补充统一媒体类型。
 
-        仅在站点配置显式声明 result_media_type=requested 时生效，避免把普通混合搜索结果误分类。
+        显式请求类型始终覆盖结果；纯媒体站点在分类缺失或未知时兜底，
+        保持 Python 与 Rust 两条解析路径的分类行为一致。
         """
         results = torrents or []
-        if not self.requested_result_media_type:
+        if self.requested_result_media_type:
+            for torrent in results:
+                torrent["category"] = self.requested_result_media_type.value
+            return results
+        if not self.site_media_type:
             return results
         for torrent in results:
-            torrent["category"] = self.requested_result_media_type.value
+            if torrent.get("category") in (
+                    None,
+                    "",
+                    MediaType.UNKNOWN,
+                    MediaType.UNKNOWN.value,
+            ):
+                torrent["category"] = self.site_media_type.value
         return results
 
     def __get_subtitle_field(self, torrent: Any, field_name: str):
@@ -1117,7 +1126,7 @@ class SiteSpider:
                 result_num=self.result_num
             )
             if rust_torrents is not None:
-                return self.__apply_requested_result_media_type(rust_torrents)
+                return self.__apply_result_media_type(rust_torrents)
 
         # 清空旧结果
         self.torrents_info_array = []
@@ -1148,7 +1157,7 @@ class SiteSpider:
                     torrent_query.clear()
                     del torrent_query
             # 返回数组的副本，防止被后续清理操作影响
-            return self.__apply_requested_result_media_type(self.torrents_info_array.copy())
+            return self.__apply_result_media_type(self.torrents_info_array.copy())
         except Exception as err:
             self.is_error = True
             logger.warn(f"错误：{self.indexername} {str(err)}")

tests/test_dicmusic_indexer.py

@@ -1,5 +1,6 @@
 from typing import Any, NoReturn
 from unittest.mock import patch
+from urllib.parse import parse_qs, urlparse
 
 import pytest
 
@@ -45,6 +46,9 @@ def _dicmusic_indexer() -> dict[str, Any]:
                 "searchstr": "{keyword}",
             },
         },
+        "browse": {
+            "path": "torrents.php?searchsubmit=1&group_results=0&page={page}",
+        },
         "torrents": {
             "list": {"selector": "table#torrent_table > tbody > tr.torrent"},
             "fields": {
@@ -79,6 +83,22 @@ def _dicmusic_indexer() -> dict[str, Any]:
     }
 
 
+def test_dicmusic_browse_requests_ungrouped_torrent_rows() -> None:
+    """DIC Music 无关键词浏览也应显式关闭 Gazelle 结果分组。"""
+    spider = SiteSpider(_dicmusic_indexer(), page=2)
+
+    browse_url = spider._SiteSpider__get_search_url()
+    parsed_url = urlparse(browse_url)
+    query = parse_qs(parsed_url.query)
+
+    assert parsed_url.path == "/torrents.php"
+    assert query == {
+        "searchsubmit": ["1"],
+        "group_results": ["0"],
+        "page": ["2"],
+    }
+
+
 def test_dicmusic_search_uses_complete_music_title_in_python_fallback() -> None:
     """DIC Music Python 兜底解析应从非分组行保留完整音乐名。"""
     with patch(
@@ -91,6 +111,7 @@ def test_dicmusic_search_uses_complete_music_title_in_python_fallback() -> None:
 
     assert len(results) == 1
     assert results[0]["title"] == "BTS - ARIRANG"
+    assert results[0]["category"] == MediaType.MUSIC.value
     assert results[0]["page_url"].endswith("torrents.php?id=123&torrentid=456")
     assert results[0]["enclosure"].endswith("torrents.php?action=download&id=456")
 
@@ -116,3 +137,4 @@ def test_dicmusic_search_uses_complete_music_title_in_rust_parser(
 
     assert len(results) == 1
     assert results[0]["title"] == "BTS - ARIRANG"
+    assert results[0]["category"] == MediaType.MUSIC.value

tests/test_music_indexer_category.py

@@ -1,3 +1,5 @@
+from unittest.mock import patch
+
 from app.modules.indexer.spider import (
     SiteSpider,
     resolve_category_media_type,
@@ -63,6 +65,28 @@ def test_site_level_music_type_fills_missing_torrent_category():
     assert spider.torrents_info["category"] == MediaType.MUSIC.value
 
 
+def test_site_level_music_type_fills_missing_rust_result_category():
+    """Rust 解析未返回分类时应使用纯音乐站点的站点级类型。"""
+    indexer = {
+        "id": "music",
+        "name": "Music",
+        "domain": "https://music.example/",
+        "media_type": "music",
+        "search": {"paths": [{"path": "torrents.php"}]},
+        "torrents": {"list": {}, "fields": {}},
+    }
+
+    with patch(
+        "app.modules.indexer.spider.rust_accel.parse_indexer_torrents",
+        return_value=[{"title": "Artist - Album", "category": None}],
+    ):
+        results = SiteSpider(indexer=indexer, mtype=MediaType.MUSIC).parse(
+            "<html><body></body></html>"
+        )
+
+    assert results[0]["category"] == MediaType.MUSIC.value
+
+
 def test_requested_result_media_type_overrides_unrepresentable_site_category():
     """音乐专属查询使用非主分类筛选时，应按显式契约把结果标记为音乐。"""
     spider = SiteSpider(