tcjlj/MeBox
1
0
Fork 0
mirror of https://github.com/truewhile/MeBox.git synced 2026-09-29 19:36:36 +08:00
Code Issues Packages Projects Releases Wiki Activity
361 Commits 3 Branches 247 Tags
7efa5d20b034309da904752e5de35fcce1472526
Commit Graph

220 Commits

Author SHA1 Message Date
ShukeBta 8a57410626 fix: keep explicit source category after rejected scrape 2026-06-25 04:24:35 +08:00
ShukeBta 4e00c4c14d fix: refresh stale scan-derived media metadata 2026-06-25 03:40:50 +08:00
ShukeBta 5eb0d4c8af fix: reject low-confidence automatic metadata matches 2026-06-25 02:58:07 +08:00
ShukeBta 9021007027 fix: harden playback and security edge cases 2026-06-25 02:33:30 +08:00
ShukeBta 3d3679c21c fix: prevent episode metadata from splitting series 2026-06-25 02:32:55 +08:00
ShukeBta baa4a38b0a fix: refresh emby sessions in realtime 2026-06-25 02:32:32 +08:00
ShukeBta 29f5df673c fix: refresh emby realtime user sessions 2026-06-25 00:36:41 +08:00
ShukeBta f49ef7f1df refactor: split organize source file flow 2026-06-25 00:17:09 +08:00
ShukeBta ab6b8178f4 refactor: split storage config responsibilities 2026-06-25 00:05:52 +08:00
ShukeBta 9462f3b021 refactor: split storage upload clients 2026-06-25 00:00:20 +08:00
ShukeBta 996030eae8 refactor: split local scan flow 2026-06-24 22:06:39 +08:00
ShukeBta 070fd7459e refactor: split rss subscription run flow 2026-06-24 21:58:16 +08:00
ShukeBta bbc1d542bb refactor: split cloud scan import flow 2026-06-24 21:50:47 +08:00
ShukeBta d344446406 refactor: split subscription site search completion 2026-06-24 21:42:30 +08:00
ShukeBta dd3d4c1aec fix: load real discover posters 2026-06-24 21:30:44 +08:00
ShukeBta 9bdf1a73e7 refactor: split ffmpeg transcode args 2026-06-24 20:43:36 +08:00
ShukeBta 68e6b5050b refactor: split cloud scan candidate collection 2026-06-24 20:28:01 +08:00
ShukeBta a7c8d46589 fix: show latest discover posters 2026-06-24 20:17:38 +08:00
ShukeBta e5b5ecce93 refactor: split subscription site search enqueue 2026-06-24 19:48:22 +08:00
ShukeBta 35f8a166c6 refactor: split image proxy tests 2026-06-24 19:36:08 +08:00
ShukeBta aa506d6566 fix: harden discover artwork caching 2026-06-24 19:26:32 +08:00
ShukeBta d93a28f18f fix: retry failed discover artwork 2026-06-24 17:24:15 +08:00
ShukeBta 55673b1675 refactor: split subscription run tests 2026-06-24 16:58:40 +08:00
ShukeBta 13d61200fc refactor: split scanner cloud tests 2026-06-24 16:48:20 +08:00
ShukeBta e898605f3a refactor: split emby compatibility tests 2026-06-24 16:11:44 +08:00
ShukeBta e56eed0758 fix: improve discover and manual scrape controls 2026-06-24 15:56:35 +08:00
ShukeBta c212207f74 refactor: split scraper service tests 2026-06-24 15:18:57 +08:00
ShukeBta 40cb8b07a2 refactor: split local metadata tests 2026-06-24 15:02:48 +08:00
ShukeBta 6be3409e3e fix: improve discover and series special handling 2026-06-24 14:49:26 +08:00
ShukeBta 3957349dc8 refactor: split bot feature tests 2026-06-24 14:18:30 +08:00
ShukeBta 7483b10853 refactor: split cloud provider tests 2026-06-24 14:11:21 +08:00
ShukeBta 34c9cb4244 fix: improve discover and manual scrape flows 2026-06-24 13:58:04 +08:00
ShukeBta 34396a3a8b refactor: split cloud mount modules 2026-06-24 13:23:28 +08:00
ShukeBta 7d7f2662dc refactor: split subscription candidate tests 2026-06-24 13:11:40 +08:00
ShukeBta 1b2cbabc71 refactor: split subscription planner modules 2026-06-24 13:06:39 +08:00
ShukeBta 0e79763462 refactor: split organizer scrape tests 2026-06-24 12:58:16 +08:00
ShukeBta 47cd20f2ea refactor: isolate download organize tests 2026-06-24 12:45:23 +08:00
ShukeBta aec5914846 refactor: isolate cloud auto category tests 2026-06-24 12:21:39 +08:00
ShukeBta f7e6ae7b87 refactor: split oversized service tests 2026-06-24 12:18:39 +08:00
ShukeBta 192f35d9fa refactor: split modules and harden scraping workflows 2026-06-24 11:59:18 +08:00
ShukeBta efca3cbe69 fix media playback and library workflows 2026-06-21 12:53:52 +08:00
ShukeBta 7b881d031c fix organizer diagnostics and site integrations 2026-06-21 09:44:29 +08:00
ShukeBta 6e61c8b545 feat(telegram): 订阅/下载补齐 year/rating/genres/原名/语言 富通知字段 
- 模型: Subscription/DownloadTask 新增 OriginalName/OriginalLanguage/Year/
  Rating/Genres 列(GORM AutoMigrate 自动建列, 无需手写迁移)
- enrich 链路: displayMetadata 与 DownloadTaskMeta 携带新字段, 从
  SearchExternalMedia(TMDb/豆瓣/Bangumi)结果回填并持久化; 放宽提前返回
  门槛, 已有海报但缺新字段时仍补齐
- 通知: 订阅/下载完成事件 data 注入 year/rating/genres/original_title/
  original_language, 喂给作者富模板 formatTelegramMediaNotification
- 测试: 新增端到端断言, 验证补齐字段透传进 Telegram caption
 2026-06-20 01:45:40 +08:00
ShukeBta 9ef7721527 fix(telegram): 订阅/下载通知补充媒体模板所需字段 
合并 #39 媒体通知模板后, formatTelegramMediaNotification 需要事件携带
片名/简介/外链/资源标题等字段才能渲染丰富内容。补充数据源中现成可用的字段:
- 订阅命中: title(订阅名)/overview/imdb_url(由 IMDBID 拼)/resource_title(首个命中资源)
- 下载完成: title/overview/resource_title(torrent 原始名, 供模板提取季集与版本)
缺失项(年份/评分/tmdb 链接等无现成来源)模板会自动略过, 不影响回退。
 2026-06-20 01:16:48 +08:00
yebuwudong 8e43f66513 fix(telegram): improve media notification template 2026-06-20 01:08:55 +08:00
ShukeBta dff19801b6 fix(series): 彻底修复剧集被按单集显示且无法修复 
根因(用真实数据确诊, 两个互相独立):
- Web 端拆集: episode NFO 的 <uniqueid type=tmdb> 是【单集 episode id】、
  <title> 是【单集名】, mergeEpisodeMetadata 却无条件覆盖了整剧 tm_db_id /
  original_name → 同剧每集 id/原名各不相同, 被 getSeriesKey 按 tmdb_id 拆成
  N 张单集卡(实测「遮天」90 集 = 89 个不同 tmdb_id)。
- Emby 散装单集: 电影库混入按 Season/SxxE 整理的剧集结构内容, 被判成 Episode
  却因 movie 库不聚 Series 而漏成散装单集。
- 无法修复: EnrichLibrary 只处理 pending/no_match, 跳过 matched 脏行。

修复:
- local_metadata.mergeEpisodeMetadata: 单集 NFO 的外部 id 不再写入整剧字段,
  单集名不再污染 original_name; 整剧 id/原名只认 tvshow.nfo。
- groupSeries.ts: 剧集分组改「库+路径剧名」优先(整剧目录对全剧一致), 外部 id
  仅作回退, 对存量脏数据也能合并。
- emby_compat: 电影库常规浏览新增 movieLibraryItems, 把剧集结构内容聚成 Series
  卡片与真电影并列(方案 B), 不再漏散装单集。
- cloud_path_repair: 「修复+重刮」先把脏的 matched 剧集行重置为 pending 再重刮,
  使其能被重新刮削; 结果新增 reset 计数。
- 新增 NormalizePollutedEpisodeMetadata 启动迁移(幂等): 按整剧目录聚合, 清洗
  tmdb/原名散乱的剧组并重置 pending。

测试: local_metadata/emby_compat/cloud_path_repair/episode_metadata_cleanup
新增或更新断言; 修正 scanner_cloud_test 中依赖旧错误行为的用例。
 2026-06-20 01:04:19 +08:00
ShukeBta b40e2c7280 fix: 订阅恢复补缺集/单媒体手动整理入口/媒体库排序/网盘媒体分类迁移 
- 订阅(问题1): selectPreparedSubscriptionCandidates 在本地已有部分集时不再
  无条件丢弃整季/全集包,单集候选不足时用全集包兜底;Restore 重置 total_episodes
  让下次 run 从权威元数据重算,避免被 feed 低估值锁死误判'无缺集'
- 单媒体整理(问题2): 后端已具备,新增 OrganizeMediaDialog 组件并在媒体详情页
  接入'整理入库'按钮(可选类型/二级分类/目标路径/转移方式/预览)
- 媒体库排序(问题3): ListByLibrariesFiltered 改多级排序 year/updated_at/
  created_at/id,id 作稳定 tie-breaker 消除云盘批量扫描同批时间戳的随机观感
- 网盘媒体消失(问题4): 云盘媒体 upsert 时允许 library_id 迁移到重新挂载的
  更精确分类库(本地媒体物理位置固定不迁移),修复一键挂载子目录后媒体消失
 2026-06-19 21:29:02 +08:00
ShukeBta 9b4884359b fix(scrape): 单集名不再覆盖整剧 original_name 修复合集被拆集 + 媒体库页新增单库修复重刮 
- 根因: 单集刮削把 TMDB episode.Name 写入 original_name(应为整剧原名),
  导致同剧每集 original_name 不同, 合集分组键回退到标题时被拆成多集无法合并。
  移除该覆盖, 保留单集 overview/剧照/评分/时长回填(不影响合集键)。
- 新增 RepairAndRescrapeLibrary 服务方法 + POST /admin/libraries/:id/repair-rescrape
  路由 + 媒体库页「修复+重刮本库」按钮, 可按库单独触发修复(回填占位符外部 ID)并重刮,
  既修空 ID 又借重刮覆盖此前被单集名污染的 original_name。
 2026-06-19 21:29:02 +08:00
ShukeBta 87bb94d4df feat(repair): 新增全库修复+重刮入口，剧集类路径正确归类 
- 服务层 RepairAndRescrapeAllLibraries：先从路径占位符回填缺失/错误外部ID，再逐库重刮(含 no_match 重试)
- 新增 handler 与路由 POST /api/admin/media/repair-rescrape，异步执行经 WS 推送进度
- 前端媒体库页新增「全库修复+重刮」按钮调用该接口
- groupSeries：路径落在剧集类目录(电视剧/动漫/特别篇等)时按剧集卡处理，跳转分类视图
 2026-06-19 12:23:56 +08:00
ShukeBta ce38fd257f chore: 移除源码注释与测试名中的第三方项目引用 
将 cloud/external_search/filemanager/subscription_planner 注释及若干测试
函数名中的外部项目名替换为中性表述,不改变任何逻辑与行为。
 2026-06-19 12:23:56 +08:00