恢复所有项目文件

source/.test_output/all_i_need/unfollow_mids_list.txt

@@ -0,0 +1 @@
+1044673687,1481344732,1858861103,444728505,23947287,35807625,111714204,1587138171,440798355,33291981,11914415,436700803,3493282273299102,612593877,2125857107,2000819931,507448807,505935166,14524124,385200931,1769820463,1562896062,3493285194632125,3493264275540254,479424216,604710494,1016523517,1428318343,700817047,543931674,1590538073,1574721168,432752294,3494376355400290,1795221360,4848323,495224316,3493258518858434,379247856,32360194,381653678,274928598,475656605,365212208,3546378525477862,35339643,1747335,1263990139,3493263038220393,251642119,387412319,1212367465,589747109,1025542770,23770618,3494350482836026,54091976,599449178,1715594148,3493127266503448,1767282898,487505057,630874464,1264711195,3537118481615036,319358609,518742534,385172962,4401694,474803476,525382468,3546595513600180,295993972,476819048,21435789,1725223092,2114928296,174471602,1480366563,17095888,295100453,1305776725,25694274,14797570,166828,385126080,3461582166166488,3537120815745590,489302782,73674032,1500074803,68134500,1047158092,3546571071293861,124806013,26055664,441631812,243680430,601300995,108526737,2100151539,3546603229023143,1749224369,3493133887211865,56300844,255139870,23244398,3493291869866324,3494354444355822,3546593938639500,1098004826,94577838,21849780,35105301,423319981,535023713,224560702,3546637651675315,3494361759221832,1640934198,1710911403,14342271,2031277323,603430640,3546568640694467,1741962246,1304346514,283389925,3461575868418125,3546622413768823,3494364269513335,185549749,502539494,73528331,510767506,3461579156752681,238171381,3546627212052911,448165099,1975692083,542824499,16243913,3494354016537425,316627722,1944667205,1433031509,3546387566299549,496787581,3546643550963789,382423121,600428973,430426421,325848853,735958,35162124,668794433,3546390949005555,478548163,3546672034482563,250584301,485234598,1555665460,6776617,108709998,437840703,28378491,67079745,1606682745,629101318,452161580,3493089637305282,374377163,213845897,323713206,272107494,622986240,1773278179,3546656899336980,67141499,318331,285027361,114366178,203983793,1283676771,1965933018,470624011,3546583482239276,3493281239402498,1475977561,2016676980,1209319826,1335124945,416206486,129860965,1780480185,1809567655,245645656,1937416537,1060544882,1335713025,3546617688886097,3546752326044595,3546613148551357,652060948,2116071253,97407861,3546908731639909,3546693165386233,278761367,323588182,486989780,3494353494345852,96609715,264869770,478849208,1679822121,19414347,3493127314737312,702915816,482867012,3546969421122388,3546590214097572,501642082,458165375,3546662484052067,481153145,1159873315,3546857594685834,1508100119,111900,1732848825,3546606469123022,106685726,490494088,1511660367

source/.test_output/all_i_need/unfollow_mids_list_1.txt

@@ -0,0 +1 @@
+1044673687,1481344732,1858861103,444728505,23947287,35807625,111714204,1587138171,440798355,33291981,11914415,436700803,3493282273299102,612593877,2125857107,2000819931,507448807,505935166,14524124,385200931,1769820463,1562896062,3493285194632125,3493264275540254,479424216,604710494,1016523517,1428318343,700817047,543931674,1590538073,1574721168,432752294,3494376355400290,1795221360,4848323,495224316,3493258518858434,379247856,32360194,381653678,274928598,475656605,365212208,3546378525477862,35339643,1747335,1263990139,3493263038220393,251642119,387412319,1212367465,589747109,1025542770,23770618,3494350482836026,54091976,599449178,1715594148,3493127266503448,1767282898,487505057,630874464,1264711195,3537118481615036,319358609,518742534,385172962,4401694,474803476,525382468,3546595513600180,295993972,476819048,21435789,1725223092,2114928296,174471602,1480366563,17095888,295100453,1305776725,25694274,14797570,166828,385126080,3461582166166488,3537120815745590,489302782,73674032,1500074803,68134500,1047158092,3546571071293861,124806013,26055664,441631812,243680430,601300995,108526737

source/.test_output/all_i_need/unfollow_mids_list_2.txt

@@ -0,0 +1 @@
+2100151539,3546603229023143,1749224369,3493133887211865,56300844,255139870,23244398,3493291869866324,3494354444355822,3546593938639500,1098004826,94577838,21849780,35105301,423319981,535023713,224560702,3546637651675315,3494361759221832,1640934198,1710911403,14342271,2031277323,603430640,3546568640694467,1741962246,1304346514,283389925,3461575868418125,3546622413768823,3494364269513335,185549749,502539494,73528331,510767506,3461579156752681,238171381,3546627212052911,448165099,1975692083,542824499,16243913,3494354016537425,316627722,1944667205,1433031509,3546387566299549,496787581,3546643550963789,382423121,600428973,430426421,325848853,735958,35162124,668794433,3546390949005555,478548163,3546672034482563,250584301,485234598,1555665460,6776617,108709998,437840703,28378491,67079745,1606682745,629101318,452161580,3493089637305282,374377163,213845897,323713206,272107494,622986240,1773278179,3546656899336980,67141499,318331,285027361,114366178,203983793,1283676771,1965933018,470624011,3546583482239276,3493281239402498,1475977561,2016676980,1209319826,1335124945,416206486,129860965,1780480185,1809567655,245645656,1937416537,1060544882,1335713025

source/.test_output/all_i_need/unfollow_mids_list_3.txt

@@ -0,0 +1 @@
+3546617688886097,3546752326044595,3546613148551357,652060948,2116071253,97407861,3546908731639909,3546693165386233,278761367,323588182,486989780,3494353494345852,96609715,264869770,478849208,1679822121,19414347,3493127314737312,702915816,482867012,3546969421122388,3546590214097572,501642082,458165375,3546662484052067,481153145,1159873315,3546857594685834,1508100119,111900,1732848825,3546606469123022,106685726,490494088,1511660367

source/.test_output/all_i_need/up_keep_follow_only.md

@@ -0,0 +1,126 @@
+# 保留关注UP主分析与分组建议
+
+- 生成时间: 2026-04-26 15:12:23
+- 来源文件: up_analysis_full_auto.md
+- 条目数: 88
+
+## 1. AI光影社 (mid: 1957113893)
+
+### AI分析
+
+AI光影社专注于人工智能领域的最新进展与深度思考，涵盖GPT模型迭代、AI智能体协作、科学应用及社会影响等硬核内容。其标题体现对技术本质和行业趋势的剖析，兼具前沿性和批判性。
+
+### 分组建议
+
+- 预设分组: A_硬核知识保留
+- 建议动作: 保留关注
+- 判断依据: 内容聚焦AI技术原理与科学应用，关键词'模型''科学''技术'命中硬核知识分组规则，但缺乏编程实操细节，未达更高分组标准，建议保留。
+
+## 2. DeeparchWorks (mid: 276483261)
+
+### AI分析
+
+DeeparchWorks专注于Linux内核深度技术解析，涵盖调度器、网络、文件系统等核心模块优化与问题修复。内容专业性强，每期结合具体补丁或新特性进行多层原理剖析，适合对系统底层有深入研究的开发者。
+
+### 分组建议
+
+- 预设分组: A_硬核知识保留
+- 建议动作: 保留关注
+- 判断依据: 标题均涉及Linux内核机制、性能优化与算法细节，命中'硬核知识'关键词（如调度器、零拷贝、BPF验证器），内容深度符合技术保留标准，但未达到每日必读的核心级别。
+
+## 3. Ele实验室 (mid: 481434238)
+
+### AI分析
+
+Ele实验室专注于计算机科学领域的硬核知识科普，内容涵盖向量数据库、自然语言处理、概率模拟等技术深度解析，结合历史视角与实用工具开发，适合对计算机原理和前沿技术有深入兴趣的观众。
+
+### 分组建议
+
+- 预设分组: A_硬核知识保留
+- 建议动作: 保留关注
+- 判断依据: 标题涉及向量数据库、NLP、加密技术等高密度技术主题，符合硬核知识定位，内容优质且无娱乐化倾向，值得保留。
+
+## 4. EnglishVista2023 (mid: 488218866)
+
+### AI分析
+
+UP主EnglishVista2023专注于英语词汇学习，内容涵盖雅思词汇系列（如第137-143天）和主题词汇速记（如天气预报、打印机相关）。视频标题规律性强，以教育类硬核知识为主，无娱乐或营销内容，适合长期英语学习者。
+
+### 分组建议
+
+- 预设分组: B_技能学习保留
+- 建议动作: 保留关注
+- 判断依据: 内容为系统性英语词汇教学，专业度中等且更新稳定，符合技能学习分组的定位，适合有明确学习需求的用户持续关注。
+
+## 5. Git源宝 (mid: 38061207)
+
+### AI分析
+
+Git源宝专注于AI编程与科技热点，内容涵盖DeepSeek解析、AI编程教学、AIGC率优化等硬核技术，同时涉及AI应用娱乐化观察，整体偏向实用技能与前沿资讯混合。
+
+### 分组建议
+
+- 预设分组: A_硬核知识保留
+- 建议动作: 保留关注
+- 判断依据: 标题高频命中'AI编程''教学''技术解析'等硬核关键词，但部分内容偏向娱乐观察，未达每日必读或纯干货标准，建议保留核心知识价值内容。
+
+## 6. Java面试突击-Mic (mid: 1031543543)
+
+### AI分析
+
+该UP主专注于Java面试技巧、职业发展建议及行业趋势分析，内容涵盖简历优化、面试问题解析、AI对程序员影响等，针对求职者和学习者提供实用指导。
+
+### 分组建议
+
+- 预设分组: B_技能学习保留
+- 建议动作: 保留关注
+- 判断依据: 标题高频命中'面试'、'Java'、'技能'等关键词，内容聚焦求职技能提升，符合B组定位；虽非顶级硬核知识，但对学习者有持续价值，建议保留。
+
+## 7. math也是柠檬精 (mid: 457983079)
+
+### AI分析
+
+该UP主专注于计算机科学和数学领域，内容涵盖408考试相关知识点、数据结构与算法、计算机组成原理等硬核技术主题。视频标题显示其内容具有高度专业性和学习价值，适合需要系统学习相关知识的观众。
+
+### 分组建议
+
+- 预设分组: B_技能学习保留
+- 建议动作: 保留关注
+- 判断依据: 内容聚焦技能学习，标题如'补码的顶级理解'、'红黑树'等命中B组关键词，专业性强，适合长期学习参考。
+
+## 8. NotOnlySuccess (mid: 3546647317448859)
+
+### AI分析
+
+UP主NotOnlySuccess专注于算法与数据结构深度解析，涵盖红黑树、动态规划优化、数学思维与编程结合等硬核主题。内容专业性强，适合有算法基础的观众，但缺乏日常实用技巧或娱乐性内容，定位较为垂直。
+
+### 分组建议
+
+- 预设分组: A_硬核知识保留
+- 建议动作: 保留关注
+- 判断依据: 标题高频命中'红黑树''DP优化''数学思维'等硬核关键词，符合A组规则。内容质量高但受众狭窄，适合算法爱好者长期学习，故建议保留。
+
+## 9. OK语文 (mid: 3493281113573620)
+
+### AI分析
+
+UP主OK语文专注高考语文备考，内容覆盖阅读、作文、选择题等核心题型，提供实用技巧与素材。标题高频出现“高考”“满分”“神技”等词，专业性强且稳定聚焦应试提分。
+
+### 分组建议
+
+- 预设分组: A_硬核知识保留
+- 建议动作: 保留关注
+- 判断依据: 内容垂直高考语文，专业度高且持续输出解题技巧与素材，符合硬核知识定位，适合备考学生长期参考。
+
+## 10. Python万能胶 (mid: 1577562848)
+
+### AI分析
+
+UP主Python万能胶专注于机器学习、深度学习、强化学习等实战案例与流程讲解，内容涵盖特征工程、模型调参、Tensorflow应用等硬核技术，适合编程学习者系统跟进。
+
+### 分组建议
+
+- 预设分组: AA_编程信息干货必留
+- 建议动作: 保留关注
+- 判断依据: 标题高频命中'机器学习''实战案例''特征工程'等编程干货关键词，内容垂直且实用，符合AA组核心规则，属于优质技术资源。
+