116 lines
3.1 KiB
Markdown
116 lines
3.1 KiB
Markdown
# B站关注清理工具 - Scripts 版
|
||
|
||
> 一键命令运行全流程：`python source/scripts/run_pipeline.py`
|
||
|
||
python source/scripts/run_pipeline.py --input-json source/resources/export_uids_test5.json
|
||
|
||
本工具包含7个步骤的完整流水线：
|
||
|
||
1. 抓取视频标题
|
||
2. 分批AI分析
|
||
3. 生成保留关注报告
|
||
4. 生成取关UID列表
|
||
5. 按首字母排序
|
||
6. 提取分组信息
|
||
7. 删除最近10条标题
|
||
|
||
## 快速开始
|
||
|
||
```powershell
|
||
# 完整流程（推荐）
|
||
python source/scripts/run_pipeline.py
|
||
|
||
# 速度优先
|
||
python source/scripts/run_pipeline.py --workers 8 --batch-size 30 --sleep-seconds 0
|
||
|
||
# 试跑30个UP
|
||
python source/scripts/run_pipeline.py --max-ups 30
|
||
|
||
# 跳过抓取，使用已有标题报告
|
||
python source/scripts/run_pipeline.py --skip-fetch
|
||
|
||
# 跳过分析，仅生成产物
|
||
python source/scripts/run_pipeline.py --skip-analyze
|
||
|
||
# 跳过排序/分组/删除
|
||
python source/scripts/run_pipeline.py --skip-sort --skip-group --skip-remove
|
||
```
|
||
|
||
## 输出文件
|
||
|
||
| 文件 | 说明 |
|
||
|------|------|
|
||
| `source/output/reports/1_up_titles_report.md` | 标题抓取报告 |
|
||
| `source/output/reports/2_up_analysis_full_auto.md` | AI分析报告（完整） |
|
||
| `source/output/reports/3_up_keep_follow_only.md` | 保留关注报告 |
|
||
| `source/output/uids/4_unfollow_mids_list.txt` | 取关UID列表 |
|
||
| `source/output/reports/5_sorted_up_analysis.md` | 按首字母排序报告 |
|
||
| `source/output/reports/6_group_info.md` | 提取分组信息报告 |
|
||
| `source/output/reports/7_no_titles.md` | 最终报告（删除最近10条） |
|
||
|
||
## 常用参数
|
||
|
||
| 参数 | 默认值 | 说明 |
|
||
|------|--------|------|
|
||
| `--workers` | 6 | 并发请求数 |
|
||
| `--batch-size` | 20 | 每批分析条数 |
|
||
| `--max-ups` | 0（全部） | 限制处理UP数量 |
|
||
| `--split-size` | 100 | UID拆分大小 |
|
||
| `--sleep-seconds` | 0 | 任务间隔秒数 |
|
||
|
||
### 跳过参数
|
||
|
||
| 参数 | 说明 |
|
||
|------|------|
|
||
| `--skip-fetch` | 跳过抓取阶段 |
|
||
| `--skip-analyze` | 跳过分析阶段 |
|
||
| `--skip-sort` | 跳过排序阶段 |
|
||
| `--skip-group` | 跳过提取分组阶段 |
|
||
| `--skip-remove` | 跳过删除最近10条阶段 |
|
||
|
||
## 分步执行
|
||
|
||
### 步骤1：抓取标题
|
||
```powershell
|
||
python source/scripts/analyze_up_content.py --skip-ai
|
||
```
|
||
|
||
### 步骤2：分批AI分析
|
||
```powershell
|
||
python source/scripts/batch_ai_summary_from_report.py --run-all-batches
|
||
```
|
||
|
||
### 步骤3：生成保留关注报告
|
||
```powershell
|
||
python source/scripts/extract_keep_follow_doc.py
|
||
```
|
||
|
||
### 步骤4：生成取关UID
|
||
```powershell
|
||
python source/scripts/extract_unfollow_list.py --format mid-only --split-size 100
|
||
```
|
||
|
||
### 步骤5：按首字母排序
|
||
```powershell
|
||
python source/scripts/sort_up_main.py
|
||
```
|
||
|
||
### 步骤6：提取分组信息
|
||
```powershell
|
||
python source/scripts/extract_group_info.py
|
||
```
|
||
|
||
### 步骤7：删除最近10条标题
|
||
```powershell
|
||
python source/scripts/remove_10content.py
|
||
```
|
||
|
||
## 先配置API
|
||
|
||
编辑 [source/scripts/analyze_up_content.py](source/scripts/analyze_up_content.py) 顶部配置：
|
||
|
||
```python
|
||
VOLCENGINE_API_KEY = "你的火山引擎API Key"
|
||
VOLCENGINE_MODEL = "deepseek-v3-1-terminus"
|
||
VOLCENGINE_BASE_URL = "https://ark.cn-beijing.volces.com/api/v3"
|
||
``` |