first_test

Co-authored-by: Copilot <copilot@github.com>

source/run_pipeline.py

@@ -0,0 +1,163 @@
+#!/usr/bin/env python3
+"""One-command pipeline: fetch titles -> batch analyze -> outputs.
+
+Pipeline outputs:
+1) source/output/reports/up_titles_report.md
+2) source/output/reports/up_analysis_full_auto.md
+3) source/output/reports/up_keep_follow_only.md
+4) source/output/uids/unfollow_mids_list.txt (+ split files)
+"""
+
+from __future__ import annotations
+
+import argparse
+import subprocess
+import sys
+from pathlib import Path
+
+
+def parse_args() -> argparse.Namespace:
+    parser = argparse.ArgumentParser(description="一键运行完整功能链")
+    parser.add_argument(
+        "--input-json",
+        default="source/resources/export_uids.json",
+        help="UP资源文件路径，默认: source/resources/export_uids.json",
+    )
+    parser.add_argument(
+        "--titles-report",
+        default="source/output/reports/up_titles_report.md",
+        help="标题抓取报告输出路径",
+    )
+    parser.add_argument(
+        "--analysis-report",
+        default="source/output/reports/up_analysis_full_auto.md",
+        help="分批分析报告输出路径",
+    )
+    parser.add_argument(
+        "--keep-report",
+        default="source/output/reports/up_keep_follow_only.md",
+        help="保留关注报告输出路径",
+    )
+    parser.add_argument(
+        "--unfollow-uids",
+        default="source/output/uids/unfollow_mids_list.txt",
+        help="取关UID输出路径",
+    )
+    parser.add_argument("--titles-per-up", type=int, default=10, help="每个UP抓取标题数量")
+    parser.add_argument("--batch-size", type=int, default=20, help="分批分析每批条数")
+    parser.add_argument("--workers", type=int, default=6, help="并发请求数")
+    parser.add_argument("--max-retries", type=int, default=2, help="单条分析重试次数")
+    parser.add_argument("--request-timeout", type=float, default=60.0, help="单次请求超时")
+    parser.add_argument("--split-size", type=int, default=100, help="取关UID拆分分组大小")
+    parser.add_argument("--sleep-seconds", type=float, default=0.0, help="任务间隔秒数")
+    parser.add_argument("--retry-times", type=int, default=3, help="抓取重试次数")
+    parser.add_argument("--fetch-mode", choices=["auto", "api", "html"], default="auto", help="标题抓取模式")
+    parser.add_argument("--only-tag", default="", help="可选：仅处理包含该标签的UP")
+    parser.add_argument("--max-ups", type=int, default=0, help="可选：限制处理UP数量")
+    parser.add_argument("--bili-cookie", default="", help="可选：运行时传入B站Cookie")
+    parser.add_argument("--skip-fetch", action="store_true", help="跳过抓取阶段，直接使用已有标题报告")
+    parser.add_argument("--skip-analyze", action="store_true", help="跳过分析阶段，直接做产物提取")
+    parser.add_argument("--python", default=sys.executable, help="指定Python解释器")
+    return parser.parse_args()
+
+
+def run_cmd(cmd: list[str], title: str) -> None:
+    print(f"\n=== {title} ===")
+    print("$", " ".join(cmd))
+    subprocess.run(cmd, check=True)
+
+
+def main() -> int:
+    args = parse_args()
+
+    for p in [
+        Path(args.titles_report).parent,
+        Path(args.analysis_report).parent,
+        Path(args.keep_report).parent,
+        Path(args.unfollow_uids).parent,
+    ]:
+        p.mkdir(parents=True, exist_ok=True)
+
+    if not args.skip_fetch:
+        fetch_cmd = [
+            args.python,
+            "source/analyze_up_content.py",
+            "--input",
+            args.input_json,
+            "--output",
+            args.titles_report,
+            "--titles-per-up",
+            str(max(1, args.titles_per_up)),
+            "--retry-times",
+            str(max(1, args.retry_times)),
+            "--fetch-mode",
+            args.fetch_mode,
+            "--sleep-seconds",
+            str(max(0.0, args.sleep_seconds)),
+            "--skip-ai",
+        ]
+        if args.only_tag:
+            fetch_cmd += ["--only-tag", args.only_tag]
+        if args.max_ups > 0:
+            fetch_cmd += ["--max-ups", str(args.max_ups)]
+        if args.bili_cookie:
+            fetch_cmd += ["--bili-cookie", args.bili_cookie]
+
+        run_cmd(fetch_cmd, "步骤1/4 抓取视频标题")
+
+    if not args.skip_analyze:
+        analyze_cmd = [
+            args.python,
+            "source/batch_ai_summary_from_report.py",
+            "--input-report",
+            args.titles_report,
+            "--output-report",
+            args.analysis_report,
+            "--batch-size",
+            str(max(1, args.batch_size)),
+            "--run-all-batches",
+            "--workers",
+            str(max(1, args.workers)),
+            "--max-retries",
+            str(max(1, args.max_retries)),
+            "--request-timeout",
+            str(max(1.0, args.request_timeout)),
+            "--sleep-seconds",
+            str(max(0.0, args.sleep_seconds)),
+        ]
+        run_cmd(analyze_cmd, "步骤2/4 分批AI分析")
+
+    keep_cmd = [
+        args.python,
+        "source/extract_keep_follow_doc.py",
+        "--input-report",
+        args.analysis_report,
+        "--output-report",
+        args.keep_report,
+    ]
+    run_cmd(keep_cmd, "步骤3/4 生成保留关注报告")
+
+    uid_cmd = [
+        args.python,
+        "source/extract_unfollow_list.py",
+        "--input-report",
+        args.analysis_report,
+        "--output-csv",
+        args.unfollow_uids,
+        "--format",
+        "mid-only",
+        "--split-size",
+        str(max(0, args.split_size)),
+    ]
+    run_cmd(uid_cmd, "步骤4/4 生成取关UID列表")
+
+    print("\n流水线完成。")
+    print(f"- 标题报告: {args.titles_report}")
+    print(f"- 分析报告: {args.analysis_report}")
+    print(f"- 保留报告: {args.keep_report}")
+    print(f"- 取关UID: {args.unfollow_uids}")
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())