feat(monitor): 抖音 Web 接口客户端 —— 绕开爬虫子进程，直接发 HTTP

移植自 mac-agent-os 的 mediacrawler_adapter：不起子进程、不开页面，用浏览器里那份
登录态直接调抖音 Web 接口。产物键名照抄 store/douyin，所以 ingest 那条链路一个字不用改。

**目前能用的（真环境实测，非推断）**：

    profile/other : 200, 7075 字节  —— 博主主页指标（粉丝/获赞/作品数/昵称）
    aweme/detail  : 200, 45425 字节 —— 单条作品详情（含点赞/评论/收藏/分享）

**目前不能用的：作品列表 `aweme/post`。** 两个互相独立的原因：
  1. 这个接口被抖音单独升级成了真校验：不带 x-tt-argus 回 403「Uifid Not Found」，
     带上 dummy 值回 200 + **空 body**。也就是说「头在不在」骗得过，「真校验」过不了。
     同一套头打 profile/other 和 aweme/detail 都是通的 —— 抖音是挑着接口加保护的，
     挑中的恰好是「批量拉作品列表」这个最敏感的动作。
  2. 改走页面截获也不行：CDP 浏览器打开博主主页会落到「验证码中间页」（当天大量探测的
     代价，过几小时要重测）。

  所以现在的边界是：**已知作品的指标刷新能做，自动发现新作品做不了**。

**排查中控住变量后得到的两条事实**（都写进注释了）：
  · `Accept` / `Accept-Language` / `Referer` 才是主页接口能返回真数据的原因 —— 只有
    UA+client hints+Cookie 时是 200 但仅 121 字节的空壳，补上这三个头变 7074 字节。
    （我先前猜的 sec-ch-ua 不是关键。）
  · 因此 UA 与 client hints 必须**成套地取自同一个浏览器**，所以 BrowserIdentity 一次
    从 CDP 取齐 cookie + UA + hints，而不是各自写死。

「200 + 空 body 必须当场报错」也是刻意写死的：放过去它会在下游变成「这个博主没作品」，
把一次失败伪装成一条正常结果 —— 爬虫那条路正是这么栽的，还被翻译成「账号被封」。

测试 +11：cookie 解析、请求头成套性（含 uifid 缺失/回退）、产物键名与 store 对齐、
以及 _get 的三条失败路径（空 body / 403 带网关原话 / 正常返回）。

mac-agent-os-main/05_tools/07_matrix/scripts/nurture_daily.py

@@ -0,0 +1,311 @@
+#!/usr/bin/env python3
+"""
+日常养号调度器 v1.0
+
+规则:
+  1. 最多 3 个浏览器并行，每个间隔 15 秒启动
+  2. 步骤顺序每次随机打乱
+  3. 每身份每平台约运行 10 分钟
+  4. 身份完成后休息 30 秒
+  5. 动作池: 浏览/搜索/点赞/收藏/评论/关注
+
+用法:
+  python nurture_daily.py                     # 所有身份按规则开始养号
+  python nurture_daily.py --dry-run           # 只看计划，不执行
+  python nurture_daily.py --phone 185xxx      # 只养指定手机号
+"""
+
+import argparse, asyncio, json, random, subprocess, sys, time
+from datetime import datetime
+from pathlib import Path
+
+TOOL_DIR = Path(__file__).resolve().parent.parent
+SCRIPTS_DIR = TOOL_DIR / "scripts"
+BP_DIR = TOOL_DIR / "blueprints"
+from matrix_mgmt import AGENT_LOCAL
+LOCAL_ROOT = AGENT_LOCAL / "tools" / "matrix"
+CONFIG_DIR = LOCAL_ROOT / "config"
+sys.path.insert(0, str(SCRIPTS_DIR))
+
+# ── 配置 ──
+MAX_BROWSERS = 3          # 最大并发浏览器
+STAGGER_SECS = 15         # 启动间隔
+SESSION_MINS = 10         # 每身份每平台预计分钟
+REST_SECS = 30            # 身份间休息
+WATCH_MIN = 6             # 浏览最小时长(秒)
+WATCH_MAX = 25            # 浏览最大时长(秒)
+GAUSS_JITTER = 0.3        # 高斯抖动系数
+
+# ── 搜索关键词池（每次随机选） ──
+KEYWORDS_DOUYIN = [
+    "美食推荐", "旅行日记", "搞笑视频", "电影解说", "宠物日常",
+    "穿搭分享", "美妆教程", "健身打卡", "音乐现场", "游戏解说",
+    "科技数码", "汽车评测", "家居装修", "户外运动", "亲子日常",
+    "职场干货", "学习技巧", "手工制作", "摄影教程", "美食探店",
+]
+
+KEYWORDS_XHS = [
+    "穿搭推荐", "美妆教程", "旅行攻略", "美食打卡", "家居好物",
+    "护肤心得", "读书笔记", "运动打卡", "育儿经验", "摄影技巧",
+]
+
+
+def gaussian_sleep(base):
+    """base 秒 ±30% 高斯抖动"""
+    delay = base * (1 + random.uniform(-GAUSS_JITTER, GAUSS_JITTER))
+    return max(2, delay)
+
+
+def load_accounts():
+    """从 registry + override 读取本机所有账号（跨机兼容）"""
+    from matrix_mgmt import MatrixManager
+    mgr = MatrixManager()
+    all_accts = mgr.list_accounts()
+    return [a for a in all_accts if a.get("is_local") and a.get("enabled", True)]
+
+
+def load_blueprints():
+    """加载所有蓝图"""
+    bps = {}
+    for f in sorted(BP_DIR.glob("*.json")):
+        try:
+            bp = json.loads(f.read_text())
+            name = bp.get("name", bp.get("id", f.stem))
+            platform = bp.get("platform", bp.get("type", "douyin"))
+            steps = bp.get("steps", [])
+            if steps:
+                bps[name] = {"name": name, "platform": platform, "steps": steps}
+        except: pass
+    return bps
+
+
+def group_by_identity(accounts):
+    """按手机号分组"""
+    groups = {}
+    for a in accounts:
+        phone = a.get("phone", "")
+        if not phone:
+            continue
+        if phone not in groups:
+            groups[phone] = {
+                "phone": phone,
+                "display_name": a.get("display_name", phone),
+                "accounts": [],
+            }
+        groups[phone]["accounts"].append(a)
+    return list(groups.values())
+
+
+def shuffle_steps(steps, platform="douyin"):
+    """随机打乱步骤顺序 + 关键词替换 + 时间抖动"""
+    shuffled = steps.copy()
+    random.shuffle(shuffled)
+
+    # 替换 @keyword 和 @corpus
+    keywords = KEYWORDS_DOUYIN if platform == "douyin" else KEYWORDS_XHS
+    for s in shuffled:
+        args = s.get("args", {})
+        for k, v in args.items():
+            if isinstance(v, str) and "@keyword" in v:
+                args[k] = random.choice(keywords)
+            elif isinstance(v, str) and "@corpus" in v:
+                # 评论语料
+                texts = [
+                    "这个太棒了！", "学到了学到了", "收藏了～",
+                    "好厉害！", "这是什么神仙", "太实用了吧",
+                    "爱了爱了", "求教程！", "这个真的绝",
+                    "已收藏慢慢看", "讲得太清楚了", "干货满满",
+                ]
+                args[k] = random.choice(texts)
+
+        # 等待时长加抖动
+        if "seconds" in args and args["seconds"] and isinstance(args["seconds"], (int, float)):
+            args["seconds"] = round(gaussian_sleep(args["seconds"]), 1)
+
+    # 控制步数在 15-30 步之间（保证约10分钟）
+    target_steps = random.randint(15, 30)
+    if len(shuffled) > target_steps:
+        shuffled = shuffled[:target_steps]
+
+    return shuffled
+
+
+def pick_blueprint(blueprints, platform):
+    """为平台选一个合适的蓝图"""
+    candidates = [b for b in blueprints.values() if b["platform"] == platform]
+    if not candidates:
+        candidates = [b for b in blueprints.values()]  # 不限平台
+    if not candidates:
+        return None
+    return random.choice(candidates)
+
+
+def build_nurture_command(phone, display_name, account, blueprint_name, rounds=1):
+    """构建养号命令 — 用 mc run（新引擎，跨机兼容）"""
+    mc_path = TOOL_DIR / "mc"
+    return [
+        "bash", str(mc_path), "run",
+        "--accounts", account["id"],
+        "--blueprints", blueprint_name,
+        "--rounds", str(rounds),
+        "--mix",
+        "--interval", "45-90",
+    ]
+
+
+async def run_nurture_session(identity, blueprints, session_id):
+    """运行一个身份的养号"""
+    phone = identity["phone"]
+    name = identity["display_name"]
+    accounts = identity["accounts"]
+    results = []
+
+    print(f"\n{'='*55}")
+    print(f" 📱 [{name}]({phone}) 开始养号")
+    print(f" {'='*55}")
+
+    for acct in accounts:
+        platform = acct.get("platform", "douyin")
+        bp = pick_blueprint(blueprints, platform)
+        if not bp:
+            print(f"   ⚠️ {platform} 无可用蓝图，跳过")
+            continue
+
+        # 选蓝图并随机打乱步骤
+        shuffled = shuffle_steps(bp["steps"], platform)
+        bp_name = bp["name"]
+
+        # 预估时长
+        est_secs = len(shuffled) * 20 + 30  # 每步~20s + 缓冲
+        est_min = round(est_secs / 60, 1)
+
+        print(f"   🎯 {acct['id']} ({platform})")
+        print(f"      蓝图: {bp_name} → 打乱后 {len(shuffled)} 步")
+        print(f"      预计: ~{est_min}分钟")
+
+        results.append({
+            "account": acct["id"],
+            "platform": platform,
+            "blueprint": bp_name,
+            "steps": len(shuffled),
+            "estimated_min": est_min,
+        })
+
+        # 实际执行（通过 yanghao_runner）
+        cmd = build_nurture_command(phone, name, acct, bp_name, rounds=random.randint(1, 3))
+        print(f"      执行: {' '.join(cmd[:6])}...")
+        try:
+            p = subprocess.run(cmd, capture_output=True, text=True, timeout=900)
+            results[-1]["returncode"] = p.returncode
+            results[-1]["stdout_len"] = len(p.stdout)
+        except subprocess.TimeoutExpired:
+            print(f"       ⏰ {acct['id']} 超时")
+            results[-1]["returncode"] = -1
+        except Exception as e:
+            print(f"       ❌ {e}")
+            results[-1]["error"] = str(e)
+
+    return results
+
+
+async def main():
+    parser = argparse.ArgumentParser(description="日常养号调度器")
+    parser.add_argument("--dry-run", action="store_true", help="只看计划，不执行")
+    parser.add_argument("--phone", default="", help="只养指定手机号")
+    args = parser.parse_args()
+
+    print(f"\n{'='*60}")
+    print(f" 📊 日常养号调度器 v1.0")
+    print(f" {'='*60}")
+    print(f" 规则:")
+    print(f"   - 最多 {MAX_BROWSERS} 个浏览器并行")
+    print(f"   - 每个间隔 {STAGGER_SECS} 秒启动")
+    print(f"   - 步骤顺序随机打乱")
+    print(f"   - 每身份约 {SESSION_MINS} 分钟")
+    if args.dry_run:
+        print(f"   - 🔍 仅查看计划 (--dry-run)")
+
+    accounts = load_accounts()
+    blueprints = load_blueprints()
+    groups = group_by_identity(accounts)
+
+    if args.phone:
+        groups = [g for g in groups if g["phone"] == args.phone]
+
+    print(f"\n 📋 账号: {len(accounts)} | 蓝图: {len(blueprints)} | 身份: {len(groups)}")
+
+    if not groups:
+        print(" ❌ 没有匹配的身份")
+        return
+
+    # 排期表
+    print(f"\n{'='*60}")
+    print(f" 📅 排期表")
+    print(f" {'='*60}")
+
+    total_est_min = 0
+    for i, g in enumerate(groups):
+        batch = i // MAX_BROWSERS + 1
+        pos = i % MAX_BROWSERS
+        start_delay = pos * STAGGER_SECS + (batch - 1) * (SESSION_MINS * 60 + REST_SECS * MAX_BROWSERS)
+        start_min = round(start_delay / 60, 1)
+
+        plats = ", ".join(a["platform"] for a in g["accounts"])
+        bps = []
+        for a in g["accounts"]:
+            bp = pick_blueprint(blueprints, a["platform"])
+            bps.append(bp["name"] if bp else "无蓝图")
+
+        est = SESSION_MINS
+        total_est_min += est
+
+        print(f" 批次{batch}#{pos+1} | +{start_min}min | {g['display_name']:16s} | {plats:20s} | ~{est}min")
+
+    total_min = round(total_est_min / MAX_BROWSERS + (len(groups) / MAX_BROWSERS) * 0.5, 1)
+    print(f"\n 📊 预估总耗时: ~{total_min} 分钟 ({round(total_min/60,1)} 小时)")
+    print(f"    身份数: {len(groups)} | 批次: {(len(groups)+MAX_BROWSERS-1)//MAX_BROWSERS}")
+
+    if args.dry_run:
+        print(f"\n 🔍 这是计划预览，加 --dry-run 去掉就是实际执行")
+        return
+
+    # 实际执行
+    print(f"\n{'='*60}")
+    print(f" 🚀 开始执行 ({datetime.now().strftime('%H:%M')})")
+    print(f" {'='*60}")
+
+    all_results = []
+    batch_num = 0
+
+    for i in range(0, len(groups), MAX_BROWSERS):
+        batch = groups[i:i + MAX_BROWSERS]
+        batch_num += 1
+
+        print(f"\n--- 批次 {batch_num} ({len(batch)} 个身份) ---")
+
+        # 用 asyncio.gather 错开启动
+        async def launch_with_stagger(idx, g):
+            if idx > 0:
+                await asyncio.sleep(STAGGER_SECS * idx)
+            return await run_nurture_session(g, blueprints, f"b{batch_num}p{idx+1}")
+
+        tasks = [launch_with_stagger(idx, g) for idx, g in enumerate(batch)]
+        batch_results = await asyncio.gather(*tasks)
+        all_results.extend(batch_results)
+
+        # 批次间休息
+        if i + MAX_BROWSERS < len(groups):
+            print(f"\n ⏸️  批次 {batch_num} 完成，休息 {REST_SECS}s...")
+            await asyncio.sleep(REST_SECS)
+
+    # 汇总
+    print(f"\n{'='*60}")
+    print(f" ✅ 全部完成! ({datetime.now().strftime('%H:%M')})")
+    print(f" {'='*60}")
+    ok = sum(1 for r in all_results for rr in r if rr.get("returncode") == 0)
+    total = sum(len(r) for r in all_results)
+    print(f" 成功: {ok}/{total}")
+
+
+if __name__ == "__main__":
+    asyncio.run(main())