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)。
This commit is contained in:
@@ -0,0 +1,320 @@
|
||||
#!/usr/bin/env python3
|
||||
"""
|
||||
sync_manager.py — 知识库备份、导出、同步状态检查
|
||||
|
||||
模式:
|
||||
backup — 全量备份 03_knowledge/ + 04_memory/long_term/ 到 memory_backup/
|
||||
export — 导出可分发知识包(可按领域/标签过滤)
|
||||
status — 检查坚果云同步路径状态
|
||||
|
||||
使用:
|
||||
python3 sync_manager.py --mode backup
|
||||
python3 sync_manager.py --mode backup --keep 5
|
||||
python3 sync_manager.py --mode export --domain 10_concepts 20_methods
|
||||
python3 sync_manager.py --mode export --tag AI工具
|
||||
python3 sync_manager.py --mode status
|
||||
python3 sync_manager.py --mode backup --dry-run
|
||||
"""
|
||||
|
||||
import argparse
|
||||
import os
|
||||
import re
|
||||
import shutil
|
||||
import sys
|
||||
import tarfile
|
||||
from datetime import datetime
|
||||
from pathlib import Path
|
||||
|
||||
|
||||
# ── 配置 ──────────────────────────────────────────────────────────────────────
|
||||
|
||||
BACKUP_SOURCES = [
|
||||
"03_knowledge",
|
||||
"04_memory/long_term",
|
||||
]
|
||||
|
||||
KNOWLEDGE_DOMAINS = [
|
||||
"10_concepts", "20_methods", "30_facts",
|
||||
"40_references", "50_resources", "60_opinions",
|
||||
]
|
||||
|
||||
DEFAULT_NUTSTORE_PATH = "~/NutstoreCloudBridge"
|
||||
DEFAULT_KEEP = 10
|
||||
|
||||
|
||||
# ── 参数解析 ───────────────────────────────────────────────────────────────────
|
||||
|
||||
def parse_args():
|
||||
parser = argparse.ArgumentParser(description="AgentOS 同步管理器")
|
||||
parser.add_argument("--root", default="~/workbuddy-agent-os/agent-sync", help="agent-os 根目录")
|
||||
parser.add_argument("--mode", choices=["backup", "export", "status"], required=True,
|
||||
help="运行模式")
|
||||
parser.add_argument("--keep", type=int, default=DEFAULT_KEEP,
|
||||
help="备份模式:保留最近N份(默认10)")
|
||||
parser.add_argument("--domain", nargs="+", default=None,
|
||||
help="导出模式:只导出指定领域目录(如 10_concepts 20_methods)")
|
||||
parser.add_argument("--tag", default=None,
|
||||
help="导出模式:只导出含有指定 tag 的文件")
|
||||
parser.add_argument("--nutstore", default=DEFAULT_NUTSTORE_PATH,
|
||||
help="坚果云同步路径(用于 status 模式)")
|
||||
parser.add_argument("--dry-run", action="store_true", help="模拟运行,不实际写入")
|
||||
return parser.parse_args()
|
||||
|
||||
|
||||
# ── 工具函数 ───────────────────────────────────────────────────────────────────
|
||||
|
||||
def ensure_dir(path: Path, dry_run: bool = False):
|
||||
if not dry_run:
|
||||
path.mkdir(parents=True, exist_ok=True)
|
||||
|
||||
|
||||
def format_size(size_bytes: int) -> str:
|
||||
for unit in ["B", "KB", "MB", "GB"]:
|
||||
if size_bytes < 1024:
|
||||
return f"{size_bytes:.1f} {unit}"
|
||||
size_bytes //= 1024
|
||||
return f"{size_bytes:.1f} TB"
|
||||
|
||||
|
||||
def get_md_tag(file_path: Path) -> list:
|
||||
"""从 frontmatter 提取 tags 列表"""
|
||||
try:
|
||||
content = file_path.read_text(encoding="utf-8", errors="ignore")
|
||||
if not content.startswith("---"):
|
||||
return []
|
||||
end = content.find("\n---", 3)
|
||||
if end == -1:
|
||||
return []
|
||||
frontmatter = content[3:end]
|
||||
# 匹配 tags: [a, b] 或多行列表
|
||||
m = re.search(r'tags\s*:\s*\[([^\]]+)\]', frontmatter)
|
||||
if m:
|
||||
return [t.strip().strip('"\'') for t in m.group(1).split(",")]
|
||||
# 多行格式
|
||||
tags = []
|
||||
in_tags = False
|
||||
for line in frontmatter.splitlines():
|
||||
if re.match(r'\s*tags\s*:', line):
|
||||
in_tags = True
|
||||
continue
|
||||
if in_tags:
|
||||
if line.strip().startswith("- "):
|
||||
tags.append(line.strip()[2:].strip())
|
||||
elif line.strip() and not line.startswith(" "):
|
||||
break
|
||||
return tags
|
||||
except Exception:
|
||||
return []
|
||||
|
||||
|
||||
# ── backup 模式 ────────────────────────────────────────────────────────────────
|
||||
|
||||
def run_backup(root: Path, keep: int, dry_run: bool):
|
||||
backup_dir = root / "04_memory" / "memory_backup"
|
||||
ensure_dir(backup_dir, dry_run)
|
||||
|
||||
now = datetime.now()
|
||||
filename = f"kb_backup_{now.strftime('%Y-%m-%d_%H%M')}.tar.gz"
|
||||
output_path = backup_dir / filename
|
||||
|
||||
sources = []
|
||||
for src in BACKUP_SOURCES:
|
||||
p = root / src
|
||||
if p.exists():
|
||||
sources.append(p)
|
||||
else:
|
||||
print(f" [跳过] {p} 不存在")
|
||||
|
||||
if not sources:
|
||||
print("[ERROR] 没有找到任何备份源")
|
||||
sys.exit(1)
|
||||
|
||||
total_files = sum(
|
||||
sum(1 for _ in src.rglob("*") if _.is_file())
|
||||
for src in sources
|
||||
)
|
||||
total_size = sum(
|
||||
sum(f.stat().st_size for f in src.rglob("*") if f.is_file())
|
||||
for src in sources
|
||||
)
|
||||
|
||||
print(f"备份目标: {', '.join(str(s.relative_to(root)) for s in sources)}")
|
||||
print(f"文件数量: {total_files} 预计大小: {format_size(total_size)}")
|
||||
print(f"输出路径: {output_path}")
|
||||
|
||||
if dry_run:
|
||||
print("[dry-run] 跳过实际打包")
|
||||
return
|
||||
|
||||
with tarfile.open(output_path, "w:gz") as tar:
|
||||
for src in sources:
|
||||
tar.add(src, arcname=src.relative_to(root))
|
||||
|
||||
actual_size = output_path.stat().st_size
|
||||
print(f"✅ 备份完成: {filename} ({format_size(actual_size)})")
|
||||
|
||||
# 清理旧备份
|
||||
_prune_backups(backup_dir, keep)
|
||||
|
||||
|
||||
def _prune_backups(backup_dir: Path, keep: int):
|
||||
"""保留最近 keep 份,删除更旧的"""
|
||||
pattern = re.compile(r"kb_backup_\d{4}-\d{2}-\d{2}_\d{4}\.tar\.gz")
|
||||
backups = sorted(
|
||||
[f for f in backup_dir.iterdir() if f.is_file() and pattern.match(f.name)],
|
||||
key=lambda f: f.stat().st_mtime,
|
||||
reverse=True,
|
||||
)
|
||||
to_delete = backups[keep:]
|
||||
if to_delete:
|
||||
print(f"清理旧备份(保留最近{keep}份):")
|
||||
for f in to_delete:
|
||||
print(f" 删除: {f.name}")
|
||||
f.unlink()
|
||||
else:
|
||||
print(f"备份数量: {len(backups)} 份,无需清理")
|
||||
|
||||
|
||||
# ── export 模式 ────────────────────────────────────────────────────────────────
|
||||
|
||||
def run_export(root: Path, domains: list, tag: str, dry_run: bool):
|
||||
now = datetime.now()
|
||||
suffix = f"_tag-{tag}" if tag else ""
|
||||
suffix += f"_dom-{'_'.join(domains)}" if domains else ""
|
||||
filename = f"kb_export_{now.strftime('%Y-%m-%d_%H%M')}{suffix}.tar.gz"
|
||||
export_dir = root / "04_memory" / "memory_backup"
|
||||
ensure_dir(export_dir, dry_run)
|
||||
output_path = export_dir / filename
|
||||
|
||||
# 确定要导出的目录
|
||||
kb_root = root / "03_knowledge"
|
||||
if domains:
|
||||
scan_dirs = [kb_root / d for d in domains if (kb_root / d).exists()]
|
||||
missing = [d for d in domains if not (kb_root / d).exists()]
|
||||
if missing:
|
||||
print(f"[警告] 以下领域目录不存在: {', '.join(missing)}")
|
||||
else:
|
||||
scan_dirs = [kb_root / d for d in KNOWLEDGE_DOMAINS if (kb_root / d).exists()]
|
||||
|
||||
# 收集文件
|
||||
all_files = []
|
||||
for d in scan_dirs:
|
||||
for f in d.rglob("*.md"):
|
||||
if tag:
|
||||
if tag in get_md_tag(f):
|
||||
all_files.append(f)
|
||||
else:
|
||||
all_files.append(f)
|
||||
|
||||
if not all_files:
|
||||
print("[ERROR] 没有找到符合条件的文件")
|
||||
sys.exit(1)
|
||||
|
||||
total_size = sum(f.stat().st_size for f in all_files)
|
||||
filter_desc = f" 标签过滤: {tag}" if tag else ""
|
||||
domain_desc = f" 领域: {', '.join(domains)}" if domains else " 领域: 全部"
|
||||
print(f"导出范围: {domain_desc}{filter_desc}")
|
||||
print(f"文件数量: {len(all_files)} 预计大小: {format_size(total_size)}")
|
||||
print(f"输出路径: {output_path}")
|
||||
|
||||
if dry_run:
|
||||
print("[dry-run] 跳过实际打包")
|
||||
for f in all_files[:5]:
|
||||
print(f" 示例: {f.relative_to(root)}")
|
||||
if len(all_files) > 5:
|
||||
print(f" ... 共 {len(all_files)} 个文件")
|
||||
return
|
||||
|
||||
with tarfile.open(output_path, "w:gz") as tar:
|
||||
for f in all_files:
|
||||
tar.add(f, arcname=f.relative_to(root))
|
||||
|
||||
actual_size = output_path.stat().st_size
|
||||
print(f"✅ 导出完成: {filename} ({format_size(actual_size)},{len(all_files)} 个文件)")
|
||||
|
||||
|
||||
# ── status 模式 ────────────────────────────────────────────────────────────────
|
||||
|
||||
def run_status(root: Path, nutstore_path: str):
|
||||
print("=" * 50)
|
||||
print("AgentOS 同步状态检查")
|
||||
print("=" * 50)
|
||||
|
||||
# 1. 本地知识库状态
|
||||
kb_root = root / "03_knowledge"
|
||||
facts_db = root / "04_memory" / "long_term" / "facts.db"
|
||||
backup_dir = root / "04_memory" / "memory_backup"
|
||||
|
||||
if kb_root.exists():
|
||||
md_count = sum(1 for _ in kb_root.rglob("*.md"))
|
||||
kb_size = sum(f.stat().st_size for f in kb_root.rglob("*") if f.is_file())
|
||||
print(f"\n📚 知识库: {md_count} 个 MD 文件 ({format_size(kb_size)})")
|
||||
else:
|
||||
print(f"\n📚 知识库: ❌ 目录不存在 ({kb_root})")
|
||||
|
||||
if facts_db.exists():
|
||||
facts_mtime = datetime.fromtimestamp(facts_db.stat().st_mtime)
|
||||
facts_size = facts_db.stat().st_size
|
||||
print(f"🧠 facts.db: {format_size(facts_size)} 最后修改: {facts_mtime.strftime('%Y-%m-%d %H:%M')}")
|
||||
else:
|
||||
print(f"🧠 facts.db: ❌ 不存在")
|
||||
|
||||
# 2. 备份状态
|
||||
if backup_dir.exists():
|
||||
pattern = re.compile(r"kb_backup_\d{4}-\d{2}-\d{2}_\d{4}\.tar\.gz")
|
||||
backups = sorted(
|
||||
[f for f in backup_dir.iterdir() if f.is_file() and pattern.match(f.name)],
|
||||
key=lambda f: f.stat().st_mtime,
|
||||
reverse=True,
|
||||
)
|
||||
if backups:
|
||||
latest = backups[0]
|
||||
latest_time = datetime.fromtimestamp(latest.stat().st_mtime)
|
||||
print(f"\n💾 最新备份: {latest.name}")
|
||||
print(f" 时间: {latest_time.strftime('%Y-%m-%d %H:%M')} 大小: {format_size(latest.stat().st_size)}")
|
||||
print(f" 共 {len(backups)} 份备份")
|
||||
else:
|
||||
print(f"\n💾 备份: ❌ 暂无备份文件")
|
||||
else:
|
||||
print(f"\n💾 备份目录: ❌ 不存在")
|
||||
|
||||
# 3. 坚果云同步状态
|
||||
nutstore = Path(nutstore_path).expanduser()
|
||||
print(f"\n☁️ 坚果云同步路径: {nutstore}")
|
||||
if nutstore.exists():
|
||||
all_files = [f for f in nutstore.rglob("*") if f.is_file()]
|
||||
if all_files:
|
||||
latest_sync = max(all_files, key=lambda f: f.stat().st_mtime)
|
||||
sync_time = datetime.fromtimestamp(latest_sync.stat().st_mtime)
|
||||
sync_size = sum(f.stat().st_size for f in all_files)
|
||||
print(f" 状态: ✅ 路径存在")
|
||||
print(f" 文件数: {len(all_files)} 总大小: {format_size(sync_size)}")
|
||||
print(f" 最近同步: {sync_time.strftime('%Y-%m-%d %H:%M')} ({latest_sync.name})")
|
||||
else:
|
||||
print(f" 状态: ⚠️ 路径存在但为空")
|
||||
else:
|
||||
print(f" 状态: ❌ 路径不存在(坚果云未挂载或未安装)")
|
||||
|
||||
print()
|
||||
|
||||
|
||||
# ── 入口 ───────────────────────────────────────────────────────────────────────
|
||||
|
||||
def main():
|
||||
args = parse_args()
|
||||
root = Path(args.root).expanduser()
|
||||
|
||||
if not root.exists():
|
||||
print(f"[ERROR] agent-os 根目录不存在: {root}")
|
||||
sys.exit(1)
|
||||
|
||||
if args.mode == "backup":
|
||||
run_backup(root, args.keep, args.dry_run)
|
||||
elif args.mode == "export":
|
||||
run_export(root, args.domain or [], args.tag, args.dry_run)
|
||||
elif args.mode == "status":
|
||||
run_status(root, args.nutstore)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
Reference in New Issue
Block a user