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)。
217 lines
8.2 KiB
Python
217 lines
8.2 KiB
Python
"""
|
||
v2_accounts.py — V2 统一账号管理 API 路由
|
||
|
||
基于 AccountService 提供联邦账号中心的统一数据视图。
|
||
旧 API(matrix.py /api/matrix/accounts)保留兼容。
|
||
|
||
前缀: /api/v2
|
||
"""
|
||
import json
|
||
import logging
|
||
import os
|
||
import sys
|
||
import time
|
||
from pathlib import Path
|
||
from fastapi import APIRouter, HTTPException
|
||
|
||
logger = logging.getLogger("dashboard.routes.v2_accounts")
|
||
|
||
_THIS_DIR = Path(__file__).resolve().parent.parent
|
||
AGENT_SYNC = Path(os.environ.get("AGENT_SYNC", str(Path.home() / "workbuddy-agent-os" / "agent-sync")))
|
||
AGENT_LOCAL = Path(os.environ.get("AGENT_LOCAL", str(Path.home() / "workbuddy-agent-os" / "agent-local")))
|
||
|
||
router = APIRouter(prefix="/api/v2", tags=["v2_accounts"])
|
||
|
||
|
||
def _get_svc():
|
||
"""获取 AccountService 实例(延迟导入,单例缓存)"""
|
||
if not hasattr(_get_svc, "_instance"):
|
||
sys.path.insert(0, str(_THIS_DIR))
|
||
from services.account_service import AccountService
|
||
_get_svc._instance = AccountService()
|
||
return _get_svc._instance
|
||
|
||
|
||
# ═══════════════════════════════════════════════════════════
|
||
# 账号列表与详情
|
||
# ═══════════════════════════════════════════════════════════
|
||
|
||
@router.get("/accounts")
|
||
def api_v2_accounts(
|
||
machine: str = "",
|
||
platform: str = "",
|
||
status: str = "",
|
||
q: str = "",
|
||
):
|
||
"""获取所有账号的统一视图"""
|
||
try:
|
||
svc = _get_svc()
|
||
accounts = svc.get_all_accounts()
|
||
|
||
if machine:
|
||
accounts = [a for a in accounts if a.get("owner_machine") == machine]
|
||
if platform:
|
||
accounts = [a for a in accounts if a.get("platform") == platform]
|
||
if status:
|
||
accounts = [a for a in accounts if a.get("login_status") == status]
|
||
if q:
|
||
lq = q.lower()
|
||
accounts = [a for a in accounts if lq in a["id"].lower() or lq in a.get("phone", "").lower()
|
||
or lq in a.get("nickname", "").lower() or lq in a.get("identity_dir", "").lower()
|
||
or lq in a.get("owner_machine", "").lower()]
|
||
|
||
return {"accounts": accounts, "total": len(accounts), "generated_at": time.time()}
|
||
except Exception as e:
|
||
logger.error(f"v2/accounts 查询失败: {e}", exc_info=True)
|
||
raise HTTPException(500, detail=str(e))
|
||
|
||
|
||
@router.get("/accounts/{account_id}")
|
||
def api_v2_account_detail(account_id: str):
|
||
"""获取单个账号详情"""
|
||
try:
|
||
svc = _get_svc()
|
||
acct = svc.get_account_detail(account_id)
|
||
if not acct:
|
||
raise HTTPException(404, detail=f"账号 {account_id} 不存在")
|
||
return acct
|
||
except HTTPException:
|
||
raise
|
||
except Exception as e:
|
||
raise HTTPException(500, detail=str(e))
|
||
|
||
|
||
@router.get("/status-summary")
|
||
def api_v2_status_summary():
|
||
"""获取账号状态汇总"""
|
||
try:
|
||
svc = _get_svc()
|
||
accounts = svc.get_all_accounts()
|
||
summary = {"total": len(accounts), "logged_in": 0, "cookie_expiring": 0,
|
||
"no_cookie": 0, "banned": 0, "disabled": 0, "unknown": 0}
|
||
for a in accounts:
|
||
s = a.get("login_status", "unknown")
|
||
if s in summary:
|
||
summary[s] += 1
|
||
else:
|
||
summary["unknown"] += 1
|
||
return summary
|
||
except Exception as e:
|
||
raise HTTPException(500, detail=str(e))
|
||
|
||
|
||
@router.patch("/accounts/{account_id}")
|
||
def api_v2_update_notes(account_id: str, data: dict):
|
||
"""更新账号备注/标签
|
||
|
||
- 本机账号:写入本机 accounts.yaml(MatrixManager)
|
||
- 远程账号:本地找不到时,tags 写入集中标签文件(agent-local/data/account_tags_cache.json)
|
||
"""
|
||
try:
|
||
import yaml
|
||
sys.path.insert(0, str(AGENT_SYNC / "05_tools" / "07_matrix" / "scripts"))
|
||
from matrix_mgmt import MatrixManager
|
||
mgr = MatrixManager()
|
||
update_data = {}
|
||
if "notes" in data:
|
||
update_data["notes"] = data["notes"]
|
||
if "tags" in data:
|
||
update_data["tags"] = data["tags"]
|
||
if not update_data:
|
||
return {"status": "ok"}
|
||
try:
|
||
mgr.update_account(account_id, update_data)
|
||
except ValueError:
|
||
# 远程账号(本机 MatrixManager 无此账号)→ tags 写入集中标签文件
|
||
if "tags" in update_data and update_data["tags"] is not None:
|
||
_save_remote_tags(account_id, update_data["tags"])
|
||
else:
|
||
raise HTTPException(404, detail=f"账号 {account_id} 不存在")
|
||
return {"status": "ok"}
|
||
except HTTPException:
|
||
raise
|
||
except Exception as e:
|
||
raise HTTPException(500, detail=str(e))
|
||
|
||
|
||
def _save_remote_tags(account_id: str, tags: list) -> None:
|
||
"""远程账号标签 → 集中标签文件(agent-local/data/account_tags_cache.json)
|
||
|
||
读取侧 account_service._load_tags_cache 已从此文件读远程 tags,写入即可生效。
|
||
"""
|
||
import json as _json
|
||
_CACHE_PATH = AGENT_LOCAL / "data" / "account_tags_cache.json"
|
||
data = {}
|
||
if _CACHE_PATH.exists():
|
||
try:
|
||
data = _json.loads(_CACHE_PATH.read_text())
|
||
except Exception:
|
||
data = {}
|
||
data[account_id] = [str(t) for t in (tags or [])]
|
||
_CACHE_PATH.parent.mkdir(parents=True, exist_ok=True)
|
||
_CACHE_PATH.write_text(_json.dumps(data, ensure_ascii=False, indent=1))
|
||
logger.info(" 🏷️ 远程账号 %s 标签已保存到集中文件 (%d 条)", account_id, len(data[account_id]))
|
||
|
||
|
||
# ═══════════════════════════════════════════════════════════
|
||
# 批量操作
|
||
# ═══════════════════════════════════════════════════════════
|
||
|
||
@router.post("/accounts/batch")
|
||
def api_v2_accounts_batch(data: dict):
|
||
"""批量操作选中的账号"""
|
||
account_ids = data.get("account_ids", [])
|
||
action = data.get("action", "")
|
||
params = data.get("params", {})
|
||
|
||
if not account_ids:
|
||
raise HTTPException(400, detail="account_ids 必填")
|
||
if not action:
|
||
raise HTTPException(400, detail="action 必填")
|
||
|
||
valid_actions = ("collect", "login", "nurture", "comment", "record")
|
||
if action not in valid_actions:
|
||
raise HTTPException(400, detail=f"action 必须是 {valid_actions} 之一")
|
||
|
||
try:
|
||
from services.command_bus import CommandBus
|
||
svc = _get_svc()
|
||
all_accounts = svc.get_all_accounts()
|
||
acct_map = {a["id"]: a for a in all_accounts}
|
||
machine_groups = {}
|
||
for aid in account_ids:
|
||
a = acct_map.get(aid)
|
||
if not a:
|
||
raise HTTPException(400, detail=f"账号 {aid} 不存在")
|
||
m = a.get("owner_machine", "")
|
||
if m not in machine_groups:
|
||
machine_groups[m] = []
|
||
machine_groups[m].append(aid)
|
||
|
||
results = []
|
||
for machine, ids in machine_groups.items():
|
||
r = CommandBus.dispatch(action, ids, params)
|
||
results.append(r)
|
||
return {"status": "ok", "results": results}
|
||
except HTTPException:
|
||
raise
|
||
except Exception as e:
|
||
logger.error(f"v2/accounts/batch 失败: {e}")
|
||
raise HTTPException(500, detail=str(e))
|
||
|
||
|
||
# ═══════════════════════════════════════════════════════════
|
||
# 状态刷新
|
||
# ═══════════════════════════════════════════════════════════
|
||
|
||
@router.post("/accounts/refresh")
|
||
def api_v2_accounts_refresh(data: dict):
|
||
"""强制刷新账号状态"""
|
||
try:
|
||
svc = _get_svc()
|
||
account_ids = data.get("account_ids")
|
||
result = svc.batch_refresh_status(account_ids)
|
||
return {"status": "ok", "refreshed": len(result), "accounts": result}
|
||
except Exception as e:
|
||
raise HTTPException(500, detail=str(e))
|