feat(monitor): 抖音接入博主监控
Deploy VitePress site to Pages / build (push) Canceled after 0s
Deploy VitePress site to Pages / Deploy (push) Canceled after 0s

上游爬虫本身不缺抖音能力(三模式、四项指标、二级评论都与小红书对等、指标还是同名同列),
缺的全在监控层的适配。这次把「平台之间不一样」的管子集中到一个新模块,再把散落的
xhs 硬编码接上去。

* 新增 api/monitor/adapters.py:产物目录名、jsonl 字段别名、目标链接形态与正则、
  通知链接模板。不放进 platforms.py 是因为那个模块被 describe_all() 整个序列化进
  /api/config/platforms 交给前端,塞进正则和目录名会让爬虫内部细节漏进 API 载荷。
  代价是两个注册表可能漂移,用一条测试钉住「声明接通就必须有适配器」。
* 两个必须知道的坑,都在这版里处理掉了:
  1) 抖音的平台 id 是 dy,而 store 把产物写在 douyin/ 下(store/douyin/_store_impl.py:47)。
     不改就是 ingest 一个文件都读不到 —— 不报错,只是 0 条,然后被冒充成「疑似登录失效」。
  2) 抖音的作品没有 note_id(叫 aweme_id)、评论也用 aweme_id 指作品。ingest 第一步是
     `if not note_id: continue`,不映射就逐条全丢。
  另外抖音顶层评论的 parent_comment_id 是字符串 "0",归一成空串,免得前端多出悬空的父节点。
* 顺带把「东西抓到了、只是没落在期望目录里」单独识别出来。这类故障的现象和登录失效
  一模一样,按登录失效报会把人指去查完全错误的方向。
* 修两个既有 bug(今天只有小红书所以无害,加抖音就踩响):
  - service.py update_task 换目标时漏传 task.platform,回落到默认小红书
  - scheduler.py 取 cookie 没传 platform,抖音任务会读着小红书那份 cookie 不动
* 行为变更(已与用户确认):cookie 闸门改成「没 cookie 且没开 CDP」才跳过。
  CDP 模式下登录态来自被接管的浏览器,粘不粘 cookie 由不得它决定;不放行的话,
  选了「接管已有 Chrome」却没粘 cookie 的用户会看到任务永远不触发,而且不报错。
  副作用是开启了 CDP 的小红书任务也不再被该闸门拦住 —— 语义上是对的。
* 目标输入框的示例链接与措辞改由能力矩阵提供(notes_label 抖音说「作品」、小红书说
  「笔记」;「建议只填纯 ID」是小红书专属劝告,抖音链接不带令牌,不再显示)。

测试 +22 条(858 通过),其中最关键的是「抖音作品/评论不被静默丢弃」与「产物目录名
不等于平台 id」两条 —— 都是把最难查的失败模式钉死在回归网里。

注意:抖音这条路的**端到端尚未验证**,需要一份可用的抖音登录态(CDP 那台 Chrome 里
登录,或导出一份 cookie)。单测覆盖的是解析与入库,真实抓取还没跑过。
This commit is contained in:
2026-10-10 14:55:28 +08:00
parent e348de48d3
commit 06718a1351
17 changed files with 863 additions and 105 deletions
+20 -16
View File
@@ -38,7 +38,7 @@ from ..schemas import (
SaveDataOptionEnum,
)
from ..services import crawler_manager
from . import app_settings, covers, notify
from . import adapters, app_settings, covers, notify
from .db import get_session
from .ingest import IngestResult, ingest_run
from .models import (
@@ -68,30 +68,33 @@ _PLATFORM_ENUM = {
"zhihu": PlatformEnum.ZHIHU,
}
_XHS_WEB_BASE = "https://www.xiaohongshu.com"
_CREATOR_PATH = "/user/profile"
_NOTE_PATH = "/explore"
# Timeout used when the caller does not care; tasks carry their own.
DEFAULT_RUN_TIMEOUT_SECONDS = 3600
def build_target_url(value: str, kind: str) -> str:
def build_target_url(value: str, kind: str, platform: str) -> str:
"""Turn a stored target into a URL the crawler's parser accepts.
Always emits a full URL rather than a bare id: the XHS parser accepts a bare
24-hex id only, so the URL form is the safer universal input. The
``xsec_token`` is appended when present but is deliberately optional -- it
expires, and the id alone is what keeps a long-running task alive.
Always emits a full URL rather than a bare id: both platforms' parsers accept
a bare id only in a narrower form, so the URL is the safer universal input.
The shape itself is platform-specific and comes from ``adapters``.
"""
path = _CREATOR_PATH if kind == MODE_CREATOR else _NOTE_PATH
return f"{_XHS_WEB_BASE}{path}/{value}"
spec = adapters.adapter(platform)
path = spec.creator_path if kind == MODE_CREATOR else spec.note_path
return f"{spec.web_base}{path}/{value}"
def build_target_urls(mode: str, targets: Iterable[MonitorTarget]) -> List[str]:
def build_target_urls(
mode: str, targets: Iterable[MonitorTarget], platform: str
) -> List[str]:
"""存储的目标 -> 爬虫接受的 URL。
``xsec_token`` 只有小红书有,而且是会过期的刷新令牌 —— 有就带上,没有就算了。
抖音恒为空,所以这一段对它是天然的 no-op,不需要平台分支。
"""
urls = []
for target in targets:
url = build_target_url(target.external_id, target.kind)
url = build_target_url(target.external_id, target.kind, platform)
if target.xsec_token:
url = f"{url}?xsec_token={target.xsec_token}"
if target.xsec_source:
@@ -159,7 +162,7 @@ async def execute_task(task_id: int, trigger: str = "manual") -> IngestResult:
raise ValueError(f"Monitor task {task_id} has no enabled targets")
platform = task.platform
urls = build_target_urls(task.mode, targets)
urls = build_target_urls(task.mode, targets, platform)
cookie = await get_cookie(session, platform)
strategy = await _strategy_settings(session, platform)
# System-wide switch. On a headless server the crawler must attach to the
@@ -249,7 +252,8 @@ async def execute_task(task_id: int, trigger: str = "manual") -> IngestResult:
if run is None or task is None:
raise ValueError(f"Run {run_id} or task {task_id} vanished during execution")
if exit_code == -1 and not (out_dir / "xhs").exists():
# 目录名按平台解析 —— 抖音的平台 id 是 dy 而产物目录是 douyin,写死就永远判不准。
if exit_code == -1 and not (out_dir / adapters.artifact_dir(task.platform)).exists():
# run_and_wait returns -1 when the process could not start or timed out.
run.status = RUN_TIMEOUT
run.finished_at = get_current_timestamp()