butubb
2112c1a870
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）。
362 lines
14 KiB
Python
362 lines
14 KiB
Python
"""
|
||
scrape_engine.py — 抓取系统核心引擎 v1
|
||
|
||
职责：
|
||
1. URL 解析 → 标准化目标
|
||
2. 适配器路由（按平台选择适配器）
|
||
3. 工具降级执行
|
||
4. 结果入库（SQLite，去重）
|
||
5. 多机分发（通过 command_bus）
|
||
6. 异步任务管理（大量目标用轮询）
|
||
"""
|
||
import asyncio, json, logging, os, time, uuid
|
||
from datetime import datetime
|
||
from pathlib import Path
|
||
from typing import Optional
|
||
|
||
from services.scrape_db import ScrapeDB
|
||
from services.adapters.douyin_scrape import DouyinScrapeAdapter
|
||
from services.adapters.xhs_scrape import XhsScrapeAdapter
|
||
from services.adapters.bilibili_scrape import BilibiliScrapeAdapter
|
||
from services.adapters.zhihu_scrape import ZhihuScrapeAdapter
|
||
from services.adapters.web_scrape import WebScrapeAdapter
|
||
|
||
logger = logging.getLogger("dashboard.scrape.engine")
|
||
|
||
_THIS_DIR = Path(__file__).resolve().parent
|
||
|
||
|
||
class ScrapeEngine:
|
||
"""抓取引擎"""
|
||
|
||
def __init__(self):
|
||
self.db = ScrapeDB()
|
||
self._adapters = {}
|
||
self._tasks = {} # 异步任务缓存 {run_id: status}
|
||
|
||
def _get_adapter(self, platform: str):
|
||
"""获取/缓存适配器实例"""
|
||
if platform not in self._adapters:
|
||
if platform == "douyin":
|
||
self._adapters[platform] = DouyinScrapeAdapter()
|
||
elif platform in ("xiaohongshu", "xhs"):
|
||
self._adapters[platform] = XhsScrapeAdapter()
|
||
elif platform == "bilibili":
|
||
self._adapters[platform] = BilibiliScrapeAdapter()
|
||
elif platform == "zhihu":
|
||
self._adapters[platform] = ZhihuScrapeAdapter()
|
||
elif platform == "web":
|
||
self._adapters[platform] = WebScrapeAdapter()
|
||
else:
|
||
raise ValueError(f"不支持的平台: {platform}")
|
||
return self._adapters[platform]
|
||
|
||
# ── URL 解析 ──
|
||
|
||
def resolve_target(self, raw: str) -> dict:
|
||
"""解析单个目标 → 标准化格式"""
|
||
raw = raw.strip()
|
||
import re
|
||
|
||
# ── 抖音 ──
|
||
# 抖音短链: v.douyin.com/xxx
|
||
if "v.douyin.com" in raw:
|
||
return {"input": raw, "platform": "douyin", "type": "shortlink",
|
||
"status": "pending"}
|
||
# 抖音视频页: douyin.com/video/{id}
|
||
if "douyin.com/video/" in raw:
|
||
m = re.search(r'/video/(\d+)', raw)
|
||
return {"input": raw, "platform": "douyin", "type": "video",
|
||
"target_id": m.group(1) if m else raw, "status": "resolved"}
|
||
# 抖音用户主页: douyin.com/user/{sec_uid}
|
||
if "douyin.com/user/" in raw:
|
||
sec_uid = raw.split("/user/")[-1].split("?")[0]
|
||
return {"input": raw, "platform": "douyin", "type": "user",
|
||
"target_id": sec_uid, "status": "resolved"}
|
||
|
||
# ── 小红书 ──
|
||
if "xiaohongshu.com" in raw:
|
||
m = re.search(r'/explore/([a-f0-9]+)', raw)
|
||
if m:
|
||
return {"input": raw, "platform": "xiaohongshu", "type": "note",
|
||
"target_id": m.group(1), "status": "resolved"}
|
||
|
||
# ── B站 ──
|
||
if "bilibili.com" in raw:
|
||
# bilibili.com/video/BVxxx
|
||
m = re.search(r'/video/([a-zA-Z0-9]+)', raw)
|
||
if m:
|
||
return {"input": raw, "platform": "bilibili", "type": "video",
|
||
"target_id": m.group(1), "status": "resolved"}
|
||
# bilibili.com/space/{mid}
|
||
m = re.search(r'/space/(\d+)', raw)
|
||
if m:
|
||
return {"input": raw, "platform": "bilibili", "type": "user",
|
||
"target_id": m.group(1), "status": "resolved"}
|
||
|
||
# ── 知乎 ──
|
||
if "zhihu.com" in raw:
|
||
# zhihu.com/answer/{id}
|
||
m = re.search(r'/answer/(\d+)', raw)
|
||
if m:
|
||
return {"input": raw, "platform": "zhihu", "type": "item",
|
||
"target_id": m.group(1), "status": "resolved"}
|
||
# zhihu.com/question/{id}
|
||
m = re.search(r'/question/(\d+)', raw)
|
||
if m:
|
||
return {"input": raw, "platform": "zhihu", "type": "item",
|
||
"target_id": m.group(1), "status": "resolved"}
|
||
|
||
# ── 识别 URL（通用网页）──
|
||
if raw.startswith("http://") or raw.startswith("https://"):
|
||
return {"input": raw, "platform": "web", "type": "page",
|
||
"target_id": raw, "status": "resolved"}
|
||
|
||
# ── 纯 sec_uid（抖音）──
|
||
if raw.startswith("MS4w") or len(raw) > 20:
|
||
return {"input": raw, "platform": "douyin", "type": "user",
|
||
"target_id": raw, "status": "resolved"}
|
||
# 纯数字（可能是 aweme_id）
|
||
if raw.isdigit() and len(raw) >= 15:
|
||
return {"input": raw, "platform": "douyin", "type": "video",
|
||
"target_id": raw, "status": "resolved"}
|
||
# 纯数字（知乎 answer_id）
|
||
if raw.isdigit():
|
||
return {"input": raw, "platform": "zhihu", "type": "item",
|
||
"target_id": raw, "status": "resolved"}
|
||
return {"input": raw, "platform": "unknown", "type": "unknown",
|
||
"status": "unresolved"}
|
||
|
||
def resolve_urls(self, targets: list) -> list[dict]:
|
||
"""批量解析 URL"""
|
||
return [self.resolve_target(t) for t in targets]
|
||
|
||
# ── URL 短链解析 ──
|
||
|
||
def _resolve_shortlink(self, url: str) -> Optional[str]:
|
||
"""解析抖音短链 → aweme_id"""
|
||
import subprocess, re
|
||
try:
|
||
result = subprocess.run(
|
||
["curl", "-sI", url],
|
||
capture_output=True, text=True, timeout=10
|
||
)
|
||
loc = ""
|
||
for line in result.stdout.split("\n"):
|
||
if line.lower().startswith("location:"):
|
||
loc = line.split(":", 1)[1].strip()
|
||
break
|
||
if loc:
|
||
m = re.search(r'/video/(\d+)', loc)
|
||
if m:
|
||
return m.group(1)
|
||
# iesdouyin.com 格式
|
||
m = re.search(r'/share/video/(\d+)', loc)
|
||
if m:
|
||
return m.group(1)
|
||
return None
|
||
except:
|
||
return None
|
||
|
||
# ── 核心执行 ──
|
||
|
||
async def run(self, request: dict) -> dict:
|
||
"""
|
||
执行抓取任务
|
||
|
||
request:
|
||
targets: [str] — 目标列表
|
||
platform: str — 自动识别时留空
|
||
depth: "light" — "light"/"deep"
|
||
tool_level: int — 1/2/3
|
||
machine: str — 指定机器（空=本机）
|
||
multi_machine: bool — 是否多机协同
|
||
async_mode: bool — 是否异步（大量目标用）
|
||
"""
|
||
t0 = time.time()
|
||
targets = request.get("targets", [])
|
||
platform = request.get("platform", "auto")
|
||
depth = request.get("depth", "light")
|
||
tool_level = request.get("tool_level", 2)
|
||
machine = request.get("machine", "")
|
||
multi_machine = request.get("multi_machine", False)
|
||
async_mode = request.get("async_mode", False) or len(targets) > 50
|
||
|
||
# 1. 解析目标
|
||
resolved = self.resolve_urls(targets)
|
||
unresolved = [r for r in resolved if r["status"] == "unresolved"]
|
||
shortlinks = [r for r in resolved if r["type"] == "shortlink"]
|
||
|
||
# 2. 解析短链
|
||
for sl in shortlinks:
|
||
aweme_id = self._resolve_shortlink(sl["input"])
|
||
if aweme_id:
|
||
sl["type"] = "video"
|
||
sl["target_id"] = aweme_id
|
||
sl["platform"] = "douyin"
|
||
sl["status"] = "resolved"
|
||
else:
|
||
sl["status"] = "unresolved"
|
||
|
||
ready = [r for r in resolved if r["status"] == "resolved"]
|
||
|
||
# 3. 异步模式 → 创建后台任务
|
||
if async_mode:
|
||
run_id = str(uuid.uuid4())[:8]
|
||
self._tasks[run_id] = {
|
||
"id": run_id, "status": "running",
|
||
"total": len(ready), "completed": 0,
|
||
"results": [], "errors": [],
|
||
"created_at": datetime.now().isoformat(),
|
||
}
|
||
asyncio.create_task(self._run_async(run_id, ready, platform,
|
||
depth, tool_level))
|
||
return {
|
||
"status": "async",
|
||
"run_id": run_id,
|
||
"total": len(ready),
|
||
"message": f"异步任务已创建（{len(ready)} 个目标）",
|
||
}
|
||
|
||
# 4. 同步模式（目标少，直接执行）
|
||
task_id = self.db.create_task(
|
||
"single", platform or "auto",
|
||
json.dumps([t["input"] for t in ready], ensure_ascii=False),
|
||
depth, tool_level, machine, len(ready)
|
||
)
|
||
results = []
|
||
errors = []
|
||
for target in ready:
|
||
try:
|
||
result = await self._scrape_one(
|
||
target, platform or target["platform"],
|
||
depth, tool_level
|
||
)
|
||
if result is None:
|
||
continue
|
||
# result 可能是 dict（单条）或 list（用户的所有内容）
|
||
items = result if isinstance(result, list) else [result]
|
||
for item in items:
|
||
self._save_item(task_id, item)
|
||
item.pop("comments", None)
|
||
results.append(item)
|
||
except Exception as e:
|
||
errors.append({"target": target["input"], "error": str(e)})
|
||
|
||
self.db.update_task_status(task_id, "completed" if not errors else "completed",
|
||
summary=json.dumps({"success": len(results),
|
||
"errors": len(errors)},
|
||
ensure_ascii=False))
|
||
|
||
return {
|
||
"status": "completed",
|
||
"task_id": task_id,
|
||
"duration": round(time.time() - t0, 1),
|
||
"total": len(ready),
|
||
"success": len(results),
|
||
"errors": errors,
|
||
"data": results,
|
||
}
|
||
|
||
async def _scrape_one(self, target: dict, platform: str,
|
||
depth: str, tool_level: int):
|
||
"""
|
||
抓取单个目标。
|
||
返回:
|
||
- dict → 单条内容（video/note）
|
||
- list → 多条内容（user 类型，用户的所有视频/笔记）
|
||
- None → 抓取失败
|
||
"""
|
||
try:
|
||
adapter = self._get_adapter(platform)
|
||
except ValueError:
|
||
logger.warning(f" ⚠️ 不支持的平台: {platform}")
|
||
return None
|
||
|
||
try:
|
||
if target["type"] == "user":
|
||
# 用户抓取：返回该用户的所有内容（列表）
|
||
return await adapter.collect_user(target["target_id"])
|
||
elif target["type"] == "video":
|
||
return await adapter.collect_item(target["target_id"],
|
||
depth, tool_level)
|
||
elif target["type"] == "note":
|
||
return await adapter.collect_item(target["target_id"],
|
||
depth, tool_level)
|
||
else:
|
||
return None
|
||
except Exception as e:
|
||
logger.error(f" ❌ 抓取失败 [{target['input']}]: {e}")
|
||
return None
|
||
|
||
def _save_item(self, task_id: int, item: dict) -> int:
|
||
"""将单条抓取结果写入数据库，返回 db_id"""
|
||
db_id = self.db.insert_item(
|
||
task_id=task_id,
|
||
platform=item.get("platform", ""),
|
||
item_id=item.get("item_id", ""),
|
||
url=item.get("url", ""),
|
||
title=item.get("title", ""),
|
||
author_name=item.get("author_name", ""),
|
||
author_id=item.get("author_id", ""),
|
||
published_at=item.get("published_at", ""),
|
||
text_content=item.get("text_content", ""),
|
||
tags=item.get("tags", []),
|
||
stats=item.get("stats", {}),
|
||
extra=item.get("extra", {}),
|
||
media=item.get("media", []),
|
||
)
|
||
# 评论独立入库
|
||
comments = item.get("comments", [])
|
||
if comments and db_id:
|
||
self.db.insert_comments(db_id, comments)
|
||
return db_id
|
||
|
||
async def _run_async(self, run_id: str, targets: list, platform: str,
|
||
depth: str, tool_level: int):
|
||
"""异步执行大量目标（结果写入 SQLite 持久化，不依赖内存）"""
|
||
task_data = self._tasks.get(run_id, {})
|
||
# 异步模式也创建持久化任务记录
|
||
all_targets_json = json.dumps([t["input"] for t in targets], ensure_ascii=False)
|
||
task_id = self.db.create_task("batch", platform or "auto",
|
||
all_targets_json, depth, tool_level,
|
||
"async", len(targets))
|
||
for target in targets:
|
||
try:
|
||
result = await self._scrape_one(
|
||
target, platform or target["platform"],
|
||
depth, tool_level
|
||
)
|
||
if result:
|
||
items = result if isinstance(result, list) else [result]
|
||
for item in items:
|
||
self._save_item(task_id, item)
|
||
item.pop("comments", None)
|
||
task_data["results"].append(item)
|
||
except Exception as e:
|
||
err = {"target": target["input"], "error": str(e)}
|
||
task_data["errors"].append(err)
|
||
task_data["completed"] += 1
|
||
task_data["status"] = "completed"
|
||
# 更新任务状态
|
||
self.db.update_task_status(task_id, "completed",
|
||
summary=json.dumps({
|
||
"success": len(task_data["results"]),
|
||
"errors": len(task_data["errors"]),
|
||
}, ensure_ascii=False))
|
||
|
||
def get_async_result(self, run_id: str) -> Optional[dict]:
|
||
"""查询异步任务结果"""
|
||
return self._tasks.get(run_id)
|
||
|
||
# ── 统计 ──
|
||
|
||
def get_stats(self) -> dict:
|
||
return {
|
||
"total": sum(self.db.count_by_platform().values()),
|
||
"today": self.db.count_today(),
|
||
"sources": self.db.sources_count(),
|
||
"by_platform": self.db.count_by_platform(),
|
||
"tasks": self.db.task_stats(),
|
||
}
|