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）。
458 lines
17 KiB
Python
458 lines
17 KiB
Python
"""
|
||
mediacrawler_adapter.py — 抖音视频数据采集器 v3
|
||
|
||
全新架构：不再依赖 CDP/Playwright/浏览器页面。
|
||
直接从 Chrome profile 读取 cookie，通过 HTTP 请求调用抖音 API。
|
||
全程无窗口、无标签页、无闪烁。
|
||
"""
|
||
import asyncio, json, logging, os, re, time, sqlite3, urllib.request, urllib.error
|
||
from pathlib import Path
|
||
from typing import Optional
|
||
|
||
logger = logging.getLogger("dashboard.mediacrawler_adapter")
|
||
|
||
# HTTP 请求头（模拟浏览器）
|
||
DOUYIN_HEADERS = {
|
||
"User-Agent": (
|
||
"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) "
|
||
"AppleWebKit/537.36 (KHTML, like Gecko) "
|
||
"Chrome/150.0.0.0 Safari/537.36"
|
||
),
|
||
"Accept": "application/json, text/plain, */*",
|
||
"Accept-Language": "zh-CN,zh;q=0.9,en;q=0.8",
|
||
"Referer": "https://www.douyin.com/",
|
||
"Origin": "https://www.douyin.com",
|
||
}
|
||
|
||
|
||
# ── Playwright CDP 连接管理（复用，避免反复建连） ──
|
||
# Chrome 新版把 cookie 值加密存在 SQLite 里，必须通过 CDP 读解密后的值。
|
||
|
||
_PW = None
|
||
_BROWSER = None
|
||
|
||
async def _ensure_cdp():
|
||
"""确保 Playwright CDP 连接可用（全局复用）"""
|
||
global _PW, _BROWSER
|
||
if _BROWSER and _PW:
|
||
try:
|
||
if _BROWSER.is_connected():
|
||
return _BROWSER.contexts[0] if _BROWSER.contexts else None
|
||
except:
|
||
pass
|
||
# 断连了，重建
|
||
try:
|
||
await _BROWSER.close()
|
||
except:
|
||
pass
|
||
try:
|
||
await _PW.stop()
|
||
except:
|
||
pass
|
||
_BROWSER = None
|
||
_PW = None
|
||
|
||
try:
|
||
from playwright.async_api import async_playwright
|
||
_PW = await async_playwright().start()
|
||
_BROWSER = await _PW.chromium.connect_over_cdp("http://127.0.0.1:9222")
|
||
ctx = _BROWSER.contexts[0] if _BROWSER.contexts else None
|
||
return ctx
|
||
except Exception as e:
|
||
_BROWSER = None
|
||
_PW = None
|
||
logger.warning(f"CDP 连接失败: {e}")
|
||
return None
|
||
|
||
|
||
async def _get_cookies() -> dict:
|
||
"""从 CDP 连接读取 Chrome cookie（解密后的值）"""
|
||
ctx = await _ensure_cdp()
|
||
if not ctx:
|
||
return {}
|
||
try:
|
||
all_cookies = await ctx.cookies()
|
||
result = {}
|
||
for c in all_cookies:
|
||
domain = c.get("domain", "")
|
||
if "douyin" in domain or "amemv" in domain:
|
||
result[c["name"]] = c["value"]
|
||
return result
|
||
except Exception as e:
|
||
logger.warning(f"读 cookie 失败: {e}")
|
||
return {}
|
||
|
||
|
||
def _cookie_str(cookies: dict) -> str:
|
||
"""将 cookie 字典转为 HTTP Header 字符串"""
|
||
return "; ".join(f"{k}={v}" for k, v in cookies.items())
|
||
|
||
|
||
# ── HTTP API 请求（不打开任何浏览器页面） ──
|
||
|
||
def _http_get(url: str, cookies: dict, timeout: int = 15) -> dict:
|
||
"""发起 HTTP GET 请求并返回 JSON（cookies 必传，由调用方从 CDP 获取）"""
|
||
req = urllib.request.Request(
|
||
url,
|
||
headers={
|
||
**DOUYIN_HEADERS,
|
||
"Cookie": _cookie_str(cookies),
|
||
},
|
||
)
|
||
with urllib.request.urlopen(req, timeout=timeout) as resp:
|
||
return json.loads(resp.read().decode())
|
||
|
||
|
||
# ── 工具函数 ──
|
||
|
||
def _extract_aweme_id(url: str):
|
||
m = re.search(r'douyin\.com/video/(\d+)', url)
|
||
if m:
|
||
return m.group(1)
|
||
m = re.search(r'iesdouyin\.com/share/video/(\d+)', url)
|
||
if m:
|
||
return m.group(1)
|
||
# 抖音精选页 / 发现页：modal_id 就是视频 aweme_id（jingxuan?modal_id=xxx）
|
||
m = re.search(r'(?:jingxuan|discover|modal_id)[=?/](\d+)', url)
|
||
if m:
|
||
return m.group(1)
|
||
return None
|
||
|
||
|
||
def _resolve_shortlink(url: str):
|
||
import subprocess
|
||
try:
|
||
result = subprocess.run(
|
||
["curl", "-sI", url], capture_output=True, text=True, timeout=10
|
||
)
|
||
for line in result.stdout.split("\n"):
|
||
if line.lower().startswith("location:"):
|
||
loc = line.split(":", 1)[1].strip()
|
||
m = re.search(r'/video/(\d+)', loc)
|
||
if m:
|
||
return m.group(1)
|
||
except Exception:
|
||
pass
|
||
return None
|
||
|
||
|
||
# ── 核心：获取视频数据（纯 HTTP，无浏览器） ──
|
||
|
||
async def get_video_data(url: str) -> dict:
|
||
"""
|
||
统一入口：传入抖音视频 URL，返回完整数据
|
||
通过 HTTP API 获取，不打开任何浏览器页面
|
||
"""
|
||
aweme_id = _extract_aweme_id(url)
|
||
if not aweme_id:
|
||
aweme_id = _resolve_shortlink(url)
|
||
if not aweme_id:
|
||
return {"error": f"无法解析视频 ID: {url}"}
|
||
|
||
t0 = time.time()
|
||
result = {
|
||
"aweme_id": aweme_id,
|
||
"title": "",
|
||
"author": "",
|
||
"likes": 0,
|
||
"comments": 0,
|
||
"collects": 0,
|
||
"shares": 0,
|
||
"comment_texts": [],
|
||
"collected_at": time.strftime("%Y-%m-%d %H:%M:%S"),
|
||
}
|
||
|
||
# 读取登录态 cookie（通过 CDP，解密后的值）
|
||
cookies = await _get_cookies()
|
||
has_session = bool(cookies.get("sessionid"))
|
||
if not has_session:
|
||
result["error"] = "抖音登录已过期，请点击顶部「📱 打开抖音登录」按钮重新登录"
|
||
result["login_expired"] = True
|
||
return result
|
||
|
||
# 调抖音 API 获取视频详情
|
||
try:
|
||
api_url = (
|
||
f"https://www.douyin.com/aweme/v1/web/aweme/detail/"
|
||
f"?aweme_id={aweme_id}&version_code=170400&app_name=aweme&device_platform=web"
|
||
)
|
||
detail = await asyncio.to_thread(_http_get, api_url, cookies)
|
||
except urllib.error.HTTPError as e:
|
||
result["error"] = f"HTTP {e.code}: {e.reason}"
|
||
return result
|
||
except Exception as e:
|
||
result["error"] = f"请求失败: {e}"
|
||
return result
|
||
|
||
aweme = detail.get("aweme_detail", {})
|
||
if not aweme:
|
||
if detail.get("status_code") == 2:
|
||
result["error"] = "抖音登录已过期，请重新登录"
|
||
result["login_expired"] = True
|
||
return result
|
||
if detail.get("filter_detail"):
|
||
# filter_detail = 视频被平台过滤（已删除/违规/推广限制），接口层详情不可查
|
||
result["filtered"] = True
|
||
# 兜底策略：接口被过滤，改用页面浏览方式（打开网址读页面标题/meta），补全标题/博主
|
||
fb = await _page_fallback(url, aweme_id)
|
||
if fb:
|
||
result.update({
|
||
"title": fb.get("title", ""),
|
||
"author": fb.get("author", ""),
|
||
"published_at": fb.get("published_at", ""),
|
||
"source": "page_fallback",
|
||
"api_error": "⚠️ 视频被平台过滤（接口层），已用页面兜底补全标题/博主",
|
||
})
|
||
return result
|
||
result["error"] = "⚠️ 视频被平台过滤，无法采集（可能已删除/违规/推广限制）"
|
||
return result
|
||
result["error"] = f"API 未返回数据: {str(detail)[:200]}"
|
||
return result
|
||
|
||
statistics = aweme.get("statistics", {})
|
||
author_info = aweme.get("author", {})
|
||
|
||
result.update({
|
||
"title": aweme.get("desc", ""),
|
||
"author": author_info.get("nickname", ""),
|
||
"author_uid": author_info.get("uid", ""), # 数字ID（博主精确身份）
|
||
"author_sec_uid": author_info.get("sec_uid", ""), # 加密ID
|
||
"author_unique_id": author_info.get("unique_id", ""), # 抖音号
|
||
"likes": int(statistics.get("digg_count", 0)),
|
||
"comments": int(statistics.get("comment_count", 0)),
|
||
"collects": int(statistics.get("collect_count", 0)),
|
||
"shares": int(statistics.get("share_count", 0)),
|
||
})
|
||
|
||
# 获取热评（复用已有 Chrome 页面，不创建新标签页）
|
||
# Chrome 已有 douyin.com 页面，直接用它的 JS 上下文执行 fetch
|
||
# ⚠️ 不要 new_page() — 那会在 Chrome 中闪出新标签页
|
||
try:
|
||
_ctx2 = await _ensure_cdp()
|
||
if _ctx2:
|
||
_pages = _ctx2.pages
|
||
if _pages:
|
||
_page = _pages[0] # 复用已有页面（about:blank 或 douyin.com）
|
||
_cmt_json = await _page.evaluate(f"""async () => {{
|
||
try {{
|
||
const r = await fetch(
|
||
'https://www.douyin.com/aweme/v1/web/comment/list/?aweme_id={aweme_id}&count=20&cursor=0',
|
||
{{ headers: {{ 'Accept': 'application/json, text/plain, */*', 'Referer': 'https://www.douyin.com/' }} }}
|
||
);
|
||
return JSON.stringify(await r.json());
|
||
}} catch(e) {{ return JSON.stringify({{error: e.message}}); }}
|
||
}}""")
|
||
_cmt_data = json.loads(_cmt_json)
|
||
if "error" not in _cmt_data:
|
||
for c in _cmt_data.get("comments", [])[:20]:
|
||
user = c.get("user", {})
|
||
result["comment_texts"].append({
|
||
"nickname": user.get("nickname", ""),
|
||
"text": c.get("text", ""),
|
||
"likes": c.get("digg_count", 0),
|
||
})
|
||
except Exception as e:
|
||
logger.debug(f"评论获取失败: {e}")
|
||
|
||
result["_method"] = "http_api"
|
||
result["_duration"] = round(time.time() - t0, 2)
|
||
return result
|
||
|
||
|
||
async def _page_fallback(url: str, aweme_id: str) -> Optional[dict]:
|
||
"""兜底策略：接口被平台过滤(filter_detail)时，用 Chrome 打开页面（网址层面浏览），
|
||
读页面标题 + meta description 补全标题/博主/发布时间。
|
||
|
||
返回 {title, author, published_at, source:'page_fallback'} 或 None（CDP 不可用/超时/解析失败）
|
||
"""
|
||
try:
|
||
ctx = await _ensure_cdp()
|
||
if not ctx or not ctx.pages:
|
||
return None
|
||
page = ctx.pages[0] # 复用现有页面，不新建标签页
|
||
try:
|
||
await page.goto(url, timeout=20000, wait_until="domcontentloaded")
|
||
except Exception:
|
||
await asyncio.sleep(2)
|
||
# 轮询等待页面加载出标题/meta（最多 12s，抖音页面加载/重定向较慢）
|
||
title, desc, author, published = "", "", "", ""
|
||
for _ in range(12):
|
||
try:
|
||
raw = await page.evaluate(
|
||
"""() => {
|
||
const d = { title: document.title || '', desc: '' };
|
||
const md = document.querySelector('meta[name="description"]');
|
||
if (md) d.desc = md.getAttribute('content') || '';
|
||
return JSON.stringify(d);
|
||
}"""
|
||
)
|
||
data = json.loads(raw)
|
||
title = (data.get("title") or "").strip()
|
||
desc = (data.get("desc") or "").strip()
|
||
if title or desc:
|
||
break
|
||
except Exception:
|
||
pass
|
||
await asyncio.sleep(1)
|
||
|
||
m = re.search(r"[-—]\s*([^\s于]+)于(\d{8})发布", desc)
|
||
if m:
|
||
author, published = m.group(1), m.group(2)
|
||
# 标题兜底：desc 去掉「- 作者于时间发布」后缀
|
||
if not title and desc:
|
||
title = re.sub(r"\s*[-—]\s*[^\s于]+于\d{8}发布.*$", "", desc).strip()
|
||
return {
|
||
"title": title,
|
||
"author": author,
|
||
"published_at": published,
|
||
"source": "page_fallback",
|
||
}
|
||
except Exception as e:
|
||
logger.debug(f"页面兜底失败: {e}")
|
||
return None
|
||
|
||
|
||
# ── 博主采集（抖音博主监控） ──
|
||
# ⚠️ 注意：profile/other 和 aweme/post 接口要用「数字 uid」，
|
||
# 不能用 sec_user_id（会报 UserId不合法）。
|
||
# 数字 uid 从视频详情 API 的 author.uid 字段获取。
|
||
|
||
async def get_author_profile(uid: str) -> dict:
|
||
"""获取博主主页信息（粉丝数/获赞数/作品数/昵称/签名）
|
||
|
||
Args:
|
||
uid: 博主数字 ID（从视频详情的 author.uid 获取）
|
||
|
||
Returns:
|
||
{"uid", "nickname", "unique_id", "fans", "total_favorited",
|
||
"works", "following", "signature", "avatar", "sec_uid", ...}
|
||
失败时含 "error" 字段
|
||
"""
|
||
result = {"uid": str(uid)}
|
||
cookies = await _get_cookies()
|
||
has_session = bool(cookies.get("sessionid"))
|
||
if not has_session:
|
||
result["error"] = "抖音登录已过期"
|
||
result["login_expired"] = True
|
||
return result
|
||
|
||
try:
|
||
api_url = (
|
||
f"https://www.douyin.com/aweme/v1/web/user/profile/other/"
|
||
f"?user_id={uid}&device_platform=webapp&aid=6383"
|
||
)
|
||
data = await asyncio.to_thread(_http_get, api_url, cookies)
|
||
except urllib.error.HTTPError as e:
|
||
result["error"] = f"HTTP {e.code}: {e.reason}"
|
||
return result
|
||
except Exception as e:
|
||
result["error"] = f"请求失败: {e}"
|
||
return result
|
||
|
||
user = data.get("user", {})
|
||
if not user:
|
||
result["error"] = f"未返回用户数据: {data.get('status_msg', '')} (code={data.get('status_code')})"
|
||
return result
|
||
|
||
result.update({
|
||
"nickname": user.get("nickname", ""),
|
||
"unique_id": user.get("unique_id", ""),
|
||
"sec_uid": user.get("sec_uid", ""),
|
||
"fans": user.get("follower_count", 0),
|
||
"total_favorited": user.get("total_favorited", 0),
|
||
"works": user.get("aweme_count", 0),
|
||
"following": user.get("following_count", 0),
|
||
"signature": user.get("signature", ""),
|
||
"avatar": (user.get("avatar_thumb") or {}).get("url_list", [""])[0] if user.get("avatar_thumb") else "",
|
||
"collected_at": time.strftime("%Y-%m-%d %H:%M:%S"),
|
||
})
|
||
return result
|
||
|
||
|
||
async def get_author_videos(uid: str, count: int = 20) -> dict:
|
||
"""获取博主最新发布的视频列表（按发布时间倒序）
|
||
|
||
Args:
|
||
uid: 博主数字 ID
|
||
count: 拉取条数（默认20）
|
||
|
||
Returns:
|
||
{"uid", "videos": [{aweme_id, title, likes, comments, collects,
|
||
shares, create_time, url}], "has_more"}
|
||
"""
|
||
result = {"uid": str(uid), "videos": [], "has_more": False}
|
||
cookies = await _get_cookies()
|
||
has_session = bool(cookies.get("sessionid"))
|
||
if not has_session:
|
||
result["error"] = "抖音登录已过期"
|
||
result["login_expired"] = True
|
||
return result
|
||
|
||
try:
|
||
api_url = (
|
||
f"https://www.douyin.com/aweme/v1/web/aweme/post/"
|
||
f"?user_id={uid}&count={count}&max_cursor=0&device_platform=webapp&aid=6383"
|
||
)
|
||
data = await asyncio.to_thread(_http_get, api_url, cookies)
|
||
except urllib.error.HTTPError as e:
|
||
result["error"] = f"HTTP {e.code}: {e.reason}"
|
||
return result
|
||
except Exception as e:
|
||
result["error"] = f"请求失败: {e}"
|
||
return result
|
||
|
||
result["has_more"] = bool(data.get("has_more"))
|
||
for a in data.get("aweme_list", [])[:count]:
|
||
st = a.get("statistics", {})
|
||
result["videos"].append({
|
||
"aweme_id": str(a.get("aweme_id", "")),
|
||
"title": a.get("desc", ""),
|
||
"likes": int(st.get("digg_count", 0)),
|
||
"comments": int(st.get("comment_count", 0)),
|
||
"collects": int(st.get("collect_count", 0)),
|
||
"shares": int(st.get("share_count", 0)),
|
||
"create_time": a.get("create_time", 0),
|
||
"url": f"https://www.douyin.com/video/{a.get('aweme_id', '')}",
|
||
})
|
||
return result
|
||
|
||
|
||
# ── 登录状态检测（从 cookie 文件直接读取，无窗口操作） ──
|
||
|
||
async def check_login_status() -> dict:
|
||
"""检测抖音登录状态（通过 CDP，无页面操作）"""
|
||
cookies = await _get_cookies()
|
||
has_session = bool(cookies.get("sessionid"))
|
||
return {
|
||
"logged_in": has_session,
|
||
"cookie_sessionid": has_session,
|
||
"cookies_count": len(cookies),
|
||
}
|
||
|
||
|
||
# ── 打开登录页（用 AppleScript 控制 Chrome，不需要 CDP） ──
|
||
|
||
async def open_login_page():
|
||
"""在 Chrome 中打开抖音首页，让用户登录
|
||
|
||
使用 macOS 的 open 命令（比 AppleScript 更可靠，不依赖辅助功能权限）。
|
||
登录后 cookie 自动保存到 Chrome profile，两个 Chrome 都会检测。
|
||
"""
|
||
import subprocess as _sp
|
||
|
||
try:
|
||
# 先用 open 命令让 Chrome 激活并打开抖音首页
|
||
# open -a 会激活已有的 Chrome 窗口（无论是日常还是采集）
|
||
_sp.run(
|
||
["open", "-a", "Google Chrome", "https://www.douyin.com/"],
|
||
capture_output=True, timeout=10
|
||
)
|
||
logger.info("已请求 Chrome 打开抖音首页")
|
||
return {
|
||
"success": True,
|
||
"message": "已打开抖音首页，请在 Chrome 中点击右上角「登录」扫码",
|
||
}
|
||
except Exception as e:
|
||
logger.error(f"打开抖音首页失败: {e}")
|
||
return {"success": False, "error": f"打开失败: {e}"}
|