| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535 |
- import fs from 'node:fs/promises';
- import path from 'node:path';
- import type { Browser } from 'playwright';
- import { AppDataSource, PlatformAccount } from '../models/index.js';
- import { BrowserManager } from '../automation/browser.js';
- import { logger } from '../utils/logger.js';
- import { UserDayStatisticsService } from './UserDayStatisticsService.js';
- import { AccountService } from './AccountService.js';
- import type { ProxyConfig } from '@media-manager/shared';
- import { WS_EVENTS } from '@media-manager/shared';
- import { wsManager } from '../websocket/index.js';
- import { launchBrowser } from '../automation/browserProvider.js';
- type PlaywrightCookie = {
- name: string;
- value: string;
- domain?: string;
- path?: string;
- url?: string;
- expires?: number;
- httpOnly?: boolean;
- secure?: boolean;
- sameSite?: 'Lax' | 'None' | 'Strict';
- };
- function ensureDir(p: string) {
- return fs.mkdir(p, { recursive: true });
- }
- function normalizeDateText(input: unknown): Date | null {
- if (!input) return null;
- if (input instanceof Date && !Number.isNaN(input.getTime())) {
- const d = new Date(input);
- d.setHours(0, 0, 0, 0);
- return d;
- }
- const s = String(input).trim();
- if (!s) return null;
- // 2026-01-27 / 2026/01/27
- const m1 = s.match(/(\d{4})\D(\d{1,2})\D(\d{1,2})/);
- if (m1) {
- const yyyy = Number(m1[1]);
- const mm = Number(m1[2]);
- const dd = Number(m1[3]);
- if (!yyyy || !mm || !dd) return null;
- const d = new Date(yyyy, mm - 1, dd);
- d.setHours(0, 0, 0, 0);
- return d;
- }
- // 01-27(兜底:用当前年份)
- const m2 = s.match(/^(\d{1,2})[-/](\d{1,2})$/);
- if (m2) {
- const yyyy = new Date().getFullYear();
- const mm = Number(m2[1]);
- const dd = Number(m2[2]);
- const d = new Date(yyyy, mm - 1, dd);
- d.setHours(0, 0, 0, 0);
- return d;
- }
- return null;
- }
- function toRatePercentStringFromValue(val: unknown): string | undefined {
- const n = typeof val === 'number' ? val : Number(val);
- if (!Number.isFinite(n)) return undefined;
- if (n === 0) return '0';
- const scaled = n * 100;
- const rounded = Math.round(scaled * 100) / 100;
- const s = rounded.toFixed(2).replace(/\.00$/, '').replace(/(\.\d)0$/, '$1');
- return `${s}%`;
- }
- function parseCookiesFromAccount(cookieData: string | null): PlaywrightCookie[] {
- if (!cookieData) return [];
- const raw = cookieData.trim();
- if (!raw) return [];
- // 1) JSON array(最常见:浏览器插件导出/前端保存)
- if (raw.startsWith('[') || raw.startsWith('{')) {
- try {
- const parsed = JSON.parse(raw);
- const arr = Array.isArray(parsed) ? parsed : (parsed?.cookies ? parsed.cookies : []);
- if (!Array.isArray(arr)) return [];
- return arr
- .map((c: any) => {
- const name = String(c?.name ?? '').trim();
- const value = String(c?.value ?? '').trim();
- if (!name) return null;
- const domain = c?.domain ? String(c.domain) : undefined;
- const pathVal = c?.path ? String(c.path) : '/';
- const url = !domain ? 'https://creator.douyin.com' : undefined;
- const sameSiteRaw = c?.sameSite;
- const sameSite =
- sameSiteRaw === 'Lax' || sameSiteRaw === 'None' || sameSiteRaw === 'Strict'
- ? sameSiteRaw
- : undefined;
- return {
- name,
- value,
- domain,
- path: pathVal,
- url,
- expires: typeof c?.expires === 'number' ? c.expires : undefined,
- httpOnly: typeof c?.httpOnly === 'boolean' ? c.httpOnly : undefined,
- secure: typeof c?.secure === 'boolean' ? c.secure : undefined,
- sameSite,
- } satisfies PlaywrightCookie;
- })
- .filter(Boolean) as PlaywrightCookie[];
- } catch {
- // fallthrough
- }
- }
- // 2) "a=b; c=d" 拼接格式
- const pairs = raw.split(';').map((p) => p.trim()).filter(Boolean);
- const cookies: PlaywrightCookie[] = [];
- for (const p of pairs) {
- const idx = p.indexOf('=');
- if (idx <= 0) continue;
- const name = p.slice(0, idx).trim();
- const value = p.slice(idx + 1).trim();
- if (!name) continue;
- cookies.push({ name, value, url: 'https://creator.douyin.com' });
- }
- return cookies;
- }
- async function createBrowserForAccount(proxy: ProxyConfig | null): Promise<{ browser: Browser; shouldClose: boolean }> {
- // 静默同步:默认一律 headless,不弹窗
- // 只有在“引导登录/验证”时(DY_STORAGE_STATE_BOOTSTRAP=1 且 DY_IMPORT_HEADLESS=0)才允许 headful
- const allowHeadfulForBootstrap = process.env.DY_STORAGE_STATE_BOOTSTRAP === '1' && process.env.DY_IMPORT_HEADLESS === '0';
- const headless = !allowHeadfulForBootstrap;
- if (proxy?.enabled) {
- const server = `${proxy.type}://${proxy.host}:${proxy.port}`;
- const browser = await launchBrowser({
- headless,
- proxy: {
- server,
- username: proxy.username,
- password: proxy.password,
- },
- args: ['--no-sandbox', '--disable-setuid-sandbox', '--disable-dev-shm-usage', '--disable-gpu', '--window-size=1920,1080'],
- });
- return { browser, shouldClose: true };
- }
- const browser = await BrowserManager.getBrowser({ headless });
- return { browser, shouldClose: false };
- }
- type DashboardMetricTrendPoint = {
- date_time?: string; // YYYYMMDD
- value?: number;
- douyin_value?: number;
- xigua_value?: number;
- yumme_value?: number;
- change_rate?: number;
- };
- type DashboardMetricItem = {
- english_metric_name?: string;
- metric_name?: string;
- metric_value?: number;
- trends?: DashboardMetricTrendPoint[];
- };
- type DashboardResponse = {
- status_code?: number;
- status_msg?: string;
- metrics?: DashboardMetricItem[];
- };
- function parseYmdCompactToDate(ymd: unknown): Date | null {
- const s = String(ymd || '').trim();
- if (!/^\d{8}$/.test(s)) return null;
- const yyyy = Number(s.slice(0, 4));
- const mm = Number(s.slice(4, 6));
- const dd = Number(s.slice(6, 8));
- if (!yyyy || !mm || !dd) return null;
- const d = new Date(yyyy, mm - 1, dd);
- d.setHours(0, 0, 0, 0);
- return d;
- }
- function pickTrendValue(pt: DashboardMetricTrendPoint): number | undefined {
- // 优先使用聚合 value;若不存在则兜底 douyin_value
- const v =
- typeof pt?.value === 'number'
- ? pt.value
- : typeof pt?.douyin_value === 'number'
- ? pt.douyin_value
- : undefined;
- if (!Number.isFinite(v as number)) return undefined;
- return v as number;
- }
- export class DouyinAccountOverviewImportService {
- private accountRepository = AppDataSource.getRepository(PlatformAccount);
- private userDayStatisticsService = new UserDayStatisticsService();
- private stateDir = path.resolve(process.cwd(), 'tmp', 'douyin-storage-state');
- private getStatePath(accountId: number) {
- return path.join(this.stateDir, `${accountId}.json`);
- }
- private async ensureStorageState(account: PlatformAccount, cookies: PlaywrightCookie[]): Promise<string | null> {
- const statePath = this.getStatePath(account.id);
- try {
- await fs.access(statePath);
- return statePath;
- } catch {
- // no state
- }
- // 需要你在弹出的浏览器里完成一次登录/验证,然后脚本会自动保存 storageState
- // 启用方式:DY_IMPORT_HEADLESS=0 且 DY_STORAGE_STATE_BOOTSTRAP=1
- if (!(process.env.DY_IMPORT_HEADLESS === '0' && process.env.DY_STORAGE_STATE_BOOTSTRAP === '1')) {
- return null;
- }
- await ensureDir(this.stateDir);
- logger.warn(
- `[DY Import] No storageState for accountId=${account.id}. Bootstrapping... 请在弹出的浏览器中完成登录/验证。`
- );
- const { browser, shouldClose } = await createBrowserForAccount(account.proxyConfig);
- try {
- const context = await browser.newContext({
- viewport: { width: 1920, height: 1080 },
- locale: 'zh-CN',
- timezoneId: 'Asia/Shanghai',
- });
- await context.addCookies(cookies as any);
- const page = await context.newPage();
- await page.goto('https://creator.douyin.com/creator-micro/data-center/operation', {
- waitUntil: 'domcontentloaded',
- });
- // 最长等 5 分钟:让你手动完成登录/滑块/短信等
- await page
- .waitForFunction(() => {
- const t = document.body?.innerText || '';
- return t.includes('数据中心') || t.includes('账号总览') || t.includes('短视频');
- }, { timeout: 5 * 60_000 })
- .catch(() => undefined);
- await context.storageState({ path: statePath });
- logger.info(`[DY Import] storageState saved: ${statePath}`);
- await context.close();
- return statePath;
- } finally {
- if (shouldClose) await browser.close().catch(() => undefined);
- }
- }
- /**
- * 统一入口:定时任务与添加账号均调用此方法,执行“账号总览-短视频-数据表现-近30天”
- */
- static async runDailyImport(): Promise<void> {
- const svc = new DouyinAccountOverviewImportService();
- await svc.runDailyImportForAllDouyinAccounts();
- }
- /**
- * 单账号入口:仅为指定抖音账号执行近30天账号总览导入(用于账号从失效恢复为 active 时补数)
- */
- static async runDailyImportForAccount(accountId: number): Promise<void> {
- const svc = new DouyinAccountOverviewImportService();
- const account = await svc.accountRepository.findOne({
- where: { id: accountId, platform: 'douyin' as any },
- });
- if (!account) {
- throw new Error(`未找到抖音账号 id=${accountId}`);
- }
- await svc.importAccountLast30Days(account);
- }
- /**
- * 为所有抖音账号导出“账号总览-短视频-数据表现-近30天”并导入 user_day_statistics
- */
- async runDailyImportForAllDouyinAccounts(): Promise<void> {
- const accounts = await this.accountRepository.find({
- where: { platform: 'douyin' as any },
- });
- logger.info(`[DY Import] Start. total_accounts=${accounts.length}`);
- for (const account of accounts) {
- try {
- await this.importAccountLast30Days(account);
- } catch (e) {
- logger.error(
- `[DY Import] Account failed. accountId=${account.id} name=${account.accountName || ''}`,
- e
- );
- }
- }
- logger.info('[DY Import] Done.');
- }
- /**
- * 单账号:导出 Excel → 解析 → 入库 → 删除文件
- */
- async importAccountLast30Days(account: PlatformAccount, isRetry = false): Promise<void> {
- const cookies = parseCookiesFromAccount(account.cookieData);
- if (!cookies.length) {
- throw new Error('cookieData 为空或无法解析');
- }
- const { browser, shouldClose } = await createBrowserForAccount(account.proxyConfig);
- try {
- const statePath = await this.ensureStorageState(account, cookies);
- const context = await browser.newContext({
- acceptDownloads: true,
- viewport: { width: 1920, height: 1080 },
- locale: 'zh-CN',
- timezoneId: 'Asia/Shanghai',
- userAgent:
- 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36',
- ...(statePath ? { storageState: statePath } : {}),
- });
- context.setDefaultTimeout(60_000);
- // 如果没 state,就退回 cookie-only(可能导出为 0)
- if (!statePath) {
- await context.addCookies(cookies as any);
- }
- const page = await context.newPage();
- logger.info(`[DY Import] accountId=${account.id} goto data-center...`);
- await page.goto('https://creator.douyin.com/creator-micro/data-center/operation', {
- waitUntil: 'domcontentloaded',
- });
- await page.waitForTimeout(1500);
- if (page.url().includes('login')) {
- // 第一次检测到登录失效时,尝试刷新账号
- if (!isRetry) {
- logger.info(`[DY Import] Login expired detected for account ${account.id}, attempting to refresh...`);
- await context.close();
- if (shouldClose) await browser.close();
-
- try {
- const accountService = new AccountService();
- const refreshResult = await accountService.refreshAccount(account.userId, account.id);
-
- if (refreshResult.needReLogin) {
- // 刷新后仍需要重新登录,走原先的失效流程
- logger.warn(`[DY Import] Account ${account.id} refresh failed, still needs re-login`);
- throw new Error('未登录/需要重新登录(跳转到 login)');
- }
-
- // 刷新成功,重新获取账号信息并重试导入
- logger.info(`[DY Import] Account ${account.id} refreshed successfully, retrying import...`);
- const refreshedAccount = await this.accountRepository.findOne({ where: { id: account.id } });
- if (!refreshedAccount) {
- throw new Error('账号刷新后未找到');
- }
-
- // 递归调用,标记为重试
- return await this.importAccountLast30Days(refreshedAccount, true);
- } catch (refreshError) {
- logger.error(`[DY Import] Account ${account.id} refresh failed:`, refreshError);
- throw new Error('未登录/需要重新登录(跳转到 login)');
- }
- } else {
- // 已经是重试了,不再尝试刷新
- throw new Error('未登录/需要重新登录(跳转到 login)');
- }
- }
- // 检测“暂无访问权限 / 权限申请中 / 暂无数据”提示:标记账号 expired + 推送提示
- const bodyText = (await page.textContent('body').catch(() => '')) || '';
- if (
- bodyText.includes('暂无访问权限') ||
- bodyText.includes('权限申请中') ||
- bodyText.includes('暂无数据权限') ||
- bodyText.includes('暂无数据,请稍后再试')
- ) {
- await this.accountRepository.update(account.id, { status: 'expired' as any });
- wsManager.sendToUser(account.userId, WS_EVENTS.ACCOUNT_UPDATED, {
- account: { id: account.id, status: 'expired', platform: 'douyin' },
- });
- wsManager.sendToUser(account.userId, WS_EVENTS.SYSTEM_MESSAGE, {
- level: 'warning',
- message: `抖音账号「${account.accountName || account.accountId || account.id}」暂无数据看板访问权限,请到抖音创作者中心申请数据权限(通过后一般次日生效)。`,
- platform: 'douyin',
- accountId: account.id,
- });
- throw new Error('抖音数据看板暂无访问权限/申请中,已标记 expired 并通知用户');
- }
- // 已直达账号总览页(data-center/operation),无需再点「数据中心/账号总览」,直接点「短视频」和「近30天」
- await page.waitForTimeout(500);
- logger.info(`[DY Import] accountId=${account.id} on 账号总览页, click 短视频 tab (#semiTabaweme)...`);
- const shortVideoById = page.locator('#semiTabaweme');
- if ((await shortVideoById.count().catch(() => 0)) > 0) {
- await shortVideoById.first().click();
- } else {
- const shortVideoCandidates = ['短视频', '短视频数据'];
- let shortVideoClicked = false;
- for (const text of shortVideoCandidates) {
- const loc = page.getByText(text, { exact: false }).first();
- if ((await loc.count().catch(() => 0)) > 0) {
- await loc.click().catch(() => undefined);
- shortVideoClicked = true;
- break;
- }
- }
- if (!shortVideoClicked) {
- throw new Error('页面上未找到「短视频」入口,请确认抖音创作者后台是否改版');
- }
- }
- // 切换“近30天”(优先用 ID #addon-aoc08fi,兜底文案)
- await page.waitForTimeout(500);
- logger.info(`[DY Import] accountId=${account.id} click 近30天 (#addon-aoc08fi)...`);
- const last30DaysById = page.locator('#addon-aoc08fi');
- if ((await last30DaysById.count().catch(() => 0)) > 0) {
- await last30DaysById.first().click();
- } else {
- await page.getByText(/近\d+天?/).first().click().catch(() => undefined);
- await page.getByText('近30天', { exact: true }).click();
- }
- await page.waitForTimeout(1200);
- let totalInserted = 0;
- let totalUpdated = 0;
- const apiUrl = 'https://creator.douyin.com/janus/douyin/creator/data/overview/dashboard';
- logger.info(`[DY Import] accountId=${account.id} fetch dashboard (POST recent_days=30)...`);
- // 优先监听页面自身是否会发起该请求;若没有,则在页面上下文里手动 fetch(浏览器自动带 cookie)
- const responsePromise = page
- .waitForResponse(
- (res) => res.request().method() === 'POST' && res.url().includes('/janus/douyin/creator/data/overview/dashboard'),
- { timeout: 8000 }
- )
- .catch(() => null);
- const evalPromise = page.evaluate(async (url) => {
- try {
- const r = await fetch(url, {
- method: 'POST',
- credentials: 'include',
- headers: {
- accept: 'application/json, text/plain, */*',
- 'content-type': 'application/json',
- },
- body: JSON.stringify({ recent_days: 30 }),
- });
- const json = await r.json().catch(() => null);
- return json;
- } catch (e: any) {
- return { status_code: -1, status_msg: String(e?.message || e) };
- }
- }, apiUrl);
- const [res, evalJson] = await Promise.all([responsePromise, evalPromise]);
- const body = ((res ? await res.json().catch(() => null) : null) ?? evalJson ?? null) as DashboardResponse | null;
- if (!body || typeof body !== 'object') {
- throw new Error('overview/dashboard 响应不是 JSON');
- }
- if (Number(body.status_code) !== 0) {
- throw new Error(`overview/dashboard 返回非成功: code=${body.status_code} msg=${body.status_msg || ''}`);
- }
- const metrics = Array.isArray(body.metrics) ? body.metrics : [];
- if (!metrics.length) {
- logger.warn(`[DY Import] dashboard metrics empty. accountId=${account.id}`);
- }
- const mergedDays = new Map<string, { recordDate: Date } & Record<string, any>>();
- const setDay = (d: Date) => {
- const key = `${d.getFullYear()}-${String(d.getMonth() + 1).padStart(2, '0')}-${String(d.getDate()).padStart(2, '0')}`;
- if (!mergedDays.has(key)) mergedDays.set(key, { recordDate: d });
- return mergedDays.get(key)!;
- };
- for (const m of metrics) {
- const en = String(m?.english_metric_name || '').trim();
- const trends = Array.isArray(m?.trends) ? m.trends : [];
- if (!en || !trends.length) continue;
- for (const pt of trends) {
- const d = parseYmdCompactToDate(pt?.date_time);
- if (!d) continue;
- const obj = setDay(d);
- const v = pickTrendValue(pt);
- if (v === undefined) continue;
- // 显式排除:主页访问 / 取关粉丝(库里没有对应字段)
- if (en === 'homepage_view_cnt' || en === 'cancel_fans_cnt') continue;
- if (en === 'total_fans_cnt') (obj as any).fansCount = Math.round(v);
- else if (en === 'play_cnt') (obj as any).playCount = Math.round(v);
- else if (en === 'digg_cnt') (obj as any).likeCount = Math.round(v);
- else if (en === 'comment_cnt') (obj as any).commentCount = Math.round(v);
- else if (en === 'share_count') (obj as any).shareCount = Math.round(v);
- else if (en === 'net_fans_cnt') (obj as any).fansIncrease = Math.round(v);
- else if (en === 'cover_click_ratio') {
- const s = toRatePercentStringFromValue(v);
- if (s != null) (obj as any).coverClickRate = s;
- }
- }
- }
- // 合并完成后统一入库(避免同一天多次 update)
- for (const v of mergedDays.values()) {
- const { recordDate, ...patch } = v;
- const r = await this.userDayStatisticsService.saveStatisticsForDate(account.id, recordDate, patch);
- totalInserted += r.inserted;
- totalUpdated += r.updated;
- }
- logger.info(
- `[DY Import] short-video imported. accountId=${account.id} days=${mergedDays.size} inserted=${totalInserted} updated=${totalUpdated}`
- );
- await context.close();
- } finally {
- if (shouldClose) {
- await browser.close().catch(() => undefined);
- }
- }
- }
- }
|