hyreal-gen-pic/server/routes/crawl.js

'use strict';

// 门店采集路由：店铺清单、抓取任务控制、任务/系统日志

const fs = require('fs');
const { sendJson, readBody } = require('../util.js');

function handleShopsManual(ctx, body, res) {
  const lines = String((body && body.text) || '').split(/\r?\n/);
  const parsed = [];
  for (const line of lines) {
    const t = line.trim();
    if (!t) continue;
    const parts = t.split(/[\s,，]+/).filter(Boolean);
    const idPart = parts.find((p) => /^\d+$/.test(p));
    if (!idPart) continue;
    const shopId = Number(idPart);
    const shopName = parts.filter((p) => p !== idPart).join(' ') || String(shopId);
    parsed.push({ shopId, shopName, source: 'manual' });
  }
  const previousShops = ctx.shops.getList();
  const mergedShops = ctx.shops.mergeShops(parsed);
  try { ctx.shops.saveManualShops(mergedShops); ctx.shops.setList(mergedShops); }
  catch (e) { ctx.shops.setList(previousShops); throw e; }
  sendJson(res, 200, { shops: ctx.shops.getList(), addedCount: parsed.length });
}

function handleShopsList(ctx, root, res) {
  if (root && !ctx.workspace.isWorkspaceRoot(root)) return sendJson(res, 403, { error: '仅允许访问当前输出根目录' });
  const completed = root && fs.existsSync(root) ? ctx.crawl.readCompletionState(root) : {};
  sendJson(res, 200, {
    shops: ctx.shops.getList().map((shop) => ({ ...shop, completed: completed[String(shop.shopId)] || null })),
  });
}

function register(app, ctx) {
  app.get('/api/shops', async ({ res, query }) => {
    // 前端把 root 做 base64url 编码传入，避免中文路径在 URL 里出问题
    const encodedRoot = query.root;
    const root = encodedRoot ? Buffer.from(encodedRoot, 'base64url').toString('utf8') : '';
    return handleShopsList(ctx, root, res);
  });

  app.post('/api/shops/manual', async ({ req, res }) => {
    const body = await readBody(req);
    return handleShopsManual(ctx, body, res);
  });

  app.post('/api/crawl/start', async ({ req, res }) => {
    const body = await readBody(req);
    return ctx.crawl.startCrawl(body, res);
  });

  app.post('/api/crawl/stop', async ({ res }) => {
    return ctx.crawl.stopCrawl(res);
  });

  app.get('/api/crawl/status', async ({ res }) => {
    sendJson(res, 200, ctx.crawl.getState());
  });

  // 采集通道探测：客户端调试端口是否就绪 + Cookie 是否已配置（采集页提示用）
  app.get('/api/crawl/channel', async ({ res }) => {
    const kit = require('../../napos/kit.js');
    const p = await kit.cdp.probe().catch(() => ({ ok: false }));
    sendJson(res, 200, {
      channel: ctx.config.crawlChannel || 'auto',
      clientUsable: !!p.ok,
      clientBrowser: (p.ok && p.version && p.version.Browser) || null,
      cookieConfigured: !!ctx.session.cookie && !!ctx.session.getKsid(),
    });
  });

  app.get('/api/logs', async ({ res }) => {
    sendJson(res, 200, { logs: ctx.logs.mergeLogs(ctx.crawl.getLog()) });
  });

  app.post('/api/logs/clear', async ({ res }) => {
    ctx.logs.clearAll(ctx.crawl.getLog());
    ctx.logs.pushSystemLog('日志已清空', 'warn');
    return sendJson(res, 200, { ok: true });
  });
}

module.exports = { register };