tihao-ai/claude-code-tihao-sourcing/scripts/brief-fixture-smoke.js

#!/usr/bin/env node
const fs = require('fs');
const path = require('path');
const { parseBrief } = require('../mcp/src/features/tihao-sourcing/brief-parser');

function main() {
  const repoRoot = path.resolve(__dirname, '..', '..');
  const docsDir = path.join(repoRoot, 'docs');

  const dhaBrief = path.join(repoRoot, 'dha_brief.xlsx');
  if (fs.existsSync(dhaBrief)) {
    const parsed = parseBrief({ brief: dhaBrief });
    assert(parsed.brand === 'DHA', 'DHA xlsx should parse product as brand');
    assert(parsed.platforms.length === 1 && parsed.platforms[0] === 'xiaohongshu', 'DHA xlsx should parse Xiaohongshu platform');
    assert(parsed.keywords.includes('DHA') && parsed.keywords.includes('营养师'), 'DHA xlsx should parse maternal nutrition keywords');
    assert(parsed.performanceRequirements.includes('CPE≤3'), 'DHA xlsx should parse CPE requirement');
    assert(parsed.referenceLinks.length === 4, 'DHA xlsx should parse 4 reference account links');
    assert(parsed.referenceStyleAnchors.length >= 4, 'DHA xlsx should build reference style anchors');
    assert(parsed.referenceSignals.includes('干货科普') && parsed.referenceSignals.includes('618合集'), 'DHA xlsx should parse reference signals');
  }

  const markdownBrief = findBriefFile(docsDir, '.md');
  if (markdownBrief) {
    const parsed = parseBrief({ brief: markdownBrief });
    assert(parsed.brand.includes('花漾肌研'), 'markdown brief should parse brand');
    assert(parsed.targetCount === 32, 'markdown brief should parse target count sum');
    assert(parsed.platformTargets.xiaohongshu === 20, 'markdown brief should parse Xiaohongshu target');
    assert(parsed.platformTargets.douyin === 8, 'markdown brief should parse Douyin target');
    assert(parsed.platformTargets.bilibili === 4, 'markdown brief should parse Bilibili target');
    assert(parsed.fanRange.min === 20000 && parsed.fanRange.max === 500000, 'markdown brief should parse fan range');
    assert(parsed.referenceSignals.includes('素人感'), 'markdown brief should parse reference style preference');
  }

  const htmlXlsBrief = findBriefFile(docsDir, '.xls');
  if (htmlXlsBrief) {
    const parsed = parseBrief({ brief: htmlXlsBrief });
    assert(parsed.brand.includes('花漾肌研'), 'html xls brief should parse brand from key-value cells');
    assert(parsed.targetCount === 32, 'html xls brief should parse target count sum');
    assert(parsed.platformTargets.xiaohongshu === 20, 'html xls brief should parse Xiaohongshu target');
    assert(parsed.platformTargets.douyin === 8, 'html xls brief should parse Douyin target');
    assert(parsed.platformTargets.bilibili === 4, 'html xls brief should parse Bilibili target');
    assert(parsed.fanRange.min === 20000 && parsed.fanRange.max === 500000, 'html xls brief should parse horizontal fan range table');
    assert(parsed.referenceSignals.includes('真实测评') && parsed.referenceSignals.includes('成分党'), 'html xls brief should parse reference keyword signals');
  }

  assertNaturalLanguageBriefCoverage();

  console.log('brief fixture smoke ok');
}

function findBriefFile(dir, ext) {
  if (!fs.existsSync(dir)) return '';
  return fs.readdirSync(dir)
    .map(name => path.join(dir, name))
    .find(file => fs.statSync(file).isFile() && path.basename(file).includes('Brief') && file.endsWith(ext)) || '';
}

function assertNaturalLanguageBriefCoverage() {
  const cases = [
    {
      name: 'skincare',
      text: [
        '客户名称：明澈肌研',
        '产品：敏感肌屏障修护精华',
        '平台与人数：小红书 6 人，抖音 2 人',
        '关键词：敏感肌、修护、成分、真实测评',
        '粉丝范围：2万-50万',
        '预算范围：3000-18000',
        '偏好真实体验、素人感，不要硬广。'
      ].join('\n'),
      assert(parsed) {
        assert(parsed.brand === '明澈肌研', 'skincare should parse brand');
        assert(parsed.category === '护肤美妆', 'skincare should infer beauty category');
        assert(parsed.platformTargets.xiaohongshu === 6 && parsed.platformTargets.douyin === 2, 'skincare should parse platform targets');
        assert(parsed.fanRange.min === 20000 && parsed.fanRange.max === 500000, 'skincare should parse fan range');
        assert(parsed.keywords.includes('敏感肌') && parsed.keywords.includes('修护'), 'skincare should parse skincare keywords');
      }
    },
    {
      name: 'real-douyin-beauty-brief',
      text: [
        '平台：抖音',
        '账号类型：美妆护肤',
        '粉丝要求：1k-10w',
        '单个预算：2000内',
        '数据要求：主页互动数据200+',
        '备注：产品是客单价比较高的面膜，想找调性高的美妆护肤博主，想要主页之前合作过大牌护肤品的',
        '',
        '参考账号：',
        'https://www.douyin.com/user/MS4wLjABAAAAZ31qrM27_cxcps8cc2sGKlnUWIOUDuVmaGB7BhNEpYAcZfhxjgicQ3VVXfRNgcWv'
      ].join('\n'),
      assert(parsed) {
        assert(parsed.category === '护肤美妆', 'real douyin beauty brief should infer beauty category');
        assert(parsed.platforms.length === 1 && parsed.platforms[0] === 'douyin', 'real douyin beauty brief should parse Douyin only');
        assert(parsed.fanRange.min === 1000 && parsed.fanRange.max === 100000, 'real douyin beauty brief should parse 1k-10w fan range');
        assert(parsed.budgetRange.min === 0 && parsed.budgetRange.max === 2000, 'real douyin beauty brief should parse 2000 upper-bound budget');
        assert(parsed.performanceRequirements.includes('主页互动数据≥200'), 'real douyin beauty brief should parse homepage interaction requirement');
        assert(parsed.keywords.includes('护肤') && parsed.keywords.includes('美妆') && parsed.keywords.includes('面膜'), 'real douyin beauty brief should keep beauty keywords');
        assert(parsed.referenceLinks.length === 1, 'real douyin beauty brief should parse reference account link');
      }
    },
    {
      name: 'healthy-snack',
      text: [
        '品牌：谷物星球',
        '本次推广产品：高蛋白低脂燕麦棒',
        '目标：找小红书 5 个、抖音 3 个健身/控糖/办公室零食博主',
        '粉丝数范围：1万-30万',
        '单篇预算：1000-8000',
        '风格偏生活化开箱、真实吃法、成分党，不要医美和夸大减肥。'
      ].join('\n'),
      assert(parsed) {
        assert(parsed.brand === '谷物星球', 'healthy snack should parse brand');
        assert(parsed.category === '食品餐饮', 'healthy snack should infer food category instead of beauty');
        assert(parsed.targetCount === 8, 'healthy snack should parse platform target sum');
        assert(parsed.fanRange.min === 10000 && parsed.fanRange.max === 300000, 'healthy snack should parse fan range');
        assert(parsed.keywords.includes('控糖') && parsed.keywords.includes('办公室零食'), 'healthy snack should parse food creator keywords');
        assert(parsed.exclusions.includes('医美') && parsed.exclusions.includes('夸大减肥'), 'healthy snack should parse exclusions');
      }
    },
    {
      name: 'home-cleaning',
      text: [
        '客户：净住研究所',
        '产品：厨房重油污清洁喷雾',
        '需求人数：小红书 8 人',
        '要求：家居清洁、收纳、租房生活、妈妈人群，粉丝 5000-20万，预算 800-5000',
        '排除：硬广、低质搬运、争议账号。'
      ].join('\n'),
      assert(parsed) {
        assert(parsed.brand === '净住研究所', 'home cleaning should parse brand');
        assert(parsed.category === '家居生活', 'home cleaning should infer home category');
        assert(parsed.platforms.length === 1 && parsed.platforms[0] === 'xiaohongshu', 'home cleaning should parse Xiaohongshu only');
        assert(parsed.fanRange.min === 5000 && parsed.fanRange.max === 200000, 'home cleaning should not read budget as fan range');
        assert(parsed.budgetRange.min === 800 && parsed.budgetRange.max === 5000, 'home cleaning should parse budget on same line');
        assert(parsed.keywords.includes('清洁') && parsed.keywords.includes('收纳'), 'home cleaning should parse home keywords');
      }
    },
    {
      name: 'b2b-saas',
      text: [
        '品牌：FlowOps',
        '产品：企业项目管理 SaaS',
        '需求：找 B站 4 个、抖音 2 个效率工具/创业/企业管理类内容创作者',
        '粉丝：1万-40万',
        '预算：3000-20000',
        '偏好教程型、案例型、职场效率，不要纯娱乐账号。'
      ].join('\n'),
      assert(parsed) {
        assert(parsed.brand === 'FlowOps', 'B2B SaaS should parse brand');
        assert(parsed.category === 'B2B软件', 'B2B SaaS should infer software category');
        assert(parsed.platformTargets.bilibili === 4 && parsed.platformTargets.douyin === 2, 'B2B SaaS should parse Bilibili and Douyin targets');
        assert(parsed.keywords.includes('效率工具') && parsed.keywords.includes('企业管理'), 'B2B SaaS should parse software keywords');
        assert(parsed.exclusions.includes('纯娱乐') || parsed.exclusions.includes('纯娱乐账号'), 'B2B SaaS should parse entertainment exclusion');
      }
    },
    {
      name: 'local-food',
      text: [
        '品牌：山城小面加盟',
        '任务：找本地生活探店、餐饮创业类达人，重庆/成都优先',
        '平台：抖音 10 人',
        '粉丝 2万-80万，预算 2000-12000',
        '重点看同城转化、门店拍摄经验，排除泛娱乐流量号。'
      ].join('\n'),
      assert(parsed) {
        assert(parsed.brand === '山城小面加盟', 'local food should parse brand');
        assert(parsed.category === '食品餐饮', 'local food should infer food/local category');
        assert(parsed.platforms.length === 1 && parsed.platforms[0] === 'douyin', 'local food should parse Douyin only');
        assert(parsed.targetCount === 10, 'local food should parse target count');
        assert(parsed.fanRange.min === 20000 && parsed.fanRange.max === 800000, 'local food should parse fan range and ignore budget');
        assert(parsed.region.includes('重庆') && parsed.region.includes('成都'), 'local food should parse city priority');
        assert(parsed.keywords.includes('本地生活') && parsed.keywords.includes('同城转化'), 'local food should parse local conversion keywords');
      }
    },
    {
      name: 'fashion-product-chat-refinement',
      text: [
        '客户：南栖衣橱',
        'Brief：找精致生活/时尚穿搭类账号，小红书 4 人',
        '粉丝：1万-20万，预算：300-3000',
        '数据要求：CPE≤3，CP1≤20，CPM≤80',
        '不要纯自拍。'
      ].join('\n'),
      input: {
        productIntro: '产品是港风通勤女装，偏成熟风、轻熟高级感，不是泛生活方式。',
        chatRecords: [
          '客户说要看最近一个月有更新，平均点赞不要百赞以下。',
          '如果多篇都是同一个人评论，怀疑刷评，先不要强推。'
        ],
        referenceAccounts: ['港风穿搭Alice']
      },
      assert(parsed) {
        assert(parsed.brand === '南栖衣橱', 'fashion should parse brand');
        assert(parsed.category === '服饰穿搭', 'fashion product/chat should refine broad lifestyle into apparel category');
        assert(parsed.contextSources.product && parsed.contextSources.chat && parsed.contextSources.reference, 'fashion should mark product/chat/reference context sources');
        assert(parsed.platformTargets.xiaohongshu === 4, 'fashion should parse Xiaohongshu target');
        assert(parsed.keywords.includes('港风') && parsed.keywords.includes('成熟风'), 'fashion should extract subtype keywords from product/chat context');
        assert(parsed.keywords.includes('时尚穿搭') || parsed.keywords.includes('穿搭'), 'fashion should keep broad fashion keyword');
        assert(parsed.exclusions.includes('纯自拍') && !parsed.exclusions.includes('穿搭'), 'fashion should not treat wearing/fashion as an exclusion');
        assert(parsed.performanceRequirements.includes('CPE≤3') && parsed.performanceRequirements.includes('CP1≤20') && parsed.performanceRequirements.includes('CPM≤80'), 'fashion should parse cost metrics');
        assert(parsed.performanceRequirements.includes('近30天平均点赞≥100'), 'fashion should turn low-like warning into average-like requirement');
        assert(parsed.performanceRequirements.includes('复核数据真实性/重复评论风险'), 'fashion should parse repeated-comment authenticity review');
      }
    }
  ];

  for (const item of cases) {
    const parsed = parseBrief({ briefText: item.text, ...(item.input || {}) });
    item.assert(parsed);
  }
}

function assert(condition, message) {
  if (!condition) throw new Error(message);
}

main();