tihao-ai/claude-code-tihao-sourcing/scripts/candidate-pool-smoke.js

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

function main() {
  const root = path.resolve(__dirname, '..');
  const outputRoot = fs.mkdtempSync(path.join(os.tmpdir(), 'tihao-candidate-pool-'));

  const fullPool = buildSourcingReport({
    criteria: {
      briefId: '候选池验收',
      brand: '候选池验收',
      targetCount: 4,
      platforms: ['xiaohongshu']
    },
    candidates: sampleCandidates(),
    collectionMode: 'sample',
    outputDir: path.join(outputRoot, 'full')
  });

  assert(fullPool.summary.reviewPoolTarget === 6, 'targetCount=4 should require a 1.5x review pool of 6');
  assert(fullPool.summary.clientReadyCount === 6, 'software table should output 6 reviewable candidates');
  assert(fullPool.summary.duplicateCount === 1, 'duplicate homepage URL should be reported');
  assertNoDuplicateSoftwareRows(path.join(outputRoot, 'full', 'tihao-sourcing-client-list.csv'));

  const shortPool = buildSourcingReport({
    criteria: {
      briefId: '候选池不足验收',
      brand: '候选池不足验收',
      targetCount: 10,
      platforms: ['xiaohongshu']
    },
    candidates: sampleCandidates(),
    collectionMode: 'sample',
    outputDir: path.join(outputRoot, 'short')
  });

  const shortJson = JSON.parse(fs.readFileSync(path.join(outputRoot, 'short', 'tihao-sourcing-result.json'), 'utf8'));
  assert(shortPool.summary.reviewPoolTarget === 15, 'targetCount=10 should require a 1.5x review pool of 15');
  assert(shortPool.summary.clientReadyCount === 7, 'short software table should keep all available non-excluded candidates');
  assert(shortJson.warnings.some(item => item.includes('1.5 倍候选池 15 位')), 'short pool should warn about 1.5x candidate pool gap');
  assert(shortPool.summary.calibrationQuestions.some(item => item.includes('1.5 倍候选池')), 'calibration question should mention 1.5x pool');

  console.log(JSON.stringify({
    ok: true,
    outputRoot,
    fullPool: fullPool.summary,
    shortPool: shortPool.summary
  }, null, 2));
}

function sampleCandidates() {
  return [
    creator('账号A', 88, 'https://example.com/a'),
    creator('账号A-高分重复', 93, 'https://example.com/a'),
    creator('账号B', 87, 'https://example.com/b'),
    creator('账号C', 86, 'https://example.com/c'),
    creator('账号D', 85, 'https://example.com/d'),
    creator('账号E', 84, 'https://example.com/e'),
    creator('账号F', 83, 'https://example.com/f'),
    creator('账号G', 82, 'https://example.com/g')
  ];
}

function creator(displayName, score, profileUrl) {
  return {
    platform: 'xiaohongshu',
    displayName,
    profileUrl,
    score,
    briefFitScore: score,
    referenceStyleFitScore: 80,
    recentContentFitScore: 78,
    visualQualityScore: 76,
    toneConsistencyScore: 79,
    recommendStatus: score >= 86 ? '强推荐' : '备选',
    recommendReason: `${displayName} 命中 Brief 核心人群和参考风格。`,
    riskNote: '需商务复核报价有效期。'
  };
}

function assertNoDuplicateSoftwareRows(csvPath) {
  const lines = fs.readFileSync(csvPath, 'utf8').replace(/^\uFEFF/, '').trim().split(/\r?\n/);
  const header = lines[0].split(',');
  const rows = lines.slice(1).map(line => parseCsvLine(line, header));
  const keys = new Set();
  for (const row of rows) {
    const briefId = pick(row, ['brief编号']);
    const platform = pick(row, ['平台']);
    const profileUrl = pick(row, ['主页链接']);
    const displayName = pick(row, ['账号名称', '博主名称']);
    const key = profileUrl
      ? `${briefId}|${platform}|${profileUrl.toLowerCase()}`
      : `${briefId}|${platform}|${displayName.toLowerCase()}`;
    assert(!keys.has(key), `duplicate software row key: ${key}`);
    keys.add(key);
  }
  assert(rows.map(row => Number(pick(row, ['序号', '排名']))).join(',') === '1,2,3,4,5,6', 'software rows should be reranked continuously');
}

function pick(row, keys) {
  for (const key of keys) {
    if (row[key] !== undefined && row[key] !== '') return row[key];
  }
  return '';
}

function parseCsvLine(line, header) {
  const cells = [];
  let current = '';
  let quoted = false;
  for (let i = 0; i < line.length; i += 1) {
    const ch = line[i];
    if (ch === '"' && quoted && line[i + 1] === '"') {
      current += '"';
      i += 1;
    } else if (ch === '"') {
      quoted = !quoted;
    } else if (ch === ',' && !quoted) {
      cells.push(current);
      current = '';
    } else {
      current += ch;
    }
  }
  cells.push(current);
  return Object.fromEntries(header.map((key, index) => [key, cells[index] || '']));
}

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

main();