tihao-ai/claude-code-tihao-sourcing/scripts/optimization-pipeline.js

#!/usr/bin/env node
const fs = require('fs');
const path = require('path');
const { spawnSync } = require('child_process');

const ROOT = path.resolve(__dirname, '..');

function main() {
  const args = parseArgs(process.argv.slice(2));
  const outputRoot = path.resolve(args.output || process.env.TIHAO_PIPELINE_OUTPUT || path.join(ROOT, 'outputs', `optimization-pipeline-${Date.now()}`));
  const dryRun = Boolean(args.dryRun);
  const strict = Boolean(args.strict);
  fs.mkdirSync(outputRoot, { recursive: true });

  const context = buildContext(args, outputRoot);
  const steps = buildSteps(context);
  const results = [];
  for (const step of steps) {
    if (dryRun) {
      results.push({ id: step.id, status: 'planned', command: step.command.join(' '), output: step.output || '' });
      continue;
    }
    const result = runStep(step);
    results.push(result);
    if (strict && result.status !== 'pass') break;
  }

  const closureResult = results.find(item => item.id === 'proof-gap-closure');
  const closureSummary = closureResult?.status === 'pass' ? readJson(closureResult.output) : null;
  const proofGapClosureClaimable = isClaimableProofGapClosure(closureSummary);
  const readyForClaim = results.every(item => item.status === 'pass') &&
    Boolean(context.inputs.reviewCsv) &&
    proofGapClosureClaimable;
  const summary = {
    generatedAt: new Date().toISOString(),
    outputRoot,
    dryRun,
    strict,
    inputs: context.inputs,
    counts: countStatuses(results),
    readyForClaim,
    proofGapClosure: closureSummary ? {
      complete: Boolean(closureSummary.complete),
      openCount: Number(closureSummary.openCount || 0),
      closedCount: Number(closureSummary.closedCount || 0),
      businessGapCount: Number(closureSummary.businessGapCount || 0),
      claimable: proofGapClosureClaimable
    } : null,
    results,
    nextActions: buildNextActions(results, context, { dryRun })
  };
  const jsonPath = path.join(outputRoot, 'optimization-pipeline-summary.json');
  const reportPath = path.join(outputRoot, 'optimization-pipeline-report.md');
  fs.writeFileSync(jsonPath, JSON.stringify(summary, null, 2), 'utf8');
  fs.writeFileSync(reportPath, withBom(renderReport(summary)), 'utf8');
  console.log(JSON.stringify({
    outputRoot,
    json: jsonPath,
    report: reportPath,
    counts: summary.counts,
    readyForClaim: summary.readyForClaim,
    nextActionCount: summary.nextActions.length
  }, null, 2));
  if (strict && !summary.readyForClaim) process.exitCode = 2;
}

function buildContext(args, outputRoot) {
  const dataPack = resolveOptional(args.dataPack || process.env.TIHAO_DATA_PACK);
  const videoPack = resolveOptional(args.videoPack || process.env.TIHAO_VIDEO_PACK);
  const historyCsv = resolveOptional(args.historyCsv || process.env.TIHAO_HISTORY_CSV || (dataPack ? path.join(dataPack, 'history-data-template.csv') : ''));
  const historyDir = resolveOptional(args.historyDir || process.env.TIHAO_HISTORY_DATASET);
  const preflight = resolveOptional(args.preflight || process.env.TIHAO_PREFLIGHT_SUMMARY);
  const reviewCsv = resolveOptional(args.reviewCsv || process.env.TIHAO_REVIEW_CSV || (dataPack ? path.join(dataPack, 'manual-review-template.csv') : ''));
  const videoCsv = resolveOptional(args.videoCsv || process.env.TIHAO_VIDEO_CSV || (videoPack ? path.join(videoPack, 'video-resource-template.csv') : ''));
  const currentManualSupplementCount = args.currentManualSupplementCount || process.env.TIHAO_CURRENT_MANUAL_SUPPLEMENT_COUNT || '';
  const historyMinBriefs = args.historyMinBriefs || process.env.TIHAO_HISTORY_MIN_BRIEFS || '';
  const generatedHistoryDir = path.join(outputRoot, 'history-dataset');
  const intakeOutput = path.join(outputRoot, 'intake-readiness');
  const historyAuditOutput = path.join(outputRoot, 'history-audit');
  const videoReadinessOutput = path.join(outputRoot, 'video-resource-readiness');
  const reviewOutput = path.join(outputRoot, 'review-metrics');
  const customerEffectOutput = path.join(outputRoot, 'customer-effect-audit');
  const closureOutput = path.join(outputRoot, 'proof-gap-closure');
  const readinessOutput = path.join(outputRoot, 'long-run-readiness');
  const evidenceOutput = path.join(outputRoot, 'evidence-index');
  return {
    outputRoot,
    historyDatasetDir: historyCsv ? generatedHistoryDir : historyDir,
    historyAuditJson: path.join(historyAuditOutput, 'historical-dataset-audit.json'),
    intakeReadinessJson: path.join(intakeOutput, 'intake-readiness-summary.json'),
    videoReadinessJson: path.join(videoReadinessOutput, 'video-resource-readiness-summary.json'),
    inputs: {
      dataPack,
      videoPack,
      historyCsv,
      historyDir,
      videoCsv,
      preflight,
      reviewCsv,
      currentManualSupplementCount,
      historyMinBriefs
    },
    paths: {
      generatedHistoryDir,
      intakeOutput,
      historyAuditOutput,
      videoReadinessOutput,
      reviewOutput,
      customerEffectOutput,
      closureOutput,
      readinessOutput,
      evidenceOutput
    }
  };
}

function buildSteps(context) {
  const steps = [];
  if (context.inputs.dataPack || context.inputs.videoPack || (context.inputs.historyCsv && context.inputs.reviewCsv && context.inputs.videoCsv)) {
    const intakeArgs = ['npm.cmd', 'run', 'intake:readiness', '--', '--output', context.paths.intakeOutput];
    if (context.inputs.dataPack) intakeArgs.push('--data-pack', context.inputs.dataPack);
    else {
      intakeArgs.push('--history-csv', context.inputs.historyCsv);
      intakeArgs.push('--review-csv', context.inputs.reviewCsv);
    }
    if (context.inputs.videoPack) intakeArgs.push('--video-pack', context.inputs.videoPack);
    else intakeArgs.push('--video-csv', context.inputs.videoCsv);
    if (context.inputs.historyMinBriefs) intakeArgs.push('--min-briefs', context.inputs.historyMinBriefs);
    steps.push(step('intake-readiness', intakeArgs, context.intakeReadinessJson));
  }
  if (context.inputs.historyCsv) {
    steps.push(step('history-from-csv', ['npm.cmd', 'run', 'history:from-csv', '--', '--input', context.inputs.historyCsv, '--output', context.paths.generatedHistoryDir], context.paths.generatedHistoryDir));
  }
  if (context.historyDatasetDir) {
    const historyAuditArgs = ['npm.cmd', 'run', 'history:audit', '--', '--input', context.historyDatasetDir, '--output', context.paths.historyAuditOutput, '--strict'];
    if (context.inputs.historyMinBriefs) historyAuditArgs.push('--min-briefs', context.inputs.historyMinBriefs);
    steps.push(step('history-audit', historyAuditArgs, context.historyAuditJson));
  }
  if (context.inputs.videoCsv) {
    steps.push(step('video-resource-readiness', ['npm.cmd', 'run', 'video:resource-readiness', '--', '--input', context.inputs.videoCsv, '--output', context.paths.videoReadinessOutput, '--strict'], context.videoReadinessJson));
  }

  const readinessArgs = ['npm.cmd', 'run', 'longrun:readiness', '--', '--mode', context.inputs.reviewCsv ? 'customer-effect' : 'full-matrix', '--output', context.paths.readinessOutput];
  if (context.inputs.preflight) readinessArgs.push('--preflight', context.inputs.preflight);
  if (context.historyDatasetDir) readinessArgs.push('--history', context.historyAuditJson);
  if (context.inputs.dataPack || context.inputs.videoPack || context.inputs.videoCsv) readinessArgs.push('--intake-readiness', context.intakeReadinessJson);
  if (context.inputs.videoCsv) readinessArgs.push('--video-readiness', context.videoReadinessJson);
  steps.push(step('longrun-readiness', readinessArgs, path.join(context.paths.readinessOutput, 'long-run-readiness-summary.json')));

  if (context.inputs.reviewCsv) {
    steps.push(step('review-metrics', ['npm.cmd', 'run', 'review:metrics', '--', '--input', context.inputs.reviewCsv, '--output', context.paths.reviewOutput, '--strict'], path.join(context.paths.reviewOutput, 'review-metrics-summary.json')));
    const effectArgs = ['npm.cmd', 'run', 'customer-effect:audit', '--', '--review-csv', context.inputs.reviewCsv, '--history-audit', context.historyAuditJson, '--output', context.paths.customerEffectOutput, '--strict'];
    if (context.inputs.currentManualSupplementCount) effectArgs.push('--current-manual-supplement-count', context.inputs.currentManualSupplementCount);
    steps.push(step('customer-effect-audit', effectArgs, path.join(context.paths.customerEffectOutput, 'customer-effect-summary.json')));
  }

  const closureArgs = ['npm.cmd', 'run', 'proof-gap:closure', '--', '--output', context.paths.closureOutput];
  if (context.historyDatasetDir) closureArgs.push('--history-audit', context.historyAuditJson);
  if (context.inputs.dataPack || context.inputs.videoPack || context.inputs.videoCsv) closureArgs.push('--intake-readiness', context.intakeReadinessJson);
  if (context.inputs.videoCsv) closureArgs.push('--video-readiness', context.videoReadinessJson);
  if (context.inputs.reviewCsv) closureArgs.push('--customer-effect', path.join(context.paths.customerEffectOutput, 'customer-effect-summary.json'));
  steps.push(step('proof-gap-closure', closureArgs, path.join(context.paths.closureOutput, 'proof-gap-closure-summary.json')));

  steps.push(step('evidence-index', ['npm.cmd', 'run', 'evidence:index', '--', '--output', context.paths.evidenceOutput], path.join(context.paths.evidenceOutput, 'evidence-index-summary.json')));
  return steps;
}

function step(id, command, output) {
  return { id, command, output };
}

function runStep(step) {
  const [command, ...args] = step.command;
  const result = process.platform === 'win32'
    ? spawnSync('cmd.exe', ['/d', '/s', '/c', step.command.map(cmdQuote).join(' ')], { cwd: ROOT, encoding: 'utf8' })
    : spawnSync(command, args, { cwd: ROOT, encoding: 'utf8' });
  return {
    id: step.id,
    command: step.command.join(' '),
    output: step.output,
    status: result.status === 0 ? 'pass' : 'fail',
    exitCode: result.status,
    stdout: safeTail(result.stdout),
    stderr: safeTail(result.stderr)
  };
}

function cmdQuote(value) {
  const text = String(value || '');
  if (!/[()\s&|<>^"]/u.test(text)) return text;
  return `"${text.replace(/"/g, '\\"')}"`;
}

function renderReport(summary) {
  const lines = [
    '# 提号优化 Pipeline 报告',
    '',
    `- 生成时间：${summary.generatedAt}`,
    `- 输出目录：${summary.outputRoot}`,
    `- dryRun：${summary.dryRun ? '是' : '否'}`,
    `- strict：${summary.strict ? '是' : '否'}`,
    `- 是否可宣称业务效果：${summary.readyForClaim ? '是' : '否'}`,
    `- proofGapClosure.complete：${summary.proofGapClosure?.complete ? 'true' : 'false'}`,
    `- proofGapClosure.openCount：${summary.proofGapClosure?.openCount ?? 'unknown'}`,
    '',
    '## 输入',
    '',
    ...Object.entries(summary.inputs).map(([key, value]) => `- ${key}: ${value || '未提供'}`),
    '',
    '## 步骤',
    '',
    '| 步骤 | 状态 | 输出 | 命令 |',
    '| --- | --- | --- | --- |',
    ...summary.results.map(item => `| ${escapeCell(item.id)} | ${item.status} | ${escapeCell(item.output || '')} | ${escapeCell(item.command)} |`),
    '',
    '## 下一步动作',
    '',
    summary.nextActions.length
      ? '| 优先级 | 负责人 | 动作 | 触发步骤 | 执行命令 | 预期产物 | 通过标准 |\n| --- | --- | --- | --- | --- | --- | --- |\n' + summary.nextActions.map(item => `| ${item.priority} | ${escapeCell(item.owner)} | ${escapeCell(item.title)} | ${escapeCell(item.step)} | \`${escapeCell(item.command)}\` | ${escapeCell(item.expectedArtifact)} | ${escapeCell(item.acceptance)} |`).join('\n')
      : '- 暂无后续动作。',
    '',
    '## 边界',
    '',
    '- pipeline 只编排已有脚本，不自动写入 token，也不默认启动真实 live 长跑。',
    '- 没有客户选择、历史人工补号基线或本轮人工补号量时，不能宣称客户效果完成。',
    '- `readyForClaim=false` 时，只能说明链路入口或局部验收完成。'
  ];
  return lines.join('\n');
}

function buildNextActions(results, context, { dryRun }) {
  if (dryRun) {
    return [{
      priority: 1,
      owner: '技术/AI',
      title: '确认输入后执行真实 pipeline',
      step: 'dry-run',
      command: rerunCommandWithoutDryRun(context),
      expectedArtifact: path.join(context.outputRoot, 'optimization-pipeline-summary.json'),
      acceptance: 'dryRun=false，且关键步骤不再是 planned。'
    }];
  }
  const actions = [];
  for (const result of results) {
    if (result.status === 'pass') continue;
    actions.push(actionForFailedStep(result.id, context, result));
    if (context.inputs.currentManualSupplementCount && result.id === 'customer-effect-audit') continue;
  }
  if (!context.inputs.currentManualSupplementCount && context.inputs.reviewCsv) {
    actions.push({
      priority: 2,
      owner: '商务',
      title: '补本轮人工补号量',
      step: 'customer-effect-audit',
      command: '补齐 --current-manual-supplement-count <本轮人工补号量> 后重跑 optimization:pipeline',
      expectedArtifact: '<客户效果输出目录>/customer-effect-summary.json',
      acceptance: 'customer-effect:audit 输出 manualSupplementReductionRate，且 strict 审计通过。'
    });
  }
  if (!actions.length && !results.every(item => item.status === 'pass')) {
    actions.push({
      priority: 99,
      owner: '技术/AI',
      title: '检查 pipeline 未通过步骤',
      step: 'pipeline',
      command: '打开 optimization-pipeline-report.md，按首个 fail 步骤处理。',
      expectedArtifact: 'optimization-pipeline-summary.json',
      acceptance: '所有 pipeline 步骤 status=pass。'
    });
  }
  const closureResult = results.find(item => item.id === 'proof-gap-closure');
  const closureSummary = closureResult?.status === 'pass' ? readJson(closureResult.output) : null;
  if (!actions.length && closureSummary && !isClaimableProofGapClosure(closureSummary)) {
    actions.push({
      priority: 2,
      owner: '技术/AI',
      title: '关闭真实证明缺口',
      step: 'proof-gap-closure',
      command: '按 proof-gap-closure-report.md 补齐 open 项后重跑 optimization:pipeline。',
      expectedArtifact: closureResult.output || '<输出目录>/proof-gap-closure/proof-gap-closure-summary.json',
      acceptance: 'proof-gap:closure complete=true、openCount=0、closedCount>=5，且 businessGapCount>=5。'
    });
  }
  if (!actions.length && results.every(item => item.status === 'pass') && !context.inputs.reviewCsv) {
    actions.push({
      priority: 3,
      owner: '商务',
      title: '补人工复核 CSV 后验证客户效果',
      step: 'review-metrics',
      command: 'npm run optimization:pipeline -- --data-pack outputs\\data-intake-pack-latest --video-pack outputs\\video-intake-pack-latest --current-manual-supplement-count <本轮人工补号量> --output <输出目录> --strict',
      expectedArtifact: '<输出目录>/customer-effect-audit/customer-effect-summary.json',
      acceptance: 'review:metrics 和 customer-effect:audit 均通过。'
    });
  }
  return dedupeActions(actions).sort((a, b) => a.priority - b.priority);
}

function actionForFailedStep(stepId, context, result) {
  const base = {
    priority: 10,
    owner: '技术/AI',
    title: `修复 ${stepId}`,
    step: stepId,
    command: result.command,
    expectedArtifact: result.output || '',
    acceptance: `${stepId} status=pass。`
  };
  const map = {
    'intake-readiness': {
      priority: 1,
      owner: '商务/投放',
      title: '修复真实材料收集包',
      command: '按 intake-readiness-report.md 的问题修正 data-pack/video-pack 后重跑 optimization:pipeline。',
      acceptance: 'intake-readiness overallReady=true 且 failureCount=0。'
    },
    'history-from-csv': {
      priority: 2,
      owner: '商务',
      title: '修复历史数据 CSV',
      command: '检查 history-data-template.csv 表头、编码和必填字段后重跑 history:from-csv。',
      acceptance: 'history-dataset JSON 成功生成。'
    },
    'history-audit': {
      priority: 3,
      owner: '商务',
      title: '补齐历史数据审计字段',
      command: '补齐真实 Brief、参考链接、人工名单、客户选择、拒绝原因和历史人工补号量基线后重跑 history:audit --strict。',
      acceptance: 'historical-dataset-audit readyForCustomerEffectProof=true。'
    },
    'video-resource-readiness': {
      priority: 4,
      owner: '商务/投放',
      title: '补齐真实视频资源',
      command: '补齐真实参考视频、真实候选视频、视频 URL、封面、ASR、帧图或正文证据后重跑 video:resource-readiness --strict。',
      acceptance: 'video-resource-readiness readyForVideoAbPreflight=true。'
    },
    'longrun-readiness': {
      priority: 5,
      owner: '技术/AI',
      title: '修复长跑前门禁',
      command: '根据 long-run-readiness-report.md 补 live preflight、history audit、intake readiness、video readiness 或 proof gap 后重跑 longrun:readiness。',
      acceptance: 'longrun:readiness ready=true 且 fail=0。'
    },
    'review-metrics': {
      priority: 6,
      owner: '商务',
      title: '补齐人工复核标注',
      command: '补齐人工复核标签、客户选择和负样本归因后重跑 review:metrics --strict。',
      acceptance: 'review:metrics overallPass=true，负样本归因覆盖率=100%。'
    },
    'customer-effect-audit': {
      priority: 7,
      owner: '商务',
      title: '补齐客户效果证明',
      command: '补齐客户最终选择、历史人工补号量基线和本轮人工补号量后重跑 customer-effect:audit --strict。',
      acceptance: 'customer-effect:audit overallPass=true，客户选中率和人工补号减少率达标。'
    },
    'evidence-index': {
      priority: 20,
      owner: '技术/AI',
      title: '刷新证据台账',
      command: 'npm run evidence:index -- --output outputs\\evidence-index-latest',
      acceptance: 'evidence-index-summary.json 成功生成，且不把 smoke 当业务证明。'
    }
  };
  return { ...base, ...(map[stepId] || {}) };
}

function rerunCommandWithoutDryRun(context) {
  const parts = ['npm run optimization:pipeline --'];
  if (context.inputs.dataPack) parts.push(`--data-pack ${context.inputs.dataPack}`);
  if (context.inputs.videoPack) parts.push(`--video-pack ${context.inputs.videoPack}`);
  if (!context.inputs.dataPack && context.inputs.historyCsv) parts.push(`--history-csv ${context.inputs.historyCsv}`);
  if (!context.inputs.dataPack && context.inputs.reviewCsv) parts.push(`--review-csv ${context.inputs.reviewCsv}`);
  if (context.inputs.historyMinBriefs) parts.push(`--history-min-briefs ${context.inputs.historyMinBriefs}`);
  if (context.inputs.currentManualSupplementCount) parts.push(`--current-manual-supplement-count ${context.inputs.currentManualSupplementCount}`);
  parts.push(`--output ${context.outputRoot}`);
  if (context.strict) parts.push('--strict');
  return parts.join(' ');
}

function dedupeActions(actions) {
  const seen = new Set();
  return actions.filter(action => {
    const key = `${action.owner}|${action.title}|${action.step}`;
    if (seen.has(key)) return false;
    seen.add(key);
    return true;
  });
}

function countStatuses(results) {
  return results.reduce((acc, item) => {
    acc[item.status] = (acc[item.status] || 0) + 1;
    return acc;
  }, {});
}

function parseArgs(argv) {
  const args = {};
  for (let index = 0; index < argv.length; index += 1) {
    const raw = argv[index];
    if (!raw.startsWith('--')) continue;
    const key = raw.slice(2).replace(/-([a-z])/g, (_, char) => char.toUpperCase());
    const next = argv[index + 1];
    if (!next || next.startsWith('--')) args[key] = true;
    else {
      args[key] = next;
      index += 1;
    }
  }
  return args;
}

function resolveOptional(value) {
  return value ? path.resolve(value) : '';
}

function safeTail(value) {
  return String(value || '').slice(-1200);
}

function readJson(file) {
  if (!file || !fs.existsSync(file)) return null;
  try {
    return JSON.parse(fs.readFileSync(file, 'utf8').replace(/^\uFEFF/, ''));
  } catch {
    return null;
  }
}

function isClaimableProofGapClosure(summary) {
  return Boolean(
    summary?.complete === true &&
    Number(summary?.openCount || 0) === 0 &&
    Number(summary?.closedCount || 0) >= 5 &&
    Number(summary?.businessGapCount || 0) >= 5
  );
}

function escapeCell(value) {
  return String(value || '').replace(/\|/g, '/').replace(/\r?\n/g, ' ');
}

function withBom(text) {
  return `\uFEFF${text}`;
}

if (require.main === module) {
  try {
    main();
  } catch (error) {
    console.error(error && error.stack ? error.stack : String(error));
    process.exit(1);
  }
}

module.exports = {
  buildContext,
  buildSteps,
  isClaimableProofGapClosure
};