skills-smartbadge-ops/scripts/fetch-report-text.mjs

#!/usr/bin/env node
// 报告正文拉取: --id 按 reportId 定位(经 /api/reports/:id)或 --url 直链,下载 HTML → 剥标签存 out/ 并打印
// 用法:
//   node fetch-report-text.mjs --id <reportId>            # 按报告 id(当前窗口内最佳)
//   node fetch-report-text.mjs --url <ossUrl|downloadUrl> # 直链(自动跟随 302)
//   node fetch-report-text.mjs --id xxxx --preview        # 只输出标题+前 28 行摘要
//   node fetch-report-text.mjs --id xxxx --json           # 输出 {id,title,chars,file}
//   node fetch-report-text.mjs --id xxxx --out 报告.txt   # 指定输出文件
// 密码: SMARTBADGE_PASSWORD(无则复用 scripts/.token)
import fs from 'fs';
import path from 'path';
import { fileURLToPath } from 'url';
import { baseUrl, fail, loadToken, login, httpJson, htmlToText, fetchAllReports } from './lib.mjs';

const args = process.argv.slice(2);
const argVal = (name) => {
  const i = args.indexOf(name);
  return i >= 0 ? args[i + 1] : undefined;
};
const id = argVal('--id');
const url = argVal('--url');
const outPath = argVal('--out');
const preview = args.includes('--preview');
const jsonOut = args.includes('--json');
const base = baseUrl();

if (!id && !url) fail('用法: node fetch-report-text.mjs --id <reportId> | --url <http(s)>  [--preview] [--json] [--out 路径]', 3);
if (url && !/^https?:\/\//i.test(url)) fail(`--url 仅接受 http(s) 协议(拒绝 file:// 等本地协议),收到: ${url}`, 4);

async function getRedirectTarget() {
  const token = loadToken();
  let r = await httpJson('GET', `/api/reports/${id}`, undefined, { base, token, raw: true });
  if (!r.ok || !r.json?.data) {
    if (r.status === 401 && process.env.SMARTBADGE_PASSWORD) {
      await login({ base, password: process.env.SMARTBADGE_PASSWORD });
      r = await httpJson('GET', `/api/reports/${id}`, undefined, { base, token: loadToken(), raw: true });
    }
    if (!r.ok || !r.json?.data) {
      const recent = await fetchAllReports({ base, token, pageSize: 3 }).then((rs) => rs.map((x) => `${x.id}(${x.type}, ${x.periodStart})`).join('、'));
      fail(`报告 id 未找到(${id}). 近期报告参考: ${recent || '无'}`, 5);
    }
  }
  const d = r.json.data;
  const target = d.ossUrl || d.downloadUrl;
  if (!target) fail(`报告 ${id} 无 ossUrl/downloadUrl,请在服务器侧检查 OSS 配置`, 6);
  return { id: String(d.id), target, title: d.title, type: d.type, period: `${d.periodStart} ~ ${d.periodEnd}` };
}

async function fetchHtml(target) {
  const controller = new AbortController();
  const timer = setTimeout(() => controller.abort(), 30000);
  try {
    const res = await fetch(target, { signal: controller.signal });
    if (!res.ok) fail(`下载失败 HTTP ${res.status}: ${target}`, 6);
    return await res.text();
  } catch (err) {
    fail(`下载失败(超时 30s 或网络异常): ${err.message}\n  直链: ${target}`, 7);
  } finally {
    clearTimeout(timer);
  }
}

const meta = id
  ? await getRedirectTarget()
  : { id: `url-${Date.now()}`, target: url, title: path.basename(new URL(url).pathname) || 'report', type: '', period: url };

const html = await fetchHtml(meta.target);
const text = htmlToText(html);
const safeName = String(meta.id).replace(/[^a-zA-Z0-9._-]/g, '_');
const file = path.resolve(outPath || path.join(path.dirname(fileURLToPath(import.meta.url)), '..', 'out', `report-${safeName}.txt`));
fs.mkdirSync(path.dirname(file), { recursive: true });
fs.writeFileSync(file, text);

if (jsonOut) {
  console.log(JSON.stringify({ id: meta.id, title: meta.title, type: meta.type, period: meta.period, chars: text.length, file }));
  process.exit(0);
}
if (preview) {
  const lines = text.split('\n').filter(Boolean);
  const first = lines.slice(0, 28).join('\n');
  console.log(`标题: ${meta.title || '(无)'}\n${first}\n...[共 ${text.length} 字符, 全文见 ${file}]`);
} else {
  console.log(text);
  console.error(`\n[全文已另存: ${file}]`);
}