tripsense/tools/fetch-datav-admin.cjs

#!/usr/bin/env node
/**
 * 拉取全国行政区划边界（阿里 DataV GeoAtlas，免密钥）。
 *
 * 为什么不用 Overpass / OSM：
 *   1) Overpass 公共实例限流极狠，实测连续大查询后连 1KB 小查询都 502/504；
 *   2) OSM 中国的 admin_level 层级与直觉相反、归属字段不可靠；
 *   3) DataV 用的是国家民政部 adcode 体系，层级清晰（省 → 市 → 县），
 *      且**自带县级边界多边形**，能做精确点面判定。
 *
 * 流程：100000_full.json 拿省 → 省文件拿地级市 → 市文件拿县。
 * 直辖市 / 省直辖县级单位由省文件直接给出（level != 'city'）。
 *
 * 用法：node tools/fetch-datav-admin.cjs C:/tmp/datav-admin.json
 */
const fs = require('fs');

const BASE = 'https://geo.datav.aliyun.com/areas_v3/bound';
const UA = { 'User-Agent': 'Mozilla/5.0 (TripSense build script)' };

async function getJSON(adcode) {
  const url = `${BASE}/${adcode}_full.json`;
  for (let attempt = 0; attempt < 4; attempt++) {
    try {
      const res = await fetch(url, { headers: UA });
      if (!res.ok) throw new Error(`${res.status}`);
      const text = await res.text();
      return JSON.parse(text);
    } catch (e) {
      if (attempt === 3) throw new Error(`${adcode}: ${e.message}`);
      await new Promise((r) => setTimeout(r, 800 * (attempt + 1)));
    }
  }
  return null;
}

function countPts(coords) {
  let n = 0;
  const walk = (c) => {
    if (typeof c[0] === 'number') { n++; return; }
    for (const x of c) walk(x);
  };
  walk(coords);
  return n;
}

async function mapLimit(list, limit, fn) {
  const out = [];
  let idx = 0;
  const workers = Array.from({ length: limit }, async () => {
    while (idx < list.length) {
      const i = idx++;
      try { out[i] = await fn(list[i], i); } catch (e) { out[i] = null; console.error('  !', list[i], e.message); }
    }
  });
  await Promise.all(workers);
  return out;
}

function pack(f, extra) {
  const g = f.geometry || { type: '', coordinates: [] };
  return {
    adcode: String(f.properties.adcode),
    name: f.properties.name,
    level: f.properties.level,
    centroid: f.properties.centroid || f.properties.center || null,
    center: f.properties.center || null,
    geomType: g.type,
    coordinates: g.coordinates,
    pts: countPts(g.coordinates),
    ...extra,
  };
}

(async () => {
  const outFile = process.argv[2] || 'C:/tmp/datav-admin.json';
  console.error('拉取省级…');
  const provDoc = await getJSON('100000');
  const provinces = provDoc.features.map((f) => ({
    adcode: String(f.properties.adcode),
    name: f.properties.name,
    level: f.properties.level,
  }));
  console.error(`省 ${provinces.length} 个`);

  // 第一层：每个省的直接下级（地级市，或直辖市的区）
  console.error('拉取省市层…');
  const firstLevel = await mapLimit(provinces, 5, async (p) => {
    const doc = await getJSON(p.adcode);
    return (doc.features || []).map((f) => pack(f, {
      province: p.name,
      provinceAdcode: p.adcode,
      childrenNum: Number(f.properties.childrenNum || 0),
    }));
  });

  const cities = firstLevel.flat().filter(Boolean);
  const direct = cities.filter((c) => c.level !== 'city');
  const cityList = cities.filter((c) => c.level === 'city' && c.childrenNum > 0);
  console.error(`地级市 ${cityList.length} 个，直辖县级 ${direct.length} 个，拉县级…`);

  const countyDocs = await mapLimit(cityList, 6, async (c) => {
    const doc = await getJSON(c.adcode);
    return (doc.features || []).map((f) => pack(f, {
      province: c.province,
      provinceAdcode: c.provinceAdcode,
      city: c.name,
      cityAdcode: c.adcode,
    }));
  });

  const counties = countyDocs.flat().filter(Boolean);
  console.error(`县级 ${counties.length} 个，总顶点 ${counties.reduce((s, c) => s + c.pts, 0)}`);

  fs.writeFileSync(outFile, JSON.stringify({ provinces, cities, counties }));
  console.error('写入', outFile, (fs.statSync(outFile).size / 1024 / 1024).toFixed(1) + 'MB');
})();