@skypilot/scraper/lib/repos/iana.org-domains.js

Node-based scriptable web scraper

"use strict";

Object.defineProperty(exports, "__esModule", {
  value: true
});
exports.builder = exports.commands = void 0;

var _ScriptBuilder = require("../scraper/ScriptBuilder");

const collectionName = 'domains';
const commands = [{
  action: 'goTo',
  url: 'https://www.iana.org/domains/reserved'
}, {
  action: 'runOnAll',
  query: {
    sel: 'table#arpa-table > tbody > tr > td > span.domain.label',
    limit: 3
  },
  commands: [{
    action: 'follow',
    query: {
      sel: 'a'
    }
  }, {
    action: 'query',
    queryDict: {
      title: {
        sel: 'head > title'
      },
      sponsor: toCellSelector('Sponsoring Organisation'),
      adminContact: toCellSelector('Administrative Contact'),
      techContact: toCellSelector('Technical Contact')
    }
  }, {
    action: 'write'
  }],
  options: {
    collectionName
  }
}];
exports.commands = commands;

function toCellSelector(label) {
  return {
    sel: `//h2[contains(text(), '${label}')]/following-sibling::b`
  };
}

const builder = new _ScriptBuilder.ScriptBuilder().goTo('https://www.iana.org/domains/reserved').runOnAll({
  sel: 'table#arpa-table > tbody > tr > td > span.domain.label',
  limit: 3
}, new _ScriptBuilder.ScriptBuilder().follow('a').query({
  title: 'head > title',
  sponsor: toCellSelector('Sponsoring Organisation'),
  adminContact: toCellSelector('Administrative Contact'),
  techContact: toCellSelector('Technical Contact')
}).write(), {
  collectionName
});
exports.builder = builder;