@democracy-deutschland/bt-named-polls/distribution/Browser.js

scraper for the german "Bundestag" (Namentliche Abstimmungen)

'use strict';

Object.defineProperty(exports, "__esModule", {
  value: true
});

var _extends = Object.assign || function (target) { for (var i = 1; i < arguments.length; i++) { var source = arguments[i]; for (var key in source) { if (Object.prototype.hasOwnProperty.call(source, key)) { target[key] = source[key]; } } } return target; };

const request = require('request');

// fix Date.parse
const MONTH = [{ string: 'März', fix: 'March' }, { string: 'Mai', fix: 'May' }, { string: 'Oktober', fix: 'October' }, { string: 'Dezember', fix: 'December' }];

class Browser {

  constructor() {
    this.cookie = null;

    this.initialize = uri => this.request(_extends({}, this.defReqOpt, {
      uri
    }));

    this.request = opts => {
      const reqOptions = _extends({
        timeout: 15000,
        method: 'GET',
        jar: this.cookie
      }, opts);

      return new Promise((resolve, reject) => {
        request(reqOptions, (error, res, body) => {
          if (!error && res.statusCode === 200) {
            resolve({ res, body });
          } else {
            reject(error);
          }
        });
      });
    };

    this.getHead = body => {
      const [, head] = body.match(/<article class="bt-artikel col-xs-12 bt-standard-content">(.*?)<\/article>/s);
      let [, date] = head.match(/<span class="bt-dachzeile">(.*?)<\/span>/s);
      const [, title] = head.match(/<br\/>(.*?)<\/h3>/s);
      const documentsData = head.match(/<i class="icon-doc"><\/i>(\d{1,3}\/\d{1,10})<\/a>/g);
      MONTH.forEach(({ string, fix }) => {
        date = date.replace(`${string}`, `${fix}`);
      });
      let documents = [];
      if (documentsData) {
        documents = documentsData.map(doc => doc.replace('<i class="icon-doc"></i>', '').replace('</a>', ''));
      }
      return { date: new Date(Date.parse(date)), title, documents };
    };

    this.getPartyVotes = body => {
      const partyVotesHtml = body.match(/<div class="col-xs-12 col-sm-3">(.*?)<\/a>\s*<\/div>/gs);
      const partyVotes = partyVotesHtml.reduce((prev, html) => {
        const [, party] = html.match(/<h4 class="bt-chart-fraktion">(.*?)<br\/>/s);
        const voteResults = html.match(/<li class="bt-legend-(?:ja|nein|enthalten|na)">(.*?)<\/li>/gs).map(voteResultHtml => voteResultHtml.match(/>(.*?)</)[1].split(' ')).reduce((prev, [votes, decision]) => _extends({}, prev, { [decision]: votes }), {});
        return _extends({}, prev, { [party]: voteResults });
      }, {});
      return partyVotes;
    };

    this.cookie = request.jar();
  }

}

exports.default = Browser;