3
\$\begingroup\$

I already posted a not (yet) complete version of this class, but that question got closed, because it contained not passing tests. Here is the full version, all 5 tests pass.

Originally there was only processPlayer when I found a very rare case (see the last test Special case). The method player is a patch to handle this kind of cases. Maybe the two method could be merged.

class Scraper {
  /**
   * Converts a html string to a cheerio object
   * @param {String} html The html string
   * @return {Object} The cheerio object
   */
  htmlToDom(html) {
    return cheerio.load(html)
  }

  // http://jsbin.com/nojedozaxo/1/edit?js,console,output
  player(href, name) {
    let p = name.replace(/\./g, '.+')
    let r = new RegExp(".+(" + p + ")-.+", 'i')
    let R = href.match(r)[1].split('-')
      .map(w => w.charAt(0).toUpperCase() + w.slice(1))
    return {
      firstName: R[2],
      lastName: R.slice(0, 2).join('-')
    }
  }

  processPlayers(players) {
    // is it a selector function or an object
    players = typeof players === 'function' ? players('a') : players.find('a')
    let href = players.attr('href')
    let N = players.text().trim().split(/\s+-\s+/).map(n => {
      if (n.indexOf(' ') === -1) {
        return this.player(href, n)
      }
      let r = new RegExp('.+(' + n.replace(/-/g, '\\S+').replace(/\./g, '[^-]+').replace(/\s/g, '.') + ').+', 'i')
      let p = {
        lastName: n.replace(/\s\S+\./g, '')
      }
      href.match(r).map(m => {
        r = new RegExp(p.lastName.replace(/\s/g, '.') + '-', 'i')
        m = m.replace(r, '').split(/-/).map(i => {
          return i.substring(0, 1).toUpperCase() + i.substring(1)
        })
        p.firstName = n.split(p.lastName + ' ')[1].replace(/[a-z]*\./g, '').split('').map(l => {
          return m[0] && m[0].indexOf(l) === 0 ? m.shift() : l
        }).join('')
      })
      return p
    })
    return {
      p1: N[0],
      p2: N[1]
    }
  }
}

// test ===========================================================

test('simple case', function() {
  let playersCell = `
    <td>
      <a href="/t/pavlyuchenkova-anastasia-sorribes-tormo-sara/">
        <span>Pavlyuchenkova A.</span>
         - Sorribes Tormo S.
       </a>
    </td>
    `
  const scraper = new Scraper()
  const td = scraper.htmlToDom(playersCell)
  const players = scraper.processPlayers(td)

  equal(players.p1.firstName, 'Anastasia')
  equal(players.p1.lastName, 'Pavlyuchenkova')

  deepEqual(players.p2, {
    firstName: 'Sara',
    lastName: 'Sorribes Tormo',
  })
});

// =====================================================================

test('hyphen in last name', function() {
  let playersCell = `
    <td><a href="/t/kudermetova-veronika-duque-marino-mariana/">
      <span>Kudermetova V.</span> - Duque-Marino M.</a>
    </td>
    `
  const scraper = new Scraper()
  const td = scraper.htmlToDom(playersCell)
  const players = scraper.processPlayers(td)

  equal(players.p2.firstName, 'Mariana')
  equal(players.p2.lastName, 'Duque-Marino')
});

// =====================================================================

test('hyphen in first name', function() {
  let playersCell = `
    <td>
      <a href="/t/tsonga-jo-wilfried-mayer-florian/">
        <span>Tsonga J-W.</span>
        - Mayer F.
      </a>
    </td>
    `
  const scraper = new Scraper()
  const td = scraper.htmlToDom(playersCell)
  const players = scraper.processPlayers(td)

  equal(players.p1.firstName, 'Jo-Wilfried')
  equal(players.p1.lastName, 'Tsonga')
});

test('two first names', function() {
  let playersCell = `
      <td>
        <a href="/t/alexandrova-ekaterina-muguruza-blanco-garbine/">
        Alexandrova E. - <span>Muguruza</span> B. G.</a>
      </td>
    `
  const scraper = new Scraper()
  const td = scraper.htmlToDom(playersCell)
  const players = scraper.processPlayers(td)

  equal(players.p2.firstName, 'Blanco Garbine')
  equal(players.p2.lastName, 'Muguruza')
});

// =====================================================================

test('short first name has two letters', t => {
  let playersCell = `
    <a href="/t/pliskova-kristyna-babos-timea/">
      <span class="bold">Pliskova Kr.</span> - Babos T.</a>`
  const scraper = new Scraper()
  const td = scraper.htmlToDom(playersCell)
  const players = scraper.processPlayers(td)

  equal(players.p1.firstName, 'Kristyna')
  equal(players.p1.lastName, 'Pliskova')
})

// =====================================================================
test('special case', t => {
  let playersCell = `
    <a href="/t/kovalik-jozef-ramirez-hidalgo-ruben-sdgdsg/">
    <span class="bold">Kovalik J.</span> - Ramirez-H.R.</a>
    `
  const scraper = new Scraper()
  const td = scraper.htmlToDom(playersCell)
  const players = scraper.processPlayers(td)

  equal(players.p2.firstName, 'Ruben')
  equal(players.p2.lastName, 'Ramirez-Hidalgo')
});
<script src="https://wzrd.in/standalone/cheerio@latest"></script>
<script src="https://code.jquery.com/qunit/qunit-1.12.0.js"></script>
<link href="https://code.jquery.com/qunit/qunit-1.12.0.css" rel="stylesheet" />

<div id="qunit"></div>
<div id="qunit-fixture"></div>

Here is an alternative JsBin: http://jsbin.com/sizocatizi/1/edit?html,js,output

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.