zServices\ReceitaFederal\Services\Portais\AN\Crawler

Complexity

Total Complexity

15

Size/Duplication

Total Lines	100
Duplicated Lines	28 %

Coupling/Cohesion

Components	1
Dependencies	4

Importance

Changes

0

5 Methods

Rating	Name	Duplication	Size	Complexity
A	__construct()	0	6	1
A	hasError()	0	19	4
C	scraping()	28	28	8
A	clearString()	0	4	1
A	scrap()	0	5	1

<?php namespace zServices\ReceitaFederal\Services\Portais\AN;

use zServices\Miscellany\Interfaces\CrawlerInterface;
use Symfony\Component\DomCrawler\Crawler as BaseCrawler;
use zServices\Miscellany\Exceptions\NoSelectorsConfigured;
use zServices\Miscellany\Exceptions\InvalidCaptcha;
use zServices\Miscellany\Exceptions\ErrorFoundData;

/**
* 
*/
class Crawler extends BaseCrawler implements CrawlerInterface
{

	/**
	 * Selectors to crawler
	 * @var array
	 */
	private $selectors = [];

	/**
	 * Get selectors and initialize crawler in HTML
	 * @param string $html      
	 * @param array  $selectors
	 * @return void

Adding a @return annotation to constructors is generally not recommended as a constructor does not have a meaningful return value.

	 */
	public function __construct($html, $selectors)
	{
		$this->selectors = $selectors;

		parent::__construct($html);
	}

	/**
	 * Verifica antes de fazer o crawler se possui erros
	 * na requisição
	 * @return boolean|null 
	 */
	public function hasError()
	{
		if (count($this->selectors) == 0) {
			throw new NoSelectorsConfigured("NoSelectorsConfigured", 1);
		}

		// verifica se a página seguida na requisição 
		// é página de erro da receita federal
		$node = $this->filter($this->selectors['error']);

		if ($node->count()) {
			 throw new ErrorFoundData($this->clearString($node->text()), 1);
		}
        
		// CNPJ informado é válido?
		if ($this->filter('#imgCaptcha')->count()) {
			throw new InvalidCaptcha('Captcha inválido', 99);
		}
	}

	/**
	 * Extrai informações do HTML através do DOM
	 *
	 * @return array
	 */
	public function scraping()

This method seems to be duplicated in your project.

	{
		$scrapped = [];

		$this->hasError();

		foreach ($this->selectors as $name => $selector) {
			if (is_string($selector)) {
				$node = $this->scrap($selector);

				if ($node->count()) {
					$scrapped[$name] = $this->clearString($node->text());
				}
			} elseif (is_array($selector)) {
				foreach ($selector as $selector => $repeat) {

The expression $selector of type integer|string|array is not guaranteed to be traversable. How about adding an additional type check?

					$node = $this->scrap($selector);
					if ($node->count()) {
						foreach ($node->filter($repeat) as $loop)
						{
							$scrapped[$name][] = $this->clearString($loop->nodeValue);
						}
					}
				}
			}
		}

		return $scrapped;
	}

	/**
	 * Limpa o valor repassado
	 * @param  string $string
	 * @return string
	 */
	public function clearString($string)
	{
		return trim(preg_replace(['/[\s]+/mu'], ' ', $string));
	}

	/**
	 * Filtra selector no crawler
	 */
	public function scrap($selector)
	{
		$node = $this->filter($selector);
		return $node;
	}
}