Created
January 20, 2019 11:26
-
-
Save rcarvs/c2d8d1aa024d8c1f74b50fc3481e6487 to your computer and use it in GitHub Desktop.
Very little crawller in PHP
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
<?php | |
/* | |
------------------------------------ | |
HOW TO USE | |
------------------------------------ | |
IN YOUR PHP FILE, INCLUDE THE CRAWLER FILE AND DO AS EXAMPLE BELOW: | |
$crawler = new crawler(); | |
$crawler->setDomainInitial('http://www.ufsj.edu.br'); | |
//if set true, the crawler will add to list and execute just links of initial domain | |
$crawler->setJustOlyLinksInInitialDomail(true); | |
$crawler->percorreWeb(); | |
*/ | |
class pagina{ | |
private $url; | |
private $conteudo; | |
public function getUrl(){ | |
return $this->url; | |
} | |
public function getConteudo(){ | |
return $this->conteudo; | |
} | |
public function setUrl($url){ | |
$this->url = $url; | |
} | |
public function setConteudo($conteudo){ | |
$this->conteudo = $conteudo; | |
} | |
} | |
class crawler{ | |
private $somenteDominioInicial = false; | |
private $dominioInicial; | |
private $links = array(); | |
public function getSomenteDominioInicial(){ | |
return $this->somenteDominioInicial; | |
} | |
public function setJustOlyLinksInInitialDomail($somenteDominioInicial){ | |
$this->somenteDominioInicial = $somenteDominioInicial; | |
} | |
public function setDomainInitial($dominioInicial){ | |
$this->dominioInicial = $dominioInicial; | |
$this->addLink($dominioInicial); | |
} | |
public function getDominioInicial(){ | |
return $this->dominioInicial; | |
} | |
public function addLink($link){ | |
$this->links[] = $link; | |
} | |
private function buscaConteudoPagina($url) { | |
try { | |
if (!empty($url)) { | |
$opts = array('http' => | |
array( | |
'method' => 'GET', | |
'timeout' => 20 | |
) | |
); | |
$context = stream_context_create($opts); | |
$retorno = file_get_contents($url, false, $context); | |
if ($retorno != false) { | |
$pagina = new pagina(); | |
$pagina->setUrl($url); | |
$pagina->setConteudo($retorno); | |
return $pagina; | |
} else { | |
return false; | |
} | |
} else { | |
return false; | |
} | |
} catch (Exception $ex) { | |
return false; | |
} | |
} | |
public function percorreWeb(){ | |
$contador = 0; | |
$quantidadeTentativas = 0; | |
while ($contador >= 0) { | |
if (isset($this->links[$contador])) { | |
$quantidadeTentativas = 0; | |
$url = $this->links[$contador]; | |
$pagina = $this->buscaConteudoPagina($url); | |
if($pagina != false && is_object($pagina)){ | |
//echo $pagina->getConteudo(); | |
$this->buscaLinksConteudo($pagina); | |
//COLOCAR AQUI A FUNÇÃO DE CONTEUDO | |
unset($this->links[$contador]); | |
} | |
}else { | |
$quantidadeTentativas++; | |
if ($quantidadeTentativas > 100) { | |
break; | |
} | |
} | |
$contador++; | |
} | |
} | |
public function getLb() { | |
if (PHP_SAPI == "cli") | |
$lb = "\n"; | |
else | |
$lb = "</br>"; | |
return $lb; | |
} | |
public function buscaLinksConteudo($pagina){ | |
$contador = 0; | |
while ($contador <= strlen($pagina->getConteudo())) { | |
$startLink = strpos($pagina->getConteudo(), "<a", $contador); | |
if ($startLink === false) { | |
break; | |
} else { | |
$contador = $startLink + 2; | |
$startHref = strpos($pagina->getConteudo(), 'href="http', $startLink); | |
if ($startHref === false) { | |
break; | |
} else { | |
$endHref = strpos($pagina->getConteudo(), '"', ($startHref + 13)); | |
if ($endHref === false) { | |
break; | |
} else { | |
if (!in_array(substr($pagina->getConteudo(), ($startHref + 6), ($endHref - $startHref - 6)), $this->links)){ | |
if($this->getSomenteDominioInicial()){ | |
if(strlen(substr($pagina->getConteudo(), ($startHref + 6), ($endHref - $startHref - 6))) > strlen($this->getDominioInicial()) && $this->getDominioInicial() == substr(substr($pagina->getConteudo(), ($startHref + 6), ($endHref - $startHref - 6)), 0,strlen($this->getDominioInicial()))){ | |
$this->links[]=substr($pagina->getConteudo(), ($startHref + 6), ($endHref - $startHref - 6)); | |
echo substr($pagina->getConteudo(), ($startHref + 6), ($endHref - $startHref - 6))." foi adicionado na fila".$this->getLb(); | |
} | |
}else{ | |
$this->links[]=substr($pagina->getConteudo(), ($startHref + 6), ($endHref - $startHref - 6)); | |
echo substr($pagina->getConteudo(), ($startHref + 6), ($endHref - $startHref - 6))." foi adicionado na fila".$this->getLb(); | |
} | |
} | |
} | |
} | |
} | |
} | |
} | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment