CN20210324

<?php
require "./vendor/autoload.php";

define (
    "USER_AGENT_STRING_MOZ47",
    "Mozilla/5.0 (Windows NT 6.3; Win64; x64; rv:47.0) Gecko/20100101 Firefox/47.0"
);


/*
 * recebe source code HTML
 * retorna um array, indexado por URL, das âncoras encontradas
 * [
   "https://site.com/blabla" => "As melhores batatas",
   ...
   "http://outro" => "outra âncora"
  ]
 */

function extractAsFromHtml(
    string $pStrHtml
){
    $ret = [];
    $doc = new \DOMDocument();
    if ($doc){
        $bTrueOrFalse = @$doc->loadHTML($pStrHtml);
        if ($bTrueOrFalse){
            $as = $doc->getElementsByTagName('a');
            foreach ($as as $a){
                $href = $a->getAttribute('href');
                $anchor = $a->nodeValue;
                //$ret[] = ['href'=>$href, 'anchor'=>$anchor];
                $bNew = !array_key_exists($href, $ret);
                if ($bNew) $ret[$href] = $anchor;
            }
        }
    }
    return $ret;
}//extractAsFromHtml
$ch = curl_init("https://www.google.com/search?q=Alain+Prost");
//$ch = curl_init("https://arturmarques.com/");
if ($ch){
    $bTrueOrFalse = curl_setopt($ch, CURLOPT_USERAGENT, USER_AGENT_STRING_MOZ47);

    $bTrueOrFalse = curl_setopt($ch, CURLOPT_HTTPGET, true); //GET REQUEST
    $bTrueOrFalse = curl_setopt($ch, CURLOPT_POST, false); //POST REQUEST

    $bTrueOrFalse = curl_setopt($ch, CURLOPT_RETURNTRANSFER, true); //obter os dados na origem
    $bTrueOrFalse = curl_setopt($ch, CURLOPT_ENCODING, ""); //encoding automático
    $bTrueOrFalse = curl_setopt($ch, CURLOPT_VERBOSE, true); //+ informação

    $bTrueOrFalse = curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true); //curl segue http-redirections
    $bTrueOrFalse = curl_setopt($ch, CURLOPT_MAXREDIRS, 2); //curl estabelece limite máximo para redireções

    $dataAtSourceOrJustABooleanIfReturnTransferIsFalse = curl_exec($ch);

    $as = extractAsFromHtml($dataAtSourceOrJustABooleanIfReturnTransferIsFalse);
    var_dump ($as);

    //echo $dataAtSourceOrJustABooleanIfReturnTransferIsFalse;
}
else{
    echo "Could not init the CURL object";
}