<?php
require 'vendor/autoload.php';
use Goutte\Client;

function strip_tags_blacklist($html, $tags) {
    $html = preg_replace('/<'. $tags .'\b[^>]*>(.*?)<\/'. $tags .'>/is', "", $html);
    return $html;
}

$url = 'https://articlebio.com/leslie-knipfing/';
if(isset($_GET['url'])){
    $url = $_GET['url'];
}
$client = new Client();
$crawler = $client->request('GET', $url);

$name = array_shift($crawler->filter('h1.single-post-news-title')->extract('_text'));
$image = 'https://articlebio.com'.$crawler->filter('.featured-image img.img-responsive')->attr('src');
$quickfacts = $content = $crawler->filter('.biofactstable')->html();
$strip_tags = "center|style|span|ins|script|a|img|input|button|figure|noscript";
$quickfacts = preg_replace('#<i.*?>.*?</i>#i', '', $quickfacts);
$quickfacts = preg_replace('#<a.*?>.*?</a>#i', '', $quickfacts);
$quickfacts = preg_replace("#<\s*\/?(".$strip_tags.")\s*[^>]*?>#im", '', $quickfacts);
$quickfacts = preg_replace('/\s+/', ' ', $quickfacts);

$quickfacts = "<table class='table biofactstable'>".$quickfacts.'</table>';

$html = $crawler->filter('#postarticle')->html();

$html = preg_replace('/(<(script|style)\b[^>]*>).*?(<\/\2>)/is', "$1$3", $html);
// remove link
$html = preg_replace('#<a.*?>.*?</a>#i', '', $html);
$html = preg_replace("#<\s*\/?(".$strip_tags.")\s*[^>]*?>#im", '', $html);

$strip_tags = "figcaption";
/*$html = preg_replace("#<\s*\/?(".$strip_tags.")\s*[^>]*?>#im", '', $html);*/
$html = preg_replace('/(<(figcaption)\b[^>]*>).*?(<\/\2>)/is', "$1$3", $html);
$html = preg_replace('/<figcaption[^>]*>([\s\S]*?)<\/figcaption[^>]*>/', '', $html);
$html = preg_replace('/<div[^>]*>([\s\S]*?)<\/div[^>]*>/', '', $html);

$textToSpin = str_replace('</div>','',$html);

$url = 'https://api.spinbot.com';
$header = array();
// Required header:
$spinbotApiKey = '87b0d1bb9b3c420381ee141837cbc7f5';
$header[] = "x-auth-key:$spinbotApiKey";

// optional header values
$header[] = 'x-spin-cap-words:true';
$header[] = 'x-words-to-skip:rewrit,nonExistentWordPart';
$header[] = 'x-min-percent-change-per-sentence:any';
//$header[] = 'x-action:getavailablespins';

// Execute cURL request, get response
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, $url);
curl_setopt($ch, CURLOPT_POST, true);
curl_setopt($ch, CURLOPT_POSTFIELDS, $textToSpin);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
curl_setopt($ch, CURLOPT_HEADER, true);
curl_setopt($ch, CURLOPT_HTTPHEADER, $header);
curl_setopt($ch, CURLOPT_VERBOSE, 1);
$response = curl_exec($ch);
curl_close($ch);

// Make the response readable
list($strResponseHeaders, $strResponseBody) = explode("\r\n\r\n", $response, 2);
list($strResponseHeaders, $strResponseBody) = explode("\r\n\r\n", $strResponseBody, 2);
$aHeaders = putHeadersTextIntoArray($strResponseHeaders);

//$out = substr(strstr($strResponseBody, '<p>'), strlen('<p>'));

$out = $quickfacts.$strResponseBody;


$fields['name'] = trim($name);
$fields['image'] = $image;
$fields['body'] = $out;
$fields['available-spins'] = $aHeaders['available-spins'];

header('Content-Type: application/json');
echo json_encode($fields);


function putHeadersTextIntoArray($header_text) {
    $headers = array();
    foreach (explode("\r\n", $header_text) as $i => $line)
        if ($i === 0) {
            $headers['http_code'] = $line;
        } else {
            list ($key, $value) = explode(': ', $line);
            $headers[$key] = $value;
        }
    return $headers;
}
?>