File: /www/wwwroot/healthyton.com/wp-scrap/spinbot_get_medicalnews.php
<?php
require 'vendor/autoload.php';
use Goutte\Client;
function strip_tags_blacklist($html, $tags) {
$html = preg_replace('/<'. $tags .'\b[^>]*>(.*?)<\/'. $tags .'>/is', "", $html);
return $html;
}
$url = 'https://www.medicalnewstoday.com/articles/324376.php';
if(isset($_GET['url'])){
$url = $_GET['url'];
}
$client = new Client();
$crawler = $client->request('GET', $url);
$content = $crawler->filter('.article_body div')->html();
//echo $content;exit;
//$process_content = strip_tags_blacklist($content,'div');
$process_content = strip_tags_blacklist($content,'header');
$process_content = strip_tags_blacklist($process_content,'center');
$process_content = strip_tags_blacklist($process_content,'style');
$process_content = strip_tags_blacklist($process_content,'img');
$process_content = strip_tags_blacklist($process_content,'a');
$textToSpin = trim($process_content);
//print_r($textToSpin);exit;
//$descriptions = $crawler->filter('.td-post-content p,.td-post-content > h2')->each(function($node) {
// return $node->xpath();
// return $node->text();
//});
//print_r($descriptions);
//print_r($content);exit;
//exit;
echo '<div style="width:45%;float:left;">';
//$textToSpin = file_get_contents('spin.txt');
//echo $textToSpin;exit;
echo '<h3><u>Original Text:</u></h3> <br/>'. $textToSpin;
echo '</div>';
echo '<div style="width:50%;float:right;">';
$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);
$aHeaders = putHeadersTextIntoArray($strResponseHeaders);
// Display return values
echo '<h3><u>Text After Spinning:</u></h3> <br/>' . $strResponseBody;
// Notice the use of response header 'available-spins' to keep track of remaining Spinbot credits.
echo '</div>';
echo '<div style="clear:both;"></div>';
echo '<hr><br/><br/><strong>Available Spins:</strong> <br/>' . $aHeaders['available-spins'];
// helper function to process return header plain text
function putHeadersTextIntoArray($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;
}
?>