How can I scrape LinkedIn company pages with cURL and PHP? - php

Question

I want to scrape some LinkedIn company pages with cURL and PHP with login Credentials. I tried this code. But I got error like
Unauthorized
You must be authenticated to access this page.
Before scraping the company page I have to sign in at LinkedIn with a personal account via cURL, but it doesn't seems to work.
Instead of using simple_html_dom we used above fetch_value.
function fetch_value($str, $find_start = '', $find_end = '') {
if ($find_start == '') {
return '';
}
$start = strpos($str, $find_start);
if ($start === false) {
return '';
}
$length = strlen($find_start);
$substr = substr($str, $start + $length);
if ($find_end == '') {
return $substr;
}
$end = strpos($substr, $find_end);
if ($end === false) {
return $substr;
}
return substr($substr, 0, $end);
}
$linkedin_login_page = "https://www.linkedin.com/uas/login";
$linkedin_ref = "https://www.linkedin.com";
$username = 'username';
$password = 'password';
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, $linkedin_login_page);
curl_setopt($ch, CURLOPT_REFERER, $linkedin_ref);
curl_setopt($ch, CURLOPT_USERAGENT,'Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.7.12) Gecko/20050915 Firefox/1.0.7)');
curl_setopt($ch, CURLOPT_AUTOREFERER, true);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, FALSE);
curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 2);
curl_setopt($ch, CURLOPT_COOKIEJAR, 'cookie.txt');
curl_setopt($ch, CURLOPT_COOKIEFILE, 'cookie.txt');
$login_content = curl_exec($ch);
if (curl_error($ch)) {
echo 'error:' . curl_error($ch);
}
$var = array(
'isJsEnabled' => 'false',
'source_app' => '',
'clickedSuggestion' => 'false',
'session_key' => trim($username),
'session_password' => trim($password),
'signin' => 'Sign In',
'session_redirect' => '',
'trk' => '',
'fromEmail' => ''
);
$var['loginCsrfParam'] = fetch_value($login_content, 'type="hidden" name="loginCsrfParam" value="', '"');
$var['csrfToken'] = fetch_value($login_content, 'type="hidden" name="csrfToken" value="', '"');
$var['sourceAlias'] = fetch_value($login_content, 'input type="hidden" name="sourceAlias" value="', '"');
$post_array = array();
foreach ($var as $key => $value) {
$post_array[] = urlencode($key) . '=' . urlencode($value);
}
$post_string = implode('&', $post_array);
curl_setopt($ch, CURLOPT_URL, "https://www.linkedin.com/uas/login-submit");
curl_setopt($ch, CURLOPT_POST, true);
curl_setopt($ch, CURLOPT_POSTFIELDS, $post_string);
$store = curl_exec($ch);
if (stripos($store, "session_password-login-error") !== false) {
$err = trim(strip_tags(fetch_value($store, '<span class="error" id="session_password-login-error">', '</span>')));
echo "Login error : ".$err;
} elseif (stripos($store, 'profile-nav-item') !== false) {
curl_setopt($ch, CURLOPT_URL, 'https://www.linkedin.com/company-beta/10667/?pathWildcard=10667');
curl_setopt($ch, CURLOPT_POST, false);
curl_setopt($ch, CURLOPT_POSTFIELDS, "");
$content = curl_exec($ch);
curl_close($ch);
echo $content;
} else {
echo "unknown error";
}
Any suggestion please help?
Thanks!