.
*
*/
namespace Friendica\Util;
use DOMDocument;
use DOMXPath;
use Friendica\Content\OEmbed;
use Friendica\Core\Hook;
use Friendica\Core\Logger;
use Friendica\Database\Database;
use Friendica\Database\DBA;
use Friendica\DI;
use Friendica\Network\HTTPException;
/**
* Get information about a given URL
*
* Class with methods for extracting certain content from an url
*/
class ParseUrl
{
const DEFAULT_EXPIRATION_FAILURE = 'now + 1 day';
const DEFAULT_EXPIRATION_SUCCESS = 'now + 3 months';
/**
* Maximum number of characters for the description
*/
const MAX_DESC_COUNT = 250;
/**
* Minimum number of characters for the description
*/
const MIN_DESC_COUNT = 100;
/**
* Search for chached embeddable data of an url otherwise fetch it
*
* @param string $url The url of the page which should be scraped
* @param bool $no_guessing If true the parse doens't search for
* preview pictures
* @param bool $do_oembed The false option is used by the function fetch_oembed()
* to avoid endless loops
*
* @return array which contains needed data for embedding
* string 'url' => The url of the parsed page
* string 'type' => Content type
* string 'title' => (optional) The title of the content
* string 'text' => (optional) The description for the content
* string 'image' => (optional) A preview image of the content (only available if $no_geuessing = false)
* array 'images' => (optional) Array of preview pictures
* string 'keywords' => (optional) The tags which belong to the content
*
* @throws HTTPException\InternalServerErrorException
* @see ParseUrl::getSiteinfo() for more information about scraping
* embeddable content
*/
public static function getSiteinfoCached($url, $no_guessing = false, $do_oembed = true): array
{
if (empty($url)) {
return [
'url' => '',
'type' => 'error',
];
}
$urlHash = hash('sha256', $url);
$parsed_url = DBA::selectFirst('parsed_url', ['content'],
['url_hash' => $urlHash, 'guessing' => !$no_guessing, 'oembed' => $do_oembed]
);
if (!empty($parsed_url['content'])) {
$data = unserialize($parsed_url['content']);
return $data;
}
$data = self::getSiteinfo($url, $no_guessing, $do_oembed);
$expires = $data['expires'];
unset($data['expires']);
DI::dba()->insert(
'parsed_url',
[
'url_hash' => $urlHash,
'guessing' => !$no_guessing,
'oembed' => $do_oembed,
'url' => $url,
'content' => serialize($data),
'created' => DateTimeFormat::utcNow(),
'expires' => $expires,
],
Database::INSERT_UPDATE
);
return $data;
}
/**
* Parse a page for embeddable content information
*
* This method parses to url for meta data which can be used to embed
* the content. If available it prioritizes Open Graph meta tags.
* If this is not available it uses the twitter cards meta tags.
* As fallback it uses standard html elements with meta informations
* like \Awesome Title\ or
* \
*
* @param string $url The url of the page which should be scraped
* @param bool $no_guessing If true the parse doens't search for
* preview pictures
* @param bool $do_oembed The false option is used by the function fetch_oembed()
* to avoid endless loops
* @param int $count Internal counter to avoid endless loops
*
* @return array which contains needed data for embedding
* string 'url' => The url of the parsed page
* string 'type' => Content type (error, link, photo, image, audio, video)
* string 'title' => (optional) The title of the content
* string 'text' => (optional) The description for the content
* string 'image' => (optional) A preview image of the content (only available if $no_guessing = false)
* array 'images' => (optional) Array of preview pictures
* string 'keywords' => (optional) The tags which belong to the content
*
* @throws \Friendica\Network\HTTPException\InternalServerErrorException
* @todo https://developers.google.com/+/plugins/snippet/
* @verbatim
*
*
*
*
*
* Shiny Trinket
* 
* Shiny trinkets are shiny.
*
* @endverbatim
*/
public static function getSiteinfo($url, $no_guessing = false, $do_oembed = true, $count = 1)
{
if (empty($url)) {
return [
'url' => '',
'type' => 'error',
];
}
// Check if the URL does contain a scheme
$scheme = parse_url($url, PHP_URL_SCHEME);
if ($scheme == '') {
$url = 'http://' . ltrim($url, '/');
}
$url = trim($url, "'\"");
$url = Network::stripTrackingQueryParams($url);
$siteinfo = [
'url' => $url,
'type' => 'link',
'expires' => DateTimeFormat::utc(self::DEFAULT_EXPIRATION_FAILURE),
];
if ($count > 10) {
Logger::log('Endless loop detected for ' . $url, Logger::DEBUG);
return $siteinfo;
}
$curlResult = DI::httpRequest()->get($url);
if (!$curlResult->isSuccess()) {
return $siteinfo;
}
$siteinfo['expires'] = DateTimeFormat::utc(self::DEFAULT_EXPIRATION_SUCCESS);
// If the file is too large then exit
if (($curlResult->getInfo()['download_content_length'] ?? 0) > 1000000) {
return $siteinfo;
}
// Native media type, no need for HTML parsing
$type = $curlResult->getHeader('Content-Type');
if ($type) {
preg_match('#(image|video|audio)/#i', $type, $matches);
if ($matches) {
$siteinfo['type'] = array_pop($matches);
return $siteinfo;
}
}
// If it isn't a HTML file then exit
if (($curlResult->getContentType() != '') && !strstr(strtolower($curlResult->getContentType()), 'html')) {
return $siteinfo;
}
if ($cacheControlHeader = $curlResult->getHeader('Cache-Control')) {
if (preg_match('/max-age=([0-9]+)/i', $cacheControlHeader, $matches)) {
$maxAge = max(86400, (int)array_pop($matches));
$siteinfo['expires'] = DateTimeFormat::utc("now + $maxAge seconds");
}
}
$header = $curlResult->getHeader();
$body = $curlResult->getBody();
if ($do_oembed) {
$oembed_data = OEmbed::fetchURL($url);
if (!empty($oembed_data->type)) {
if (!in_array($oembed_data->type, ['error', 'rich', ''])) {
$siteinfo['type'] = $oembed_data->type;
}
// See https://github.com/friendica/friendica/pull/5763#discussion_r217913178
if ($siteinfo['type'] != 'photo') {
if (isset($oembed_data->title)) {
$siteinfo['title'] = trim($oembed_data->title);
}
if (isset($oembed_data->description)) {
$siteinfo['text'] = trim($oembed_data->description);
}
if (isset($oembed_data->thumbnail_url)) {
$siteinfo['image'] = $oembed_data->thumbnail_url;
}
}
}
}
$charset = '';
// Look for a charset, first in headers
// Expected form: Content-Type: text/html; charset=ISO-8859-4
if (preg_match('/charset=([a-z0-9-_.\/]+)/i', $header, $matches)) {
$charset = trim(trim(trim(array_pop($matches)), ';,'));
}
// Then in body that gets precedence
// Expected forms:
// -
// -
// -
// -
// We escape