src/Util/ParseUrl.php

   1 <?php
   2 /**
   3  * @copyright Copyright (C) 2010-2021, the Friendica project
   4  *
   5  * @license GNU AGPL version 3 or any later version
   6  *
   7  * This program is free software: you can redistribute it and/or modify
   8  * it under the terms of the GNU Affero General Public License as
   9  * published by the Free Software Foundation, either version 3 of the
  10  * License, or (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15  * GNU Affero General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU Affero General Public License
  18  * along with this program.  If not, see <https://www.gnu.org/licenses/>.
  19  *
  20  */
  21
  22 namespace Friendica\Util;
  23
  24 use DOMDocument;
  25 use DOMXPath;
  26 use Friendica\Content\OEmbed;
  27 use Friendica\Core\Hook;
  28 use Friendica\Core\Logger;
  29 use Friendica\Database\Database;
  30 use Friendica\Database\DBA;
  31 use Friendica\DI;
  32 use Friendica\Network\HTTPException;
  33
  34 /**
  35  * Get information about a given URL
  36  *
  37  * Class with methods for extracting certain content from an url
  38  */
  39 class ParseUrl
  40 {
  41         const DEFAULT_EXPIRATION_FAILURE = 'now + 1 day';
  42         const DEFAULT_EXPIRATION_SUCCESS = 'now + 3 months';
  43
  44         /**
  45          * Maximum number of characters for the description
  46          */
  47         const MAX_DESC_COUNT = 250;
  48
  49         /**
  50          * Minimum number of characters for the description
  51          */
  52         const MIN_DESC_COUNT = 100;
  53
  54         /**
  55          * Fetch the content type of the given url
  56          * @param string $url URL of the page
  57          * @return array content type
  58          */
  59         public static function getContentType(string $url)
  60         {
  61                 $curlResult = DI::httpRequest()->head($url);
  62                 if (!$curlResult->isSuccess()) {
  63                         return [];
  64                 }
  65
  66                 $contenttype =  $curlResult->getHeader('Content-Type');
  67                 if (empty($contenttype)) {
  68                         return [];
  69                 }
  70
  71                 return explode('/', current(explode(';', $contenttype)));
  72         }
  73
  74         /**
  75          * Search for chached embeddable data of an url otherwise fetch it
  76          *
  77          * @param string $url         The url of the page which should be scraped
  78          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
  79          *                            to avoid endless loops
  80          *
  81          * @return array which contains needed data for embedding
  82          *    string 'url'      => The url of the parsed page
  83          *    string 'type'     => Content type
  84          *    string 'title'    => (optional) The title of the content
  85          *    string 'text'     => (optional) The description for the content
  86          *    string 'image'    => (optional) A preview image of the content
  87          *    array  'images'   => (optional) Array of preview pictures
  88          *    string 'keywords' => (optional) The tags which belong to the content
  89          *
  90          * @throws HTTPException\InternalServerErrorException
  91          * @see   ParseUrl::getSiteinfo() for more information about scraping
  92          * embeddable content
  93          */
  94         public static function getSiteinfoCached($url, $do_oembed = true): array
  95         {
  96                 if (empty($url)) {
  97                         return [
  98                                 'url' => '',
  99                                 'type' => 'error',
 100                         ];
 101                 }
 102
 103                 $urlHash = hash('sha256', $url);
 104
 105                 $parsed_url = DBA::selectFirst('parsed_url', ['content'],
 106                         ['url_hash' => $urlHash, 'oembed' => $do_oembed]
 107                 );
 108                 if (!empty($parsed_url['content'])) {
 109                         $data = unserialize($parsed_url['content']);
 110                         return $data;
 111                 }
 112
 113                 $data = self::getSiteinfo($url, $do_oembed);
 114
 115                 $expires = $data['expires'];
 116
 117                 unset($data['expires']);
 118
 119                 DI::dba()->insert(
 120                         'parsed_url',
 121                         [
 122                                 'url_hash' => $urlHash,
 123                                 'oembed'   => $do_oembed,
 124                                 'url'      => $url,
 125                                 'content'  => serialize($data),
 126                                 'created'  => DateTimeFormat::utcNow(),
 127                                 'expires'  => $expires,
 128                         ],
 129                         Database::INSERT_UPDATE
 130                 );
 131
 132                 return $data;
 133         }
 134
 135         /**
 136          * Parse a page for embeddable content information
 137          *
 138          * This method parses to url for meta data which can be used to embed
 139          * the content. If available it prioritizes Open Graph meta tags.
 140          * If this is not available it uses the twitter cards meta tags.
 141          * As fallback it uses standard html elements with meta informations
 142          * like \<title\>Awesome Title\</title\> or
 143          * \<meta name="description" content="An awesome description"\>
 144          *
 145          * @param string $url         The url of the page which should be scraped
 146          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
 147          *                            to avoid endless loops
 148          * @param int    $count       Internal counter to avoid endless loops
 149          *
 150          * @return array which contains needed data for embedding
 151          *    string 'url'      => The url of the parsed page
 152          *    string 'type'     => Content type (error, link, photo, image, audio, video)
 153          *    string 'title'    => (optional) The title of the content
 154          *    string 'text'     => (optional) The description for the content
 155          *    string 'image'    => (optional) A preview image of the content
 156          *    array  'images'   => (optional) Array of preview pictures
 157          *    string 'keywords' => (optional) The tags which belong to the content
 158          *
 159          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 160          * @todo  https://developers.google.com/+/plugins/snippet/
 161          * @verbatim
 162          * <meta itemprop="name" content="Awesome title">
 163          * <meta itemprop="description" content="An awesome description">
 164          * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
 165          *
 166          * <body itemscope itemtype="http://schema.org/Product">
 167          *   <h1 itemprop="name">Shiny Trinket</h1>
 168          *   <img itemprop="image" src="{image-url}" />
 169          *   <p itemprop="description">Shiny trinkets are shiny.</p>
 170          * </body>
 171          * @endverbatim
 172          */
 173         public static function getSiteinfo($url, $do_oembed = true, $count = 1)
 174         {
 175                 if (empty($url)) {
 176                         return [
 177                                 'url' => '',
 178                                 'type' => 'error',
 179                         ];
 180                 }
 181
 182                 // Check if the URL does contain a scheme
 183                 $scheme = parse_url($url, PHP_URL_SCHEME);
 184
 185                 if ($scheme == '') {
 186                         $url = 'http://' . ltrim($url, '/');
 187                 }
 188
 189                 $url = trim($url, "'\"");
 190
 191                 $url = Network::stripTrackingQueryParams($url);
 192
 193                 $siteinfo = [
 194                         'url' => $url,
 195                         'type' => 'link',
 196                         'expires' => DateTimeFormat::utc(self::DEFAULT_EXPIRATION_FAILURE),
 197                 ];
 198
 199                 if ($count > 10) {
 200                         Logger::notice('Endless loop detected', ['url' => $url]);
 201                         return $siteinfo;
 202                 }
 203
 204                 $type = self::getContentType($url);
 205                 Logger::info('Got content-type', ['content-type' => $type, 'url' => $url]);
 206                 if (!empty($type) && in_array($type[0], ['image', 'video', 'audio'])) {
 207                         $siteinfo['type'] = $type[0];
 208                         return $siteinfo;
 209                 }
 210
 211                 if ((count($type) >= 2) && (($type[0] != 'text') || ($type[1] != 'html'))) {
 212                         Logger::info('Unparseable content-type, quitting here, ', ['content-type' => $type, 'url' => $url]);
 213                         return $siteinfo;
 214                 }
 215
 216                 $curlResult = DI::httpRequest()->get($url, false, ['content_length' => 1000000]);
 217                 if (!$curlResult->isSuccess() || empty($curlResult->getBody())) {
 218                         return $siteinfo;
 219                 }
 220
 221                 $siteinfo['expires'] = DateTimeFormat::utc(self::DEFAULT_EXPIRATION_SUCCESS);
 222
 223                 // If the file is too large then exit
 224                 if (($curlResult->getInfo()['download_content_length'] ?? 0) > 1000000) {
 225                         return $siteinfo;
 226                 }
 227
 228                 if ($cacheControlHeader = $curlResult->getHeader('Cache-Control')) {
 229                         if (preg_match('/max-age=([0-9]+)/i', $cacheControlHeader, $matches)) {
 230                                 $maxAge = max(86400, (int)array_pop($matches));
 231                                 $siteinfo['expires'] = DateTimeFormat::utc("now + $maxAge seconds");
 232                         }
 233                 }
 234
 235                 $body = $curlResult->getBody();
 236
 237                 if ($do_oembed) {
 238                         $oembed_data = OEmbed::fetchURL($url, false, false);
 239
 240                         if (!empty($oembed_data->type)) {
 241                                 if (!in_array($oembed_data->type, ['error', 'rich', 'image', 'video', 'audio', ''])) {
 242                                         $siteinfo['type'] = $oembed_data->type;
 243                                 }
 244
 245                                 // See https://github.com/friendica/friendica/pull/5763#discussion_r217913178
 246                                 if ($siteinfo['type'] != 'photo') {
 247                                         if (!empty($oembed_data->title)) {
 248                                                 $siteinfo['title'] = trim($oembed_data->title);
 249                                         }
 250                                         if (!empty($oembed_data->description)) {
 251                                                 $siteinfo['text'] = trim($oembed_data->description);
 252                                         }
 253                                         if (!empty($oembed_data->author_name)) {
 254                                                 $siteinfo['author_name'] = trim($oembed_data->author_name);
 255                                         }
 256                                         if (!empty($oembed_data->author_url)) {
 257                                                 $siteinfo['author_url'] = trim($oembed_data->author_url);
 258                                         }
 259                                         if (!empty($oembed_data->provider_name)) {
 260                                                 $siteinfo['publisher_name'] = trim($oembed_data->provider_name);
 261                                         }
 262                                         if (!empty($oembed_data->provider_url)) {
 263                                                 $siteinfo['publisher_url'] = trim($oembed_data->provider_url);
 264                                         }
 265                                         if (!empty($oembed_data->thumbnail_url)) {
 266                                                 $siteinfo['image'] = $oembed_data->thumbnail_url;
 267                                         }
 268                                 }
 269                         }
 270                 }
 271
 272                 $charset = '';
 273                 // Look for a charset, first in headers
 274                 // Expected form: Content-Type: text/html; charset=ISO-8859-4
 275                 foreach ($curlResult->getContentType() as $type) {
 276                         if (preg_match('/charset=([a-z0-9-_.\/]+)/i', $type, $matches)) {
 277                                 $charset = trim(trim(trim(array_pop($matches)), ';,'));
 278                         }
 279                 }
 280
 281                 // Then in body that gets precedence
 282                 // Expected forms:
 283                 // - <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
 284                 // - <meta charset="utf-8">
 285                 // - <meta charset=utf-8>
 286                 // - <meta charSet="utf-8">
 287                 // We escape <style> and <script> tags since they can contain irrelevant charset information
 288                 // (see https://github.com/friendica/friendica/issues/9251#issuecomment-698636806)
 289                 Strings::performWithEscapedBlocks($body, '#<(?:style|script).*?</(?:style|script)>#ism', function ($body) use (&$charset) {
 290                         if (preg_match('/charset=["\']?([a-z0-9-_.\/]+)/i', $body, $matches)) {
 291                                 $charset = trim(trim(trim(array_pop($matches)), ';,'));
 292                         }
 293                 });
 294
 295                 $siteinfo['charset'] = $charset;
 296
 297                 if ($charset && strtoupper($charset) != 'UTF-8') {
 298                         // See https://github.com/friendica/friendica/issues/5470#issuecomment-418351211
 299                         $charset = str_ireplace('latin-1', 'latin1', $charset);
 300
 301                         Logger::info('detected charset', ['charset' => $charset]);
 302                         $body = iconv($charset, 'UTF-8//TRANSLIT', $body);
 303                 }
 304
 305                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', 'UTF-8');
 306
 307                 $doc = new DOMDocument();
 308                 @$doc->loadHTML($body);
 309
 310                 XML::deleteNode($doc, 'style');
 311                 XML::deleteNode($doc, 'option');
 312                 XML::deleteNode($doc, 'h1');
 313                 XML::deleteNode($doc, 'h2');
 314                 XML::deleteNode($doc, 'h3');
 315                 XML::deleteNode($doc, 'h4');
 316                 XML::deleteNode($doc, 'h5');
 317                 XML::deleteNode($doc, 'h6');
 318                 XML::deleteNode($doc, 'ol');
 319                 XML::deleteNode($doc, 'ul');
 320
 321                 $xpath = new DOMXPath($doc);
 322
 323                 $list = $xpath->query('//meta[@content]');
 324                 foreach ($list as $node) {
 325                         $meta_tag = [];
 326                         if ($node->attributes->length) {
 327                                 foreach ($node->attributes as $attribute) {
 328                                         $meta_tag[$attribute->name] = $attribute->value;
 329                                 }
 330                         }
 331
 332                         if (@$meta_tag['http-equiv'] == 'refresh') {
 333                                 $path = $meta_tag['content'];
 334                                 $pathinfo = explode(';', $path);
 335                                 $content = '';
 336                                 foreach ($pathinfo as $value) {
 337                                         if (substr(strtolower($value), 0, 4) == 'url=') {
 338                                                 $content = substr($value, 4);
 339                                         }
 340                                 }
 341                                 if ($content != '') {
 342                                         $siteinfo = self::getSiteinfo($content, $do_oembed, ++$count);
 343                                         return $siteinfo;
 344                                 }
 345                         }
 346                 }
 347
 348                 $list = $xpath->query('//title');
 349                 if ($list->length > 0) {
 350                         $siteinfo['title'] = trim($list->item(0)->nodeValue);
 351                 }
 352
 353                 $twitter_card = false;
 354                 $twitter_image = false;
 355
 356                 $list = $xpath->query('//meta[@name]');
 357                 foreach ($list as $node) {
 358                         $meta_tag = [];
 359                         if ($node->attributes->length) {
 360                                 foreach ($node->attributes as $attribute) {
 361                                         $meta_tag[$attribute->name] = $attribute->value;
 362                                 }
 363                         }
 364
 365                         if (empty($meta_tag['content'])) {
 366                                 continue;
 367                         }
 368
 369                         $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 370
 371                         switch (strtolower($meta_tag['name'])) {
 372                                 case 'fulltitle':
 373                                         $siteinfo['title'] = trim($meta_tag['content']);
 374                                         break;
 375                                 case 'description':
 376                                         $siteinfo['text'] = trim($meta_tag['content']);
 377                                         break;
 378                                 case 'thumbnail':
 379                                         $siteinfo['image'] = $meta_tag['content'];
 380                                         break;
 381                                 case 'twitter:image':
 382                                         $siteinfo['image'] = $meta_tag['content'];
 383                                         $twitter_image = true;
 384                                         break;
 385                                 case 'twitter:image:src':
 386                                         $siteinfo['image'] = $meta_tag['content'];
 387                                         break;
 388                                 case 'twitter:card':
 389                                         // Detect photo pages
 390                                         if ($meta_tag['content'] == 'summary_large_image') {
 391                                                 $twitter_card = true;
 392                                         }
 393                                         break;
 394                                 case 'twitter:description':
 395                                         $siteinfo['text'] = trim($meta_tag['content']);
 396                                         break;
 397                                 case 'twitter:title':
 398                                         $siteinfo['title'] = trim($meta_tag['content']);
 399                                         break;
 400                                 case 'dc.title':
 401                                         $siteinfo['title'] = trim($meta_tag['content']);
 402                                         break;
 403                                 case 'dc.description':
 404                                         $siteinfo['text'] = trim($meta_tag['content']);
 405                                         break;
 406                                 case 'dc.creator':
 407                                         $siteinfo['publisher_name'] = trim($meta_tag['content']);
 408                                         break;
 409                                 case 'keywords':
 410                                         $keywords = explode(',', $meta_tag['content']);
 411                                         break;
 412                                 case 'news_keywords':
 413                                         $keywords = explode(',', $meta_tag['content']);
 414                                         break;
 415                         }
 416                 }
 417
 418                 if (isset($keywords)) {
 419                         $siteinfo['keywords'] = [];
 420                         foreach ($keywords as $keyword) {
 421                                 if (!in_array(trim($keyword), $siteinfo['keywords'])) {
 422                                         $siteinfo['keywords'][] = trim($keyword);
 423                                 }
 424                         }
 425                 }
 426
 427                 $list = $xpath->query('//meta[@property]');
 428                 foreach ($list as $node) {
 429                         $meta_tag = [];
 430                         if ($node->attributes->length) {
 431                                 foreach ($node->attributes as $attribute) {
 432                                         $meta_tag[$attribute->name] = $attribute->value;
 433                                 }
 434                         }
 435
 436                         if (!empty($meta_tag['content'])) {
 437                                 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 438
 439                                 switch (strtolower($meta_tag['property'])) {
 440                                         case 'og:image':
 441                                                 $siteinfo['image'] = $meta_tag['content'];
 442                                                 break;
 443                                         case 'og:image:url':
 444                                                 $siteinfo['image'] = $meta_tag['content'];
 445                                                 break;
 446                                         case 'og:image:secure_url':
 447                                                 $siteinfo['image'] = $meta_tag['content'];
 448                                                 break;
 449                                         case 'og:title':
 450                                                 $siteinfo['title'] = trim($meta_tag['content']);
 451                                                 break;
 452                                         case 'og:description':
 453                                                 $siteinfo['text'] = trim($meta_tag['content']);
 454                                                 break;
 455                                         case 'og:site_name':
 456                                                 $siteinfo['publisher_name'] = trim($meta_tag['content']);
 457                                                 break;
 458                                         case 'twitter:description':
 459                                                 $siteinfo['text'] = trim($meta_tag['content']);
 460                                                 break;
 461                                         case 'twitter:title':
 462                                                 $siteinfo['title'] = trim($meta_tag['content']);
 463                                                 break;
 464                                         case 'twitter:image':
 465                                                 $siteinfo['image'] = $meta_tag['content'];
 466                                                 $twitter_image = true;
 467                                                 break;
 468                                 }
 469                         }
 470                 }
 471
 472                 $list = $xpath->query("//script[@type='application/ld+json']");
 473                 foreach ($list as $node) {
 474                         if (!empty($node->nodeValue)) {
 475                                 if ($jsonld = json_decode($node->nodeValue, true)) {
 476                                         $siteinfo = self::parseParts($siteinfo, $jsonld);
 477                                 }
 478                         }
 479                 }
 480
 481 // Currently deactivated, see https://github.com/friendica/friendica/pull/10148#issuecomment-821512658
 482                 // Prevent to have a photo type without an image
 483 //              if ($twitter_card && $twitter_image && !empty($siteinfo['image'])) {
 484 //                      $siteinfo['type'] = 'photo';
 485 //              }
 486
 487                 if (!empty($siteinfo['image'])) {
 488                         $siteinfo['images'] = $siteinfo['images'] ?? [];
 489                         array_unshift($siteinfo['images'], ['url' => $siteinfo['image']]);
 490                         unset($siteinfo['image']);
 491                 }
 492
 493                 $siteinfo = self::checkMedia($url, $siteinfo);
 494
 495                 if (!empty($siteinfo['text']) && mb_strlen($siteinfo['text']) > self::MAX_DESC_COUNT) {
 496                         $siteinfo['text'] = mb_substr($siteinfo['text'], 0, self::MAX_DESC_COUNT) . '…';
 497                         $pos = mb_strrpos($siteinfo['text'], '.');
 498                         if ($pos > self::MIN_DESC_COUNT) {
 499                                 $siteinfo['text'] = mb_substr($siteinfo['text'], 0, $pos + 1);
 500                         }
 501                 }
 502
 503                 Logger::info('Siteinfo fetched', ['url' => $url, 'siteinfo' => $siteinfo]);
 504
 505                 Hook::callAll('getsiteinfo', $siteinfo);
 506
 507                 return $siteinfo;
 508         }
 509
 510         /**
 511          * Check the attached media elements.
 512          * Fix existing data and add missing data.
 513          *
 514          * @param string $page_url
 515          * @param array $siteinfo
 516          * @return void
 517          */
 518         private static function checkMedia(string $page_url, array $siteinfo)
 519         {
 520                 if (!empty($siteinfo['images'])) {
 521                         array_walk($siteinfo['images'], function (&$image) use ($page_url) {
 522                                 // According to the specifications someone could place a picture url into the content field as well.
 523                                 // But this doesn't seem to happen in the wild, so we don't cover it here.
 524                                 if (!empty($image['url'])) {
 525                                         $image['url'] = self::completeUrl($image['url'], $page_url);
 526                                         $photodata = Images::getInfoFromURLCached($image['url']);
 527                                         if (!empty($photodata) && ($photodata[0] > 50) && ($photodata[1] > 50)) {
 528                                                 $image['src'] = $image['url'];
 529                                                 $image['width'] = $photodata[0];
 530                                                 $image['height'] = $photodata[1];
 531                                                 $image['contenttype'] = $photodata['mime'];
 532                                                 unset($image['url']);
 533                                                 ksort($image);
 534                                         } else {
 535                                                 $image = [];
 536                                         }
 537                                 } else {
 538                                         $image = [];
 539                                 }
 540                         });
 541
 542                         $siteinfo['images'] = array_values(array_filter($siteinfo['images']));
 543                 }
 544
 545                 foreach (['audio', 'video'] as $element) {
 546                         if (!empty($siteinfo[$element])) {
 547                                 array_walk($siteinfo[$element], function (&$media) use ($page_url, &$siteinfo) {
 548                                         $url = '';
 549                                         $embed = '';
 550                                         $content = '';
 551                                         $contenttype = '';
 552                                         foreach (['embed', 'content', 'url'] as $field) {
 553                                                 if (!empty($media[$field])) {
 554                                                         $media[$field] = self::completeUrl($media[$field], $page_url);
 555                                                         $type = self::getContentType($media[$field]);
 556                                                         if (($type[0] ?? '') == 'text') {
 557                                                                 if ($field == 'embed') {
 558                                                                         $embed = $media[$field];
 559                                                                 } else {
 560                                                                         $url = $media[$field];
 561                                                                 }
 562                                                         } elseif (!empty($type[0])) {
 563                                                                 $content = $media[$field];
 564                                                                 $contenttype = implode('/', $type);
 565                                                         }
 566                                                 }
 567                                                 unset($media[$field]);
 568                                         }
 569
 570                                         foreach (['image', 'preview'] as $field) {
 571                                                 if (!empty($media[$field])) {
 572                                                         $media[$field] = self::completeUrl($media[$field], $page_url);
 573                                                 }
 574                                         }
 575
 576                                         if (!empty($url)) {
 577                                                 $media['url'] = $url;
 578                                         }
 579                                         if (!empty($embed)) {
 580                                                 $media['embed'] = $embed;
 581                                                 if (!empty($media['main'])) {
 582                                                         $siteinfo['embed'] = $embed;
 583                                                 }
 584                                         }
 585                                         if (!empty($content)) {
 586                                                 $media['src'] = $content;
 587                                         }
 588                                         if (!empty($contenttype)) {
 589                                                 $media['contenttype'] = $contenttype;
 590                                         }
 591                                         if (empty($url) && empty($content) && empty($embed)) {
 592                                                 $media = [];
 593                                         }
 594                                         ksort($media);
 595                                 });
 596
 597                                 $siteinfo[$element] = array_values(array_filter($siteinfo[$element]));
 598                         }
 599                         if (empty($siteinfo[$element])) {
 600                                 unset($siteinfo[$element]);
 601                         }
 602                 }
 603                 return $siteinfo;
 604         }
 605
 606         /**
 607          * Convert tags from CSV to an array
 608          *
 609          * @param string $string Tags
 610          * @return array with formatted Hashtags
 611          */
 612         public static function convertTagsToArray($string)
 613         {
 614                 $arr_tags = str_getcsv($string);
 615                 if (count($arr_tags)) {
 616                         // add the # sign to every tag
 617                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 618
 619                         return $arr_tags;
 620                 }
 621         }
 622
 623         /**
 624          * Add a hasht sign to a string
 625          *
 626          * This method is used as callback function
 627          *
 628          * @param string $tag The pure tag name
 629          * @param int    $k   Counter for internal use
 630          * @return void
 631          */
 632         private static function arrAddHashes(&$tag, $k)
 633         {
 634                 $tag = "#" . $tag;
 635         }
 636
 637         /**
 638          * Add a scheme to an url
 639          *
 640          * The src attribute of some html elements (e.g. images)
 641          * can miss the scheme so we need to add the correct
 642          * scheme
 643          *
 644          * @param string $url    The url which possibly does have
 645          *                       a missing scheme (a link to an image)
 646          * @param string $scheme The url with a correct scheme
 647          *                       (e.g. the url from the webpage which does contain the image)
 648          *
 649          * @return string The url with a scheme
 650          */
 651         private static function completeUrl($url, $scheme)
 652         {
 653                 $urlarr = parse_url($url);
 654
 655                 // If the url does allready have an scheme
 656                 // we can stop the process here
 657                 if (isset($urlarr["scheme"])) {
 658                         return($url);
 659                 }
 660
 661                 $schemearr = parse_url($scheme);
 662
 663                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 664
 665                 if (!empty($schemearr["port"])) {
 666                         $complete .= ":".$schemearr["port"];
 667                 }
 668
 669                 if (!empty($urlarr["path"])) {
 670                         if (strpos($urlarr["path"], "/") !== 0) {
 671                                 $complete .= "/";
 672                         }
 673
 674                         $complete .= $urlarr["path"];
 675                 }
 676
 677                 if (!empty($urlarr["query"])) {
 678                         $complete .= "?".$urlarr["query"];
 679                 }
 680
 681                 if (!empty($urlarr["fragment"])) {
 682                         $complete .= "#".$urlarr["fragment"];
 683                 }
 684
 685                 return($complete);
 686         }
 687
 688         /**
 689          * Parse the Json-Ld parts of a web page
 690          *
 691          * @param array $siteinfo
 692          * @param array $jsonld
 693          * @return array siteinfo
 694          */
 695         private static function parseParts(array $siteinfo, array $jsonld)
 696         {
 697                 if (!empty($jsonld['@graph']) && is_array($jsonld['@graph'])) {
 698                         foreach ($jsonld['@graph'] as $part) {
 699                                 if (!empty($part) && is_array($part)) {
 700                                         $siteinfo = self::parseParts($siteinfo, $part);
 701                                 }
 702                         }
 703                 } elseif (!empty($jsonld['@type'])) {
 704                         $siteinfo = self::parseJsonLd($siteinfo, $jsonld);
 705                 } elseif (!empty($jsonld)) {
 706                         $keys = array_keys($jsonld);
 707                         $numeric_keys = true;
 708                         foreach ($keys as $key) {
 709                                 if (!is_int($key)) {
 710                                         $numeric_keys = false;
 711                                 }
 712                         }
 713                         if ($numeric_keys) {
 714                                 foreach ($jsonld as $part) {
 715                                         if (!empty($part) && is_array($part)) {
 716                                                 $siteinfo = self::parseParts($siteinfo, $part);
 717                                         }
 718                                 }
 719                         }
 720                 }
 721
 722                 array_walk_recursive($siteinfo, function (&$element) {
 723                         if (is_string($element)) {
 724                                 $element = trim(strip_tags(html_entity_decode($element, ENT_COMPAT, 'UTF-8')));
 725                         }
 726                 });
 727
 728                 return $siteinfo;
 729         }
 730
 731         /**
 732          * Improve the siteinfo with information from the provided JSON-LD information
 733          * @see https://jsonld.com/
 734          * @see https://schema.org/
 735          *
 736          * @param array $siteinfo
 737          * @param array $jsonld
 738          * @return array siteinfo
 739          */
 740         private static function parseJsonLd(array $siteinfo, array $jsonld)
 741         {
 742                 $type = JsonLD::fetchElement($jsonld, '@type');
 743                 if (empty($type)) {
 744                         Logger::info('Empty type', ['url' => $siteinfo['url']]);
 745                         return $siteinfo;
 746                 }
 747
 748                 // Silently ignore some types that aren't processed
 749                 if (in_array($type, ['SiteNavigationElement', 'JobPosting', 'CreativeWork', 'MusicAlbum',
 750                         'WPHeader', 'WPSideBar', 'WPFooter', 'LegalService', 'MusicRecording',
 751                         'ItemList', 'BreadcrumbList', 'Blog', 'Dataset', 'Product'])) {
 752                         return $siteinfo;
 753                 }
 754
 755                 switch ($type) {
 756                         case 'Article':
 757                         case 'AdvertiserContentArticle':
 758                         case 'NewsArticle':
 759                         case 'Report':
 760                         case 'SatiricalArticle':
 761                         case 'ScholarlyArticle':
 762                         case 'SocialMediaPosting':
 763                         case 'TechArticle':
 764                         case 'ReportageNewsArticle':
 765                         case 'SocialMediaPosting':
 766                         case 'BlogPosting':
 767                         case 'LiveBlogPosting':
 768                         case 'DiscussionForumPosting':
 769                                 return self::parseJsonLdArticle($siteinfo, $jsonld);
 770                         case 'WebPage':
 771                         case 'AboutPage':
 772                         case 'CheckoutPage':
 773                         case 'CollectionPage':
 774                         case 'ContactPage':
 775                         case 'FAQPage':
 776                         case 'ItemPage':
 777                         case 'MedicalWebPage':
 778                         case 'ProfilePage':
 779                         case 'QAPage':
 780                         case 'RealEstateListing':
 781                         case 'SearchResultsPage':
 782                         case 'MediaGallery':
 783                         case 'ImageGallery':
 784                         case 'VideoGallery':
 785                         case 'RadioEpisode':
 786                         case 'Event':
 787                                 return self::parseJsonLdWebPage($siteinfo, $jsonld);
 788                         case 'WebSite':
 789                                 return self::parseJsonLdWebSite($siteinfo, $jsonld);
 790                         case 'Organization':
 791                         case 'Airline':
 792                         case 'Consortium':
 793                         case 'Corporation':
 794                         case 'EducationalOrganization':
 795                         case 'FundingScheme':
 796                         case 'GovernmentOrganization':
 797                         case 'LibrarySystem':
 798                         case 'LocalBusiness':
 799                         case 'MedicalOrganization':
 800                         case 'NGO':
 801                         case 'NewsMediaOrganization':
 802                         case 'Project':
 803                         case 'SportsOrganization':
 804                         case 'WorkersUnion':
 805                                 return self::parseJsonLdWebOrganization($siteinfo, $jsonld);
 806                         case 'Person':
 807                         case 'Patient':
 808                         case 'PerformingGroup':
 809                         case 'DanceGroup';
 810                         case 'MusicGroup':
 811                         case 'TheaterGroup':
 812                                 return self::parseJsonLdWebPerson($siteinfo, $jsonld);
 813                         case 'AudioObject':
 814                         case 'Audio':
 815                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'audio');
 816                         case 'VideoObject':
 817                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'video');
 818                         case 'ImageObject':
 819                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'images');
 820                         default:
 821                                 Logger::info('Unknown type', ['type' => $type, 'url' => $siteinfo['url']]);
 822                                 return $siteinfo;
 823                 }
 824         }
 825
 826         /**
 827          * Fetch author and publisher data
 828          *
 829          * @param array $siteinfo
 830          * @param array $jsonld
 831          * @return array siteinfo
 832          */
 833         private static function parseJsonLdAuthor(array $siteinfo, array $jsonld)
 834         {
 835                 $jsonldinfo = [];
 836
 837                 if (!empty($jsonld['publisher']) && is_array($jsonld['publisher'])) {
 838                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'name');
 839                         if (!empty($content) && is_string($content)) {
 840                                 $jsonldinfo['publisher_name'] = trim($content);
 841                         }
 842
 843                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'url');
 844                         if (!empty($content) && is_string($content)) {
 845                                 $jsonldinfo['publisher_url'] = trim($content);
 846                         }
 847
 848                         $brand = JsonLD::fetchElement($jsonld, 'publisher', 'brand', '@type', 'Organization');
 849                         if (!empty($brand) && is_array($brand)) {
 850                                 $content = JsonLD::fetchElement($brand, 'name');
 851                                 if (!empty($content) && is_string($content)) {
 852                                         $jsonldinfo['publisher_name'] = trim($content);
 853                                 }
 854
 855                                 $content = JsonLD::fetchElement($brand, 'url');
 856                                 if (!empty($content) && is_string($content)) {
 857                                         $jsonldinfo['publisher_url'] = trim($content);
 858                                 }
 859
 860                                 $content = JsonLD::fetchElement($brand, 'logo', 'url');
 861                                 if (!empty($content) && is_string($content)) {
 862                                         $jsonldinfo['publisher_img'] = trim($content);
 863                                 }
 864                         }
 865
 866                         $logo = JsonLD::fetchElement($jsonld, 'publisher', 'logo');
 867                         if (!empty($logo) && is_array($logo)) {
 868                                 $content = JsonLD::fetchElement($logo, 'url');
 869                                 if (!empty($content) && is_string($content)) {
 870                                         $jsonldinfo['publisher_img'] = trim($content);
 871                                 }
 872                         }
 873                 } elseif (!empty($jsonld['publisher']) && is_string($jsonld['publisher'])) {
 874                         $jsonldinfo['publisher_name'] = trim($jsonld['publisher']);
 875                 }
 876
 877                 if (!empty($jsonld['author']) && is_array($jsonld['author'])) {
 878                         $content = JsonLD::fetchElement($jsonld, 'author', 'name');
 879                         if (!empty($content) && is_string($content)) {
 880                                 $jsonldinfo['author_name'] = trim($content);
 881                         }
 882
 883                         $content = JsonLD::fetchElement($jsonld, 'author', 'sameAs');
 884                         if (!empty($content) && is_string($content)) {
 885                                 $jsonldinfo['author_url'] = trim($content);
 886                         }
 887
 888                         $content = JsonLD::fetchElement($jsonld, 'author', 'url');
 889                         if (!empty($content) && is_string($content)) {
 890                                 $jsonldinfo['author_url'] = trim($content);
 891                         }
 892
 893                         $logo = JsonLD::fetchElement($jsonld, 'author', 'logo');
 894                         if (!empty($logo) && is_array($logo)) {
 895                                 $content = JsonLD::fetchElement($logo, 'url');
 896                                 if (!empty($content) && is_string($content)) {
 897                                         $jsonldinfo['author_img'] = trim($content);
 898                                 }
 899                         }
 900                 } elseif (!empty($jsonld['author']) && is_string($jsonld['author'])) {
 901                         $jsonldinfo['author_name'] = trim($jsonld['author']);
 902                 }
 903
 904                 Logger::info('Fetched Author information', ['fetched' => $jsonldinfo]);
 905
 906                 return array_merge($siteinfo, $jsonldinfo);
 907         }
 908
 909         /**
 910          * Fetch data from the provided JSON-LD Article type
 911          * @see https://schema.org/Article
 912          *
 913          * @param array $siteinfo
 914          * @param array $jsonld
 915          * @return array siteinfo
 916          */
 917         private static function parseJsonLdArticle(array $siteinfo, array $jsonld)
 918         {
 919                 $jsonldinfo = [];
 920
 921                 $content = JsonLD::fetchElement($jsonld, 'headline');
 922                 if (!empty($content) && is_string($content)) {
 923                         $jsonldinfo['title'] = trim($content);
 924                 }
 925
 926                 $content = JsonLD::fetchElement($jsonld, 'alternativeHeadline');
 927                 if (!empty($content) && is_string($content) && (($jsonldinfo['title'] ?? '') != trim($content))) {
 928                         $jsonldinfo['alternative_title'] = trim($content);
 929                 }
 930
 931                 $content = JsonLD::fetchElement($jsonld, 'description');
 932                 if (!empty($content) && is_string($content)) {
 933                         $jsonldinfo['text'] = trim($content);
 934                 }
 935
 936                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 937                 if (!empty($content)) {
 938                         $jsonldinfo['image'] = trim($content);
 939                 }
 940
 941                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
 942                 if (!empty($content) && is_string($content)) {
 943                         $jsonldinfo['image'] = trim($content);
 944                 }
 945
 946                 if (!empty($jsonld['keywords']) && !is_array($jsonld['keywords'])) {
 947                         $content = JsonLD::fetchElement($jsonld, 'keywords');
 948                         if (!empty($content)) {
 949                                 $siteinfo['keywords'] = [];
 950                                 $keywords = explode(',', $content);
 951                                 foreach ($keywords as $keyword) {
 952                                         $siteinfo['keywords'][] = trim($keyword);
 953                                 }
 954                         }
 955                 } elseif (!empty($jsonld['keywords'])) {
 956                         $content = JsonLD::fetchElementArray($jsonld, 'keywords');
 957                         if (!empty($content) && is_array($content)) {
 958                                 $jsonldinfo['keywords'] = $content;
 959                         }
 960                 }
 961
 962                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 963
 964                 Logger::info('Fetched article information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 965
 966                 return array_merge($siteinfo, $jsonldinfo);
 967         }
 968
 969         /**
 970          * Fetch data from the provided JSON-LD WebPage type
 971          * @see https://schema.org/WebPage
 972          *
 973          * @param array $siteinfo
 974          * @param array $jsonld
 975          * @return array siteinfo
 976          */
 977         private static function parseJsonLdWebPage(array $siteinfo, array $jsonld)
 978         {
 979                 $jsonldinfo = [];
 980
 981                 $content = JsonLD::fetchElement($jsonld, 'name');
 982                 if (!empty($content)) {
 983                         $jsonldinfo['title'] = trim($content);
 984                 }
 985
 986                 $content = JsonLD::fetchElement($jsonld, 'description');
 987                 if (!empty($content)) {
 988                         $jsonldinfo['text'] = trim($content);
 989                 }
 990
 991                 $content = JsonLD::fetchElement($jsonld, 'image');
 992                 if (!empty($content)) {
 993                         $jsonldinfo['image'] = trim($content);
 994                 }
 995
 996                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 997                 if (!empty($content)) {
 998                         $jsonldinfo['image'] = trim($content);
 999                 }
1000
1001                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
1002
1003                 Logger::info('Fetched WebPage information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1004
1005                 return array_merge($siteinfo, $jsonldinfo);
1006         }
1007
1008         /**
1009          * Fetch data from the provided JSON-LD WebSite type
1010          * @see https://schema.org/WebSite
1011          *
1012          * @param array $siteinfo
1013          * @param array $jsonld
1014          * @return array siteinfo
1015          */
1016         private static function parseJsonLdWebSite(array $siteinfo, array $jsonld)
1017         {
1018                 $jsonldinfo = [];
1019
1020                 $content = JsonLD::fetchElement($jsonld, 'name');
1021                 if (!empty($content)) {
1022                         $jsonldinfo['publisher_name'] = trim($content);
1023                 }
1024
1025                 $content = JsonLD::fetchElement($jsonld, 'description');
1026                 if (!empty($content)) {
1027                         $jsonldinfo['publisher_description'] = trim($content);
1028                 }
1029
1030                 $content = JsonLD::fetchElement($jsonld, 'url');
1031                 if (!empty($content)) {
1032                         $jsonldinfo['publisher_url'] = trim($content);
1033                 }
1034
1035                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
1036                 if (!empty($content)) {
1037                         $jsonldinfo['image'] = trim($content);
1038                 }
1039
1040                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
1041
1042                 Logger::info('Fetched WebSite information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1043                 return array_merge($siteinfo, $jsonldinfo);
1044         }
1045
1046         /**
1047          * Fetch data from the provided JSON-LD Organization type
1048          * @see https://schema.org/Organization
1049          *
1050          * @param array $siteinfo
1051          * @param array $jsonld
1052          * @return array siteinfo
1053          */
1054         private static function parseJsonLdWebOrganization(array $siteinfo, array $jsonld)
1055         {
1056                 $jsonldinfo = [];
1057
1058                 $content = JsonLD::fetchElement($jsonld, 'name');
1059                 if (!empty($content)) {
1060                         $jsonldinfo['publisher_name'] = trim($content);
1061                 }
1062
1063                 $content = JsonLD::fetchElement($jsonld, 'description');
1064                 if (!empty($content)) {
1065                         $jsonldinfo['publisher_description'] = trim($content);
1066                 }
1067
1068                 $content = JsonLD::fetchElement($jsonld, 'url');
1069                 if (!empty($content)) {
1070                         $jsonldinfo['publisher_url'] = trim($content);
1071                 }
1072
1073                 $content = JsonLD::fetchElement($jsonld, 'logo', 'url', '@type', 'ImageObject');
1074                 if (!empty($content)) {
1075                         $jsonldinfo['publisher_img'] = trim($content);
1076                 }
1077
1078                 $content = JsonLD::fetchElement($jsonld, 'brand', 'name', '@type', 'Organization');
1079                 if (!empty($content)) {
1080                         $jsonldinfo['publisher_name'] = trim($content);
1081                 }
1082
1083                 $content = JsonLD::fetchElement($jsonld, 'brand', 'url', '@type', 'Organization');
1084                 if (!empty($content)) {
1085                         $jsonldinfo['publisher_url'] = trim($content);
1086                 }
1087
1088                 Logger::info('Fetched Organization information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1089                 return array_merge($siteinfo, $jsonldinfo);
1090         }
1091
1092         /**
1093          * Fetch data from the provided JSON-LD Person type
1094          * @see https://schema.org/Person
1095          *
1096          * @param array $siteinfo
1097          * @param array $jsonld
1098          * @return array siteinfo
1099          */
1100         private static function parseJsonLdWebPerson(array $siteinfo, array $jsonld)
1101         {
1102                 $jsonldinfo = [];
1103
1104                 $content = JsonLD::fetchElement($jsonld, 'name');
1105                 if (!empty($content)) {
1106                         $jsonldinfo['author_name'] = trim($content);
1107                 }
1108
1109                 $content = JsonLD::fetchElement($jsonld, 'description');
1110                 if (!empty($content)) {
1111                         $jsonldinfo['author_description'] = trim($content);
1112                 }
1113
1114                 $content = JsonLD::fetchElement($jsonld, 'sameAs');
1115                 if (!empty($content) && is_string($content)) {
1116                         $jsonldinfo['author_url'] = trim($content);
1117                 }
1118
1119                 $content = JsonLD::fetchElement($jsonld, 'url');
1120                 if (!empty($content)) {
1121                         $jsonldinfo['author_url'] = trim($content);
1122                 }
1123
1124                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
1125                 if (!empty($content)) {
1126                         $jsonldinfo['author_img'] = trim($content);
1127                 }
1128
1129                 Logger::info('Fetched Person information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1130                 return array_merge($siteinfo, $jsonldinfo);
1131         }
1132
1133         /**
1134          * Fetch data from the provided JSON-LD MediaObject type
1135          * @see https://schema.org/MediaObject
1136          *
1137          * @param array $siteinfo
1138          * @param array $jsonld
1139          * @return array siteinfo
1140          */
1141         private static function parseJsonLdMediaObject(array $siteinfo, array $jsonld, string $name)
1142         {
1143                 $media = [];
1144
1145                 $content = JsonLD::fetchElement($jsonld, 'caption');
1146                 if (!empty($content)) {
1147                         $media['caption'] = trim($content);
1148                 }
1149
1150                 $content = JsonLD::fetchElement($jsonld, 'url');
1151                 if (!empty($content)) {
1152                         $media['url'] = trim($content);
1153                 }
1154
1155                 $content = JsonLD::fetchElement($jsonld, 'mainEntityOfPage');
1156                 if (!empty($content)) {
1157                         $media['main'] = Strings::compareLink($content, $siteinfo['url']);
1158                 }
1159
1160                 $content = JsonLD::fetchElement($jsonld, 'description');
1161                 if (!empty($content)) {
1162                         $media['description'] = trim($content);
1163                 }
1164
1165                 $content = JsonLD::fetchElement($jsonld, 'name');
1166                 if (!empty($content) && (($media['description'] ?? '') != trim($content))) {
1167                         $media['name'] = trim($content);
1168                 }
1169
1170                 $content = JsonLD::fetchElement($jsonld, 'contentUrl');
1171                 if (!empty($content)) {
1172                         $media['content'] = trim($content);
1173                 }
1174
1175                 $content = JsonLD::fetchElement($jsonld, 'embedUrl');
1176                 if (!empty($content)) {
1177                         $media['embed'] = trim($content);
1178                 }
1179
1180                 $content = JsonLD::fetchElement($jsonld, 'height');
1181                 if (!empty($content)) {
1182                         $media['height'] = trim($content);
1183                 }
1184
1185                 $content = JsonLD::fetchElement($jsonld, 'width');
1186                 if (!empty($content)) {
1187                         $media['width'] = trim($content);
1188                 }
1189
1190                 $content = JsonLD::fetchElement($jsonld, 'image');
1191                 if (!empty($content)) {
1192                         $media['image'] = trim($content);
1193                 }
1194
1195                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
1196                 if (!empty($content) && (($media['image'] ?? '') != trim($content))) {
1197                         if (!empty($media['image'])) {
1198                                 $media['preview'] = trim($content);
1199                         } else {
1200                                 $media['image'] = trim($content);
1201                         }
1202                 }
1203
1204                 Logger::info('Fetched Media information', ['url' => $siteinfo['url'], 'fetched' => $media]);
1205                 $siteinfo[$name][] = $media;
1206                 return $siteinfo;
1207         }
1208 }