src/Util/ParseUrl.php

   1 <?php
   2 /**
   3  * @copyright Copyright (C) 2020, Friendica
   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::log('Endless loop detected for ' . $url, Logger::DEBUG);
 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);
 217                 if (!$curlResult->isSuccess()) {
 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                 $header = $curlResult->getHeader();
 236                 $body = $curlResult->getBody();
 237
 238                 if ($do_oembed) {
 239                         $oembed_data = OEmbed::fetchURL($url, false, false);
 240
 241                         if (!empty($oembed_data->type)) {
 242                                 if (!in_array($oembed_data->type, ['error', 'rich', 'image', 'video', 'audio', ''])) {
 243                                         $siteinfo['type'] = $oembed_data->type;
 244                                 }
 245
 246                                 // See https://github.com/friendica/friendica/pull/5763#discussion_r217913178
 247                                 if ($siteinfo['type'] != 'photo') {
 248                                         if (!empty($oembed_data->title)) {
 249                                                 $siteinfo['title'] = trim($oembed_data->title);
 250                                         }
 251                                         if (!empty($oembed_data->description)) {
 252                                                 $siteinfo['text'] = trim($oembed_data->description);
 253                                         }
 254                                         if (!empty($oembed_data->author_name)) {
 255                                                 $siteinfo['author_name'] = trim($oembed_data->author_name);
 256                                         }
 257                                         if (!empty($oembed_data->author_url)) {
 258                                                 $siteinfo['author_url'] = trim($oembed_data->author_url);
 259                                         }
 260                                         if (!empty($oembed_data->provider_name)) {
 261                                                 $siteinfo['publisher_name'] = trim($oembed_data->provider_name);
 262                                         }
 263                                         if (!empty($oembed_data->provider_url)) {
 264                                                 $siteinfo['publisher_url'] = trim($oembed_data->provider_url);
 265                                         }
 266                                         if (!empty($oembed_data->thumbnail_url)) {
 267                                                 $siteinfo['image'] = $oembed_data->thumbnail_url;
 268                                         }
 269                                 }
 270                         }
 271                 }
 272
 273                 $charset = '';
 274                 // Look for a charset, first in headers
 275                 // Expected form: Content-Type: text/html; charset=ISO-8859-4
 276                 if (preg_match('/charset=([a-z0-9-_.\/]+)/i', $header, $matches)) {
 277                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 278                 }
 279
 280                 // Then in body that gets precedence
 281                 // Expected forms:
 282                 // - <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
 283                 // - <meta charset="utf-8">
 284                 // - <meta charset=utf-8>
 285                 // - <meta charSet="utf-8">
 286                 // We escape <style> and <script> tags since they can contain irrelevant charset information
 287                 // (see https://github.com/friendica/friendica/issues/9251#issuecomment-698636806)
 288                 Strings::performWithEscapedBlocks($body, '#<(?:style|script).*?</(?:style|script)>#ism', function ($body) use (&$charset) {
 289                         if (preg_match('/charset=["\']?([a-z0-9-_.\/]+)/i', $body, $matches)) {
 290                                 $charset = trim(trim(trim(array_pop($matches)), ';,'));
 291                         }
 292                 });
 293
 294                 $siteinfo['charset'] = $charset;
 295
 296                 if ($charset && strtoupper($charset) != 'UTF-8') {
 297                         // See https://github.com/friendica/friendica/issues/5470#issuecomment-418351211
 298                         $charset = str_ireplace('latin-1', 'latin1', $charset);
 299
 300                         Logger::log('detected charset ' . $charset, Logger::DEBUG);
 301                         $body = iconv($charset, 'UTF-8//TRANSLIT', $body);
 302                 }
 303
 304                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', 'UTF-8');
 305
 306                 $doc = new DOMDocument();
 307                 @$doc->loadHTML($body);
 308
 309                 XML::deleteNode($doc, 'style');
 310                 XML::deleteNode($doc, 'option');
 311                 XML::deleteNode($doc, 'h1');
 312                 XML::deleteNode($doc, 'h2');
 313                 XML::deleteNode($doc, 'h3');
 314                 XML::deleteNode($doc, 'h4');
 315                 XML::deleteNode($doc, 'h5');
 316                 XML::deleteNode($doc, 'h6');
 317                 XML::deleteNode($doc, 'ol');
 318                 XML::deleteNode($doc, 'ul');
 319
 320                 $xpath = new DOMXPath($doc);
 321
 322                 $list = $xpath->query('//meta[@content]');
 323                 foreach ($list as $node) {
 324                         $meta_tag = [];
 325                         if ($node->attributes->length) {
 326                                 foreach ($node->attributes as $attribute) {
 327                                         $meta_tag[$attribute->name] = $attribute->value;
 328                                 }
 329                         }
 330
 331                         if (@$meta_tag['http-equiv'] == 'refresh') {
 332                                 $path = $meta_tag['content'];
 333                                 $pathinfo = explode(';', $path);
 334                                 $content = '';
 335                                 foreach ($pathinfo as $value) {
 336                                         if (substr(strtolower($value), 0, 4) == 'url=') {
 337                                                 $content = substr($value, 4);
 338                                         }
 339                                 }
 340                                 if ($content != '') {
 341                                         $siteinfo = self::getSiteinfo($content, $do_oembed, ++$count);
 342                                         return $siteinfo;
 343                                 }
 344                         }
 345                 }
 346
 347                 $list = $xpath->query('//title');
 348                 if ($list->length > 0) {
 349                         $siteinfo['title'] = trim($list->item(0)->nodeValue);
 350                 }
 351
 352                 $list = $xpath->query('//meta[@name]');
 353                 foreach ($list as $node) {
 354                         $meta_tag = [];
 355                         if ($node->attributes->length) {
 356                                 foreach ($node->attributes as $attribute) {
 357                                         $meta_tag[$attribute->name] = $attribute->value;
 358                                 }
 359                         }
 360
 361                         if (empty($meta_tag['content'])) {
 362                                 continue;
 363                         }
 364
 365                         $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 366
 367                         switch (strtolower($meta_tag['name'])) {
 368                                 case 'fulltitle':
 369                                         $siteinfo['title'] = trim($meta_tag['content']);
 370                                         break;
 371                                 case 'description':
 372                                         $siteinfo['text'] = trim($meta_tag['content']);
 373                                         break;
 374                                 case 'thumbnail':
 375                                         $siteinfo['image'] = $meta_tag['content'];
 376                                         break;
 377                                 case 'twitter:image':
 378                                         $siteinfo['image'] = $meta_tag['content'];
 379                                         break;
 380                                 case 'twitter:image:src':
 381                                         $siteinfo['image'] = $meta_tag['content'];
 382                                         break;
 383                                 case 'twitter:card':
 384                                         // Detect photo pages
 385                                         if ($meta_tag['content'] == 'summary_large_image') {
 386                                                 $siteinfo['type'] = 'photo';
 387                                         }
 388                                         break;
 389                                 case 'twitter:description':
 390                                         $siteinfo['text'] = trim($meta_tag['content']);
 391                                         break;
 392                                 case 'twitter:title':
 393                                         $siteinfo['title'] = trim($meta_tag['content']);
 394                                         break;
 395                                 case 'dc.title':
 396                                         $siteinfo['title'] = trim($meta_tag['content']);
 397                                         break;
 398                                 case 'dc.description':
 399                                         $siteinfo['text'] = trim($meta_tag['content']);
 400                                         break;
 401                                 case 'dc.creator':
 402                                         $siteinfo['publisher_name'] = trim($meta_tag['content']);
 403                                         break;
 404                                 case 'keywords':
 405                                         $keywords = explode(',', $meta_tag['content']);
 406                                         break;
 407                                 case 'news_keywords':
 408                                         $keywords = explode(',', $meta_tag['content']);
 409                                         break;
 410                         }
 411                 }
 412
 413                 if (isset($keywords)) {
 414                         $siteinfo['keywords'] = [];
 415                         foreach ($keywords as $keyword) {
 416                                 if (!in_array(trim($keyword), $siteinfo['keywords'])) {
 417                                         $siteinfo['keywords'][] = trim($keyword);
 418                                 }
 419                         }
 420                 }
 421
 422                 $list = $xpath->query('//meta[@property]');
 423                 foreach ($list as $node) {
 424                         $meta_tag = [];
 425                         if ($node->attributes->length) {
 426                                 foreach ($node->attributes as $attribute) {
 427                                         $meta_tag[$attribute->name] = $attribute->value;
 428                                 }
 429                         }
 430
 431                         if (!empty($meta_tag['content'])) {
 432                                 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 433
 434                                 switch (strtolower($meta_tag['property'])) {
 435                                         case 'og:image':
 436                                                 $siteinfo['image'] = $meta_tag['content'];
 437                                                 break;
 438                                         case 'og:image:url':
 439                                                 $siteinfo['image'] = $meta_tag['content'];
 440                                                 break;
 441                                         case 'og:image:secure_url':
 442                                                 $siteinfo['image'] = $meta_tag['content'];
 443                                                 break;
 444                                         case 'og:title':
 445                                                 $siteinfo['title'] = trim($meta_tag['content']);
 446                                                 break;
 447                                         case 'og:description':
 448                                                 $siteinfo['text'] = trim($meta_tag['content']);
 449                                                 break;
 450                                         case 'og:site_name':
 451                                                 $siteinfo['publisher_name'] = trim($meta_tag['content']);
 452                                                 break;
 453                                         case 'twitter:description':
 454                                                 $siteinfo['text'] = trim($meta_tag['content']);
 455                                                 break;
 456                                         case 'twitter:title':
 457                                                 $siteinfo['title'] = trim($meta_tag['content']);
 458                                                 break;
 459                                         case 'twitter:image':
 460                                                 $siteinfo['image'] = $meta_tag['content'];
 461                                                 break;
 462                                 }
 463                         }
 464                 }
 465
 466                 $list = $xpath->query("//script[@type='application/ld+json']");
 467                 foreach ($list as $node) {
 468                         if (!empty($node->nodeValue)) {
 469                                 $nodevalue = html_entity_decode($node->nodeValue, ENT_COMPAT, 'UTF-8');
 470                                 if ($jsonld = json_decode($nodevalue, true)) {
 471                                         $siteinfo = self::parseParts($siteinfo, $jsonld);
 472                                 }
 473                         }
 474                 }
 475
 476                 // Prevent to have a photo type without an image
 477                 if ((empty($siteinfo['image']) || !empty($siteinfo['text'])) && ($siteinfo['type'] == 'photo')) {
 478                         $siteinfo['type'] = 'link';
 479                 }
 480
 481                 if (!empty($siteinfo['image']) && empty($siteinfo['images'])) {
 482                         $src = self::completeUrl($siteinfo['image'], $url);
 483
 484                         unset($siteinfo['image']);
 485
 486                         $photodata = Images::getInfoFromURLCached($src);
 487
 488                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 489                                 $siteinfo['images'][] = ['src' => $src,
 490                                         'width' => $photodata[0],
 491                                         'height' => $photodata[1]];
 492                         }
 493                 }
 494
 495                 if (!empty($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          * Parse the Json-Ld parts
 512          *
 513          * @param array $siteinfo
 514          * @param array $jsonld
 515          * @return array siteinfo
 516          */
 517         private static function parseParts(array $siteinfo, array $jsonld)
 518         {
 519                 if (!empty($jsonld['@graph']) && is_array($jsonld['@graph'])) {
 520                         foreach ($jsonld['@graph'] as $part) {
 521                                 $siteinfo = self::parseJsonLd($siteinfo, $part);
 522                         }
 523                 } elseif (!empty($jsonld['@type'])) {
 524                         $siteinfo = self::parseJsonLd($siteinfo, $jsonld);
 525                 } elseif (!empty($jsonld)) {
 526                         $keys = array_keys($jsonld);
 527                         $numeric_keys = true;
 528                         foreach ($keys as $key) {
 529                                 if (!is_int($key)) {
 530                                         $numeric_keys = false;
 531                                 }
 532                         }
 533                         if ($numeric_keys) {
 534                                 foreach ($jsonld as $part) {
 535                                         $siteinfo = self::parseParts($siteinfo, $part);
 536                                 }
 537                         }
 538                 }
 539
 540                 return $siteinfo;
 541         }
 542
 543         /**
 544          * Improve the siteinfo with information from the provided JSON-LD information
 545          * @see https://jsonld.com/
 546          *
 547          * @param array $siteinfo
 548          * @param array $jsonld
 549          * @return array siteinfo
 550          */
 551         private static function parseJsonLd(array $siteinfo, array $jsonld)
 552         {
 553                 $type = JsonLD::fetchElement($jsonld, '@type');
 554                 if (empty($type)) {
 555                         Logger::info('Empty type', ['url' => $siteinfo['url']]);
 556                         return $siteinfo;
 557                 }
 558
 559                 // Silently ignore some types that aren't processed
 560                 if (in_array($type, ['SiteNavigationElement', 'JobPosting', 'CreativeWork',
 561                         'WPHeader', 'WPSideBar', 'WPFooter', 'LegalService',
 562                         'ItemList', 'BreadcrumbList', 'Blog', 'Dataset', 'Product'])) {
 563                         return $siteinfo;
 564                 }
 565
 566                 switch ($type) {
 567                         case 'Article':
 568                         case 'AdvertiserContentArticle':
 569                         case 'NewsArticle':
 570                         case 'Report':
 571                         case 'SatiricalArticle':
 572                         case 'ScholarlyArticle':
 573                         case 'SocialMediaPosting':
 574                         case 'TechArticle':
 575                         case 'ReportageNewsArticle':
 576                         case 'SocialMediaPosting':
 577                         case 'BlogPosting':
 578                         case 'LiveBlogPosting':
 579                         case 'DiscussionForumPosting':
 580                                 return self::parseJsonLdArticle($siteinfo, $jsonld);
 581                         case 'WebPage':
 582                         case 'AboutPage':
 583                         case 'CheckoutPage':
 584                         case 'CollectionPage':
 585                         case 'ContactPage':
 586                         case 'FAQPage':
 587                         case 'ItemPage':
 588                         case 'MedicalWebPage':
 589                         case 'ProfilePage':
 590                         case 'QAPage':
 591                         case 'RealEstateListing':
 592                         case 'SearchResultsPage':
 593                         case 'MediaGallery':
 594                         case 'ImageGallery':
 595                         case 'VideoGallery':
 596                         case 'RadioEpisode':
 597                         case 'Event':
 598                                 return self::parseJsonLdWebPage($siteinfo, $jsonld);
 599                         case 'WebSite':
 600                                 return self::parseJsonLdWebSite($siteinfo, $jsonld);
 601                         case 'Organization':
 602                         case 'Airline':
 603                         case 'Consortium':
 604                         case 'Corporation':
 605                         case 'EducationalOrganization':
 606                         case 'FundingScheme':
 607                         case 'GovernmentOrganization':
 608                         case 'LibrarySystem':
 609                         case 'LocalBusiness':
 610                         case 'MedicalOrganization':
 611                         case 'NGO':
 612                         case 'NewsMediaOrganization':
 613                         case 'Project':
 614                         case 'SportsOrganization':
 615                         case 'WorkersUnion':
 616                                 return self::parseJsonLdWebOrganization($siteinfo, $jsonld);
 617                         case 'Person':
 618                         case 'Patient':
 619                                 case 'PerformingGroup':
 620                         case 'DanceGroup';
 621                         case 'MusicGroup':
 622                         case 'TheaterGroup':
 623                                 return self::parseJsonLdWebPerson($siteinfo, $jsonld);
 624                         case 'AudioObject':
 625                         case 'Audio':
 626                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'audio');
 627                         case 'VideoObject':
 628                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'video');
 629                         case 'ImageObject':
 630                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'images');
 631                         default:
 632                                 Logger::info('Unknown type', ['type' => $type, 'url' => $siteinfo['url']]);
 633                                 return $siteinfo;
 634                 }
 635         }
 636
 637         /**
 638          * Improve the siteinfo with information from the provided JSON-LD information concerning authors and publishers
 639          *
 640          * @param array $siteinfo
 641          * @param array $jsonld
 642          * @return array siteinfo
 643          */
 644         private static function parseJsonLdAuthor(array $siteinfo, array $jsonld)
 645         {
 646                 $jsonldinfo = [];
 647
 648                 if (!empty($jsonld['publisher']) && is_array($jsonld['publisher'])) {
 649                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'name');
 650                         if (!empty($content) && is_string($content)) {
 651                                 $jsonldinfo['publisher_name'] = trim($content);
 652                         }
 653
 654                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'sameAs');
 655                         if (!empty($content) && is_string($content)) {
 656                                 $jsonldinfo['publisher_url'] = trim($content);
 657                         }
 658
 659                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'url');
 660                         if (!empty($content) && is_string($content)) {
 661                                 $jsonldinfo['publisher_url'] = trim($content);
 662                         }
 663
 664                         $brand = JsonLD::fetchElement($jsonld, 'publisher', 'brand', '@type', 'Organization');
 665                         if (!empty($brand) && is_array($brand)) {
 666                                 $content = JsonLD::fetchElement($brand, 'name', '@type', 'brand');
 667                                 if (!empty($content) && is_string($content)) {
 668                                         $jsonldinfo['publisher_name'] = trim($content);
 669                                 }
 670                                 $content = JsonLD::fetchElement($brand, 'url', '@type', 'brand');
 671                                 if (!empty($content) && is_string($content)) {
 672                                         $jsonldinfo['publisher_url'] = trim($content);
 673                                 }
 674                         }
 675                 } elseif (!empty($jsonld['publisher']) && is_string($jsonld['publisher'])) {
 676                         $jsonldinfo['publisher_name'] = trim($jsonld['publisher']);
 677                 }
 678
 679                 if (!empty($jsonld['author']) && is_array($jsonld['author'])) {
 680                         $content = JsonLD::fetchElement($jsonld, 'author', 'name');
 681                         if (!empty($content) && is_string($content)) {
 682                                 $jsonldinfo['author_name'] = trim($content);
 683                         }
 684
 685                         $content = JsonLD::fetchElement($jsonld, 'author', 'sameAs');
 686                         if (!empty($content) && is_string($content)) {
 687                                 $jsonldinfo['author_url'] = trim($content);
 688                         }
 689
 690                         $content = JsonLD::fetchElement($jsonld, 'author', 'url');
 691                         if (!empty($content) && is_string($content)) {
 692                                 $jsonldinfo['author_url'] = trim($content);
 693                         }
 694                 } elseif (!empty($jsonld['author']) && is_string($jsonld['author'])) {
 695                         $jsonldinfo['author_name'] = trim($jsonld['author']);
 696                 }
 697
 698                 Logger::info('Fetched author information', ['fetched' => $jsonldinfo]);
 699
 700                 return array_merge($siteinfo, $jsonldinfo);
 701         }
 702
 703         /**
 704          * Improve the siteinfo with information from the provided JSON-LD Article information
 705          * @see https://schema.org/Article
 706          *
 707          * @param array $siteinfo
 708          * @param array $jsonld
 709          * @return array siteinfo
 710          */
 711         private static function parseJsonLdArticle(array $siteinfo, array $jsonld)
 712         {
 713                 $jsonldinfo = [];
 714
 715                 $content = JsonLD::fetchElement($jsonld, 'headline');
 716                 if (!empty($content) && is_string($content)) {
 717                         $jsonldinfo['title'] = trim($content);
 718                 }
 719
 720                 $content = JsonLD::fetchElement($jsonld, 'alternativeHeadline');
 721                 if (!empty($content) && is_string($content) && (($jsonldinfo['title'] ?? '') != trim($content))) {
 722                         $jsonldinfo['alternative_title'] = trim($content);
 723                 }
 724
 725                 $content = JsonLD::fetchElement($jsonld, 'description');
 726                 if (!empty($content) && is_string($content)) {
 727                         $jsonldinfo['text'] = trim($content);
 728                 }
 729
 730                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 731                 if (!empty($content)) {
 732                         $jsonldinfo['image'] = trim($content);
 733                 }
 734
 735                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
 736                 if (!empty($content) && is_string($content)) {
 737                         $jsonldinfo['image'] = trim($content);
 738                 }
 739
 740                 if (!empty($jsonld['keywords']) && !is_array($jsonld['keywords'])) {
 741                         $content = JsonLD::fetchElement($jsonld, 'keywords');
 742                         if (!empty($content)) {
 743                                 $siteinfo['keywords'] = [];
 744                                 $keywords = explode(',', $content);
 745                                 foreach ($keywords as $keyword) {
 746                                         $siteinfo['keywords'][] = trim($keyword);
 747                                 }
 748                         }
 749                 } else {
 750                         $content = JsonLD::fetchElementArray($jsonld, 'keywords');
 751                         if (!empty($content) && is_array($content)) {
 752                                 $jsonldinfo['keywords'] = $content;
 753                         }
 754                 }
 755
 756                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 757
 758                 Logger::info('Fetched article information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 759
 760                 return array_merge($siteinfo, $jsonldinfo);
 761         }
 762
 763         /**
 764          * Improve the siteinfo with information from the provided JSON-LD WebPage information
 765          * @see https://schema.org/WebPage
 766          *
 767          * @param array $siteinfo
 768          * @param array $jsonld
 769          * @return array siteinfo
 770          */
 771         private static function parseJsonLdWebPage(array $siteinfo, array $jsonld)
 772         {
 773                 $jsonldinfo = [];
 774
 775                 $content = JsonLD::fetchElement($jsonld, 'name');
 776                 if (!empty($content)) {
 777                         $jsonldinfo['title'] = trim($content);
 778                 }
 779
 780                 $content = JsonLD::fetchElement($jsonld, 'description');
 781                 if (!empty($content)) {
 782                         $jsonldinfo['text'] = trim($content);
 783                 }
 784
 785                 $content = JsonLD::fetchElement($jsonld, 'image');
 786                 if (!empty($content)) {
 787                         $jsonldinfo['image'] = trim($content);
 788                 }
 789
 790                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 791                 if (!empty($content)) {
 792                         $jsonldinfo['image'] = trim($content);
 793                 }
 794
 795                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 796
 797                 Logger::info('Fetched webpage information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 798
 799                 return array_merge($siteinfo, $jsonldinfo);
 800         }
 801
 802         /**
 803          * Improve the siteinfo with information from the provided JSON-LD WebSite information
 804          * @see https://schema.org/WebSite
 805          *
 806          * @param array $siteinfo
 807          * @param array $jsonld
 808          * @return array siteinfo
 809          */
 810         private static function parseJsonLdWebSite(array $siteinfo, array $jsonld)
 811         {
 812                 $jsonldinfo = [];
 813
 814                 $content = JsonLD::fetchElement($jsonld, 'name');
 815                 if (!empty($content)) {
 816                         $jsonldinfo['publisher_name'] = trim($content);
 817                 }
 818
 819                 $content = JsonLD::fetchElement($jsonld, 'description');
 820                 if (!empty($content)) {
 821                         $jsonldinfo['publisher_description'] = trim($content);
 822                 }
 823
 824                 $content = JsonLD::fetchElement($jsonld, 'url');
 825                 if (!empty($content)) {
 826                         $jsonldinfo['publisher_url'] = trim($content);
 827                 }
 828
 829                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 830                 if (!empty($content)) {
 831                         $jsonldinfo['image'] = trim($content);
 832                 }
 833
 834                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 835
 836                 Logger::info('Fetched WebSite information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 837                 return array_merge($siteinfo, $jsonldinfo);
 838         }
 839
 840         /**
 841          * Improve the siteinfo with information from the provided JSON-LD Organization information
 842          * @see https://schema.org/Organization
 843          *
 844          * @param array $siteinfo
 845          * @param array $jsonld
 846          * @return array siteinfo
 847          */
 848         private static function parseJsonLdWebOrganization(array $siteinfo, array $jsonld)
 849         {
 850                 $jsonldinfo = [];
 851
 852                 $content = JsonLD::fetchElement($jsonld, 'name');
 853                 if (!empty($content)) {
 854                         $jsonldinfo['publisher_name'] = trim($content);
 855                 }
 856
 857                 $content = JsonLD::fetchElement($jsonld, 'description');
 858                 if (!empty($content)) {
 859                         $jsonldinfo['publisher_description'] = trim($content);
 860                 }
 861
 862                 $content = JsonLD::fetchElement($jsonld, 'sameAs');
 863                 if (!empty($content) && is_string($content)) {
 864                         $jsonldinfo['publisher_url'] = trim($content);
 865                 }
 866
 867                 $content = JsonLD::fetchElement($jsonld, 'url');
 868                 if (!empty($content)) {
 869                         $jsonldinfo['publisher_url'] = trim($content);
 870                 }
 871
 872                 $content = JsonLD::fetchElement($jsonld, 'logo', 'url', '@type', 'ImageObject');
 873                 if (!empty($content)) {
 874                         $jsonldinfo['publisher_img'] = trim($content);
 875                 }
 876
 877                 $content = JsonLD::fetchElement($jsonld, 'brand', 'name', '@type', 'Organization');
 878                 if (!empty($content)) {
 879                         $jsonldinfo['publisher_name'] = trim($content);
 880                 }
 881
 882                 $content = JsonLD::fetchElement($jsonld, 'brand', 'url', '@type', 'Organization');
 883                 if (!empty($content)) {
 884                         $jsonldinfo['publisher_url'] = trim($content);
 885                 }
 886
 887                 Logger::info('Fetched Organization information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 888                 return array_merge($siteinfo, $jsonldinfo);
 889         }
 890
 891         /**
 892          * Improve the siteinfo with information from the provided JSON-LD Person information
 893          * @see https://schema.org/Person
 894          *
 895          * @param array $siteinfo
 896          * @param array $jsonld
 897          * @return array siteinfo
 898          */
 899         private static function parseJsonLdWebPerson(array $siteinfo, array $jsonld)
 900         {
 901                 $jsonldinfo = [];
 902
 903                 $content = JsonLD::fetchElement($jsonld, 'name');
 904                 if (!empty($content)) {
 905                         $jsonldinfo['author_name'] = trim($content);
 906                 }
 907
 908                 $content = JsonLD::fetchElement($jsonld, 'description');
 909                 if (!empty($content)) {
 910                         $jsonldinfo['author_description'] = trim($content);
 911                 }
 912
 913                 $content = JsonLD::fetchElement($jsonld, 'sameAs');
 914                 if (!empty($content) && is_string($content)) {
 915                         $jsonldinfo['author_url'] = trim($content);
 916                 }
 917
 918                 $content = JsonLD::fetchElement($jsonld, 'url');
 919                 if (!empty($content)) {
 920                         $jsonldinfo['author_url'] = trim($content);
 921                 }
 922
 923                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
 924                 if (!empty($content)) {
 925                         $jsonldinfo['author_img'] = trim($content);
 926                 }
 927
 928                 Logger::info('Fetched Person information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 929                 return array_merge($siteinfo, $jsonldinfo);
 930         }
 931
 932         /**
 933          * Improve the siteinfo with information from the provided JSON-LD MediaObject
 934          * @see https://schema.org/MediaObject
 935          *
 936          * @param array $siteinfo
 937          * @param array $jsonld
 938          * @return array siteinfo
 939          */
 940         private static function parseJsonLdMediaObject(array $siteinfo, array $jsonld, string $name)
 941         {
 942                 $media = [];
 943
 944                 $content = JsonLD::fetchElement($jsonld, 'caption');
 945                 if (!empty($content)) {
 946                         $media['caption'] = trim($content);
 947                 }
 948
 949                 $content = JsonLD::fetchElement($jsonld, 'url');
 950                 if (!empty($content)) {
 951                         $media['src'] = trim($content);
 952                 }
 953
 954                 $content = JsonLD::fetchElement($jsonld, 'description');
 955                 if (!empty($content)) {
 956                         $media['description'] = trim($content);
 957                 }
 958
 959                 $content = JsonLD::fetchElement($jsonld, 'name');
 960                 if (!empty($content) && (($media['description'] ?? '') != trim($content))) {
 961                         $media['name'] = trim($content);
 962                 }
 963
 964                 $content = JsonLD::fetchElement($jsonld, 'contentUrl');
 965                 if (!empty($content)) {
 966                         $media['content'] = trim($content);
 967                 }
 968
 969                 $content = JsonLD::fetchElement($jsonld, 'embedUrl');
 970                 if (!empty($content)) {
 971                         $media['embed'] = trim($content);
 972                 }
 973
 974                 $content = JsonLD::fetchElement($jsonld, 'height');
 975                 if (!empty($content)) {
 976                         $media['height'] = trim($content);
 977                 }
 978
 979                 $content = JsonLD::fetchElement($jsonld, 'width');
 980                 if (!empty($content)) {
 981                         $media['width'] = trim($content);
 982                 }
 983
 984                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 985                 if (!empty($content)) {
 986                         $media['preview'] = trim($content);
 987                 }
 988
 989                 $content = JsonLD::fetchElement($jsonld, 'image');
 990                 if (!empty($content)) {
 991                         $media['image'] = trim($content);
 992                 }
 993
 994                 Logger::info('Fetched Media information', ['url' => $siteinfo['url'], 'fetched' => $media]);
 995                 $siteinfo[$name][] = $media;
 996                 return $siteinfo;
 997         }
 998
 999         /**
1000          * Convert tags from CSV to an array
1001          *
1002          * @param string $string Tags
1003          * @return array with formatted Hashtags
1004          */
1005         public static function convertTagsToArray($string)
1006         {
1007                 $arr_tags = str_getcsv($string);
1008                 if (count($arr_tags)) {
1009                         // add the # sign to every tag
1010                         array_walk($arr_tags, ["self", "arrAddHashes"]);
1011
1012                         return $arr_tags;
1013                 }
1014         }
1015
1016         /**
1017          * Add a hasht sign to a string
1018          *
1019          * This method is used as callback function
1020          *
1021          * @param string $tag The pure tag name
1022          * @param int    $k   Counter for internal use
1023          * @return void
1024          */
1025         private static function arrAddHashes(&$tag, $k)
1026         {
1027                 $tag = "#" . $tag;
1028         }
1029
1030         /**
1031          * Add a scheme to an url
1032          *
1033          * The src attribute of some html elements (e.g. images)
1034          * can miss the scheme so we need to add the correct
1035          * scheme
1036          *
1037          * @param string $url    The url which possibly does have
1038          *                       a missing scheme (a link to an image)
1039          * @param string $scheme The url with a correct scheme
1040          *                       (e.g. the url from the webpage which does contain the image)
1041          *
1042          * @return string The url with a scheme
1043          */
1044         private static function completeUrl($url, $scheme)
1045         {
1046                 $urlarr = parse_url($url);
1047
1048                 // If the url does allready have an scheme
1049                 // we can stop the process here
1050                 if (isset($urlarr["scheme"])) {
1051                         return($url);
1052                 }
1053
1054                 $schemearr = parse_url($scheme);
1055
1056                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
1057
1058                 if (!empty($schemearr["port"])) {
1059                         $complete .= ":".$schemearr["port"];
1060                 }
1061
1062                 if (!empty($urlarr["path"])) {
1063                         if (strpos($urlarr["path"], "/") !== 0) {
1064                                 $complete .= "/";
1065                         }
1066
1067                         $complete .= $urlarr["path"];
1068                 }
1069
1070                 if (!empty($urlarr["query"])) {
1071                         $complete .= "?".$urlarr["query"];
1072                 }
1073
1074                 if (!empty($urlarr["fragment"])) {
1075                         $complete .= "#".$urlarr["fragment"];
1076                 }
1077
1078                 return($complete);
1079         }
1080 }