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                                         if (!empty($jsonld['@graph']) && is_array($jsonld['@graph'])) {
 472                                                 foreach ($jsonld['@graph'] as $part) {
 473                                                         $siteinfo = self::parseJsonLd($siteinfo, $part);
 474                                                 }
 475                                         } else {
 476                                                 $siteinfo = self::parseJsonLd($siteinfo, $jsonld);
 477                                         }
 478                                 }
 479                         }
 480                 }
 481
 482                 // Prevent to have a photo type without an image
 483                 if ((empty($siteinfo['image']) || !empty($siteinfo['text'])) && ($siteinfo['type'] == 'photo')) {
 484                         $siteinfo['type'] = 'link';
 485                 }
 486
 487                 if (!empty($siteinfo['image'])) {
 488                         $src = self::completeUrl($siteinfo['image'], $url);
 489
 490                         unset($siteinfo['image']);
 491
 492                         $photodata = Images::getInfoFromURLCached($src);
 493
 494                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 495                                 $siteinfo['images'][] = ['src' => $src,
 496                                         'width' => $photodata[0],
 497                                         'height' => $photodata[1]];
 498                         }
 499                 }
 500
 501                 if (!empty($siteinfo['text']) && mb_strlen($siteinfo['text']) > self::MAX_DESC_COUNT) {
 502                         $siteinfo['text'] = mb_substr($siteinfo['text'], 0, self::MAX_DESC_COUNT) . '…';
 503                         $pos = mb_strrpos($siteinfo['text'], '.');
 504                         if ($pos > self::MIN_DESC_COUNT) {
 505                                 $siteinfo['text'] = mb_substr($siteinfo['text'], 0, $pos + 1);
 506                         }
 507                 }
 508
 509                 Logger::info('Siteinfo fetched', ['url' => $url, 'siteinfo' => $siteinfo]);
 510
 511                 Hook::callAll('getsiteinfo', $siteinfo);
 512
 513                 return $siteinfo;
 514         }
 515
 516         /**
 517          * Improve the siteinfo with information from the provided JSON-LD information
 518          * @see https://jsonld.com/
 519          *
 520          * @param array $siteinfo
 521          * @param array $jsonld
 522          * @return array siteinfo
 523          */
 524         private static function parseJsonLd(array $siteinfo, array $jsonld)
 525         {
 526                 $type = JsonLD::fetchElement($jsonld, '@type');
 527
 528                 switch ($type) {
 529                         case 'Article':
 530                         case 'NewsArticle':
 531                                 return self::parseJsonLdArticle($siteinfo, $jsonld);
 532                         case 'WebPage':
 533                                 return self::parseJsonLdWebPage($siteinfo, $jsonld);
 534                         case 'WebSite':
 535                                 return self::parseJsonLdWebSite($siteinfo, $jsonld);
 536                         case 'Organization':
 537                                 return self::parseJsonLdWebOrganization($siteinfo, $jsonld);
 538                         case 'Person':
 539                                 return self::parseJsonLdWebPerson($siteinfo, $jsonld);
 540                         case 'BreadcrumbList':
 541                         case 'Audio': /// @todo Can contain direct media links to audio - can be interesting in the future
 542                         case 'VideoObject':
 543                         case 'ImageObject':
 544                         case 'LiveBlogPosting':
 545                         case 'SocialMediaPosting':
 546                                         // quit silently
 547                                 return $siteinfo;
 548                         default:
 549                                 Logger::info('Unsupported or unknown type', ['type' => $type, 'url' => $siteinfo['url']]);
 550                                 return $siteinfo;
 551                 }
 552         }
 553
 554         /**
 555          * Improve the siteinfo with information from the provided JSON-LD information concerning authors and publishers
 556          *
 557          * @param array $siteinfo
 558          * @param array $jsonld
 559          * @return array siteinfo
 560          */
 561         private static function parseJsonLdAuthor(array $siteinfo, array $jsonld)
 562         {
 563                 $jsonldinfo = [];
 564
 565                 if (!empty($jsonld['publisher']) && is_array($jsonld['publisher'])) {
 566                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'name', '@type', 'Person');
 567                         if (!empty($content) && is_string($content)) {
 568                                 $jsonldinfo['publisher_name'] = trim($content);
 569                         }
 570
 571                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'url', '@type', 'Person');
 572                         if (!empty($content) && is_string($content)) {
 573                                 $jsonldinfo['publisher_url'] = trim($content);
 574                         }
 575
 576                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'name', '@type', 'Organization');
 577                         if (!empty($content) && is_string($content)) {
 578                                 $jsonldinfo['publisher_name'] = trim($content);
 579                         }
 580
 581                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'url', '@type', 'Organization');
 582                         if (!empty($content) && is_string($content)) {
 583                                 $jsonldinfo['publisher_url'] = trim($content);
 584                         }
 585
 586                         $brand = JsonLD::fetchElement($jsonld, 'publisher', 'brand', '@type', 'Organization');
 587                         if (!empty($brand) && is_array($brand)) {
 588                                 $content = JsonLD::fetchElement($brand, 'name', '@type', 'brand');
 589                                 if (!empty($content) && is_string($content)) {
 590                                         $jsonldinfo['publisher_name'] = trim($content);
 591                                 }
 592                         }
 593                 }
 594
 595                 if (!empty($jsonld['author']) && is_array($jsonld['author'])) {
 596                         $content = JsonLD::fetchElement($jsonld, 'author', 'name', '@type', 'Organization');
 597                         if (!empty($content) && is_string($content)) {
 598                                 $jsonldinfo['author_name'] = trim($content);
 599                         }
 600
 601                         $content = JsonLD::fetchElement($jsonld, 'author', 'url', '@type', 'Organization');
 602                         if (!empty($content) && is_string($content)) {
 603                                 $jsonldinfo['author_url'] = trim($content);
 604                         }
 605
 606                         $content = JsonLD::fetchElement($jsonld, 'author', 'name', '@type', 'Person');
 607                         if (!empty($content) && is_string($content)) {
 608                                 $jsonldinfo['author_name'] = trim($content);
 609                         }
 610
 611                         $content = JsonLD::fetchElement($jsonld, 'author', 'url', '@type', 'Person');
 612                         if (!empty($content) && is_string($content)) {
 613                                 $jsonldinfo['author_url'] = trim($content);
 614                         }
 615                 }
 616
 617                 Logger::info('Fetched author information', ['fetched' => $jsonldinfo]);
 618
 619                 return array_merge($siteinfo, $jsonldinfo);
 620         }
 621
 622         /**
 623          * Improve the siteinfo with information from the provided JSON-LD Article information
 624          *
 625          * @param array $siteinfo
 626          * @param array $jsonld
 627          * @return array siteinfo
 628          */
 629         private static function parseJsonLdArticle(array $siteinfo, array $jsonld)
 630         {
 631                 $jsonldinfo = [];
 632
 633                 $content = JsonLD::fetchElement($jsonld, 'headline');
 634                 if (!empty($content) && is_string($content)) {
 635                         $jsonldinfo['title'] = trim($content);
 636                 }
 637
 638                 $content = JsonLD::fetchElement($jsonld, 'alternativeHeadline');
 639                 if (!empty($content) && is_string($content)) {
 640                         $jsonldinfo['alternative_title'] = trim($content);
 641                 }
 642
 643                 $content = JsonLD::fetchElement($jsonld, 'description');
 644                 if (!empty($content) && is_string($content)) {
 645                         $jsonldinfo['text'] = trim($content);
 646                 }
 647
 648                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
 649                 if (!empty($content) && is_string($content)) {
 650                         $jsonldinfo['image'] = trim($content);
 651                 }
 652
 653 /// @todo Check for the correct separator, also check for dpublicates before adding
 654 //              $content = JsonLD::fetchElement($jsonld, 'keywords');
 655 //              if (!empty($content) && is_string($content)) {
 656 //                      $jsonldinfo['keywords'] = trim($content);
 657 //              }
 658
 659                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 660
 661                 Logger::info('Fetched article information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 662
 663                 return array_merge($siteinfo, $jsonldinfo);
 664         }
 665
 666         /**
 667          * Improve the siteinfo with information from the provided JSON-LD WebPage information
 668          *
 669          * @param array $siteinfo
 670          * @param array $jsonld
 671          * @return array siteinfo
 672          */
 673         private static function parseJsonLdWebPage(array $siteinfo, array $jsonld)
 674         {
 675                 $jsonldinfo = [];
 676
 677                 $content = JsonLD::fetchElement($jsonld, 'name');
 678                 if (!empty($content)) {
 679                         $jsonldinfo['title'] = trim($content);
 680                 }
 681
 682                 $content = JsonLD::fetchElement($jsonld, 'description');
 683                 if (!empty($content)) {
 684                         $jsonldinfo['text'] = trim($content);
 685                 }
 686
 687                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 688                 if (!empty($content)) {
 689                         $jsonldinfo['image'] = trim($content);
 690                 }
 691
 692                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 693
 694                 Logger::info('Fetched webpage information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 695
 696                 return array_merge($siteinfo, $jsonldinfo);
 697         }
 698
 699         /**
 700          * Improve the siteinfo with information from the provided JSON-LD WebSite information
 701          *
 702          * @param array $siteinfo
 703          * @param array $jsonld
 704          * @return array siteinfo
 705          */
 706         private static function parseJsonLdWebSite(array $siteinfo, array $jsonld)
 707         {
 708                 $jsonldinfo = [];
 709
 710                 $content = JsonLD::fetchElement($jsonld, 'name');
 711                 if (!empty($content)) {
 712                         $jsonldinfo['publisher_name'] = trim($content);
 713                 }
 714
 715                 $content = JsonLD::fetchElement($jsonld, 'description');
 716                 if (!empty($content)) {
 717                         $jsonldinfo['publisher_description'] = trim($content);
 718                 }
 719
 720                 $content = JsonLD::fetchElement($jsonld, 'url');
 721                 if (!empty($content)) {
 722                         $jsonldinfo['publisher_url'] = trim($content);
 723                 }
 724
 725                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 726                 if (!empty($content)) {
 727                         $jsonldinfo['image'] = trim($content);
 728                 }
 729
 730                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 731
 732                 Logger::info('Fetched WebSite information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 733                 return array_merge($siteinfo, $jsonldinfo);
 734         }
 735
 736         /**
 737          * Improve the siteinfo with information from the provided JSON-LD Organization information
 738          *
 739          * @param array $siteinfo
 740          * @param array $jsonld
 741          * @return array siteinfo
 742          */
 743         private static function parseJsonLdWebOrganization(array $siteinfo, array $jsonld)
 744         {
 745                 $jsonldinfo = [];
 746
 747                 $content = JsonLD::fetchElement($jsonld, 'name');
 748                 if (!empty($content)) {
 749                         $jsonldinfo['publisher_name'] = trim($content);
 750                 }
 751
 752                 $content = JsonLD::fetchElement($jsonld, 'url');
 753                 if (!empty($content)) {
 754                         $jsonldinfo['publisher_url'] = trim($content);
 755                 }
 756
 757                 $content = JsonLD::fetchElement($jsonld, 'logo', 'url', '@type', 'ImageObject');
 758                 if (!empty($content)) {
 759                         $jsonldinfo['publisher_img'] = trim($content);
 760                 }
 761
 762                 Logger::info('Fetched Organization information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 763                 return array_merge($siteinfo, $jsonldinfo);
 764         }
 765
 766         /**
 767          * Improve the siteinfo with information from the provided JSON-LD Person information
 768          *
 769          * @param array $siteinfo
 770          * @param array $jsonld
 771          * @return array siteinfo
 772          */
 773         private static function parseJsonLdWebPerson(array $siteinfo, array $jsonld)
 774         {
 775                 $jsonldinfo = [];
 776
 777                 $content = JsonLD::fetchElement($jsonld, 'name');
 778                 if (!empty($content)) {
 779                         $jsonldinfo['author_name'] = trim($content);
 780                 }
 781
 782                 $content = JsonLD::fetchElement($jsonld, 'description');
 783                 if (!empty($content)) {
 784                         $jsonldinfo['author_description'] = trim($content);
 785                 }
 786
 787                 $content = JsonLD::fetchElement($jsonld, 'url');
 788                 if (!empty($content)) {
 789                         $jsonldinfo['author_url'] = trim($content);
 790                 }
 791
 792                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
 793                 if (!empty($content)) {
 794                         $jsonldinfo['author_img'] = trim($content);
 795                 }
 796
 797                 Logger::info('Fetched Person information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 798                 return array_merge($siteinfo, $jsonldinfo);
 799         }
 800
 801         /**
 802          * Convert tags from CSV to an array
 803          *
 804          * @param string $string Tags
 805          * @return array with formatted Hashtags
 806          */
 807         public static function convertTagsToArray($string)
 808         {
 809                 $arr_tags = str_getcsv($string);
 810                 if (count($arr_tags)) {
 811                         // add the # sign to every tag
 812                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 813
 814                         return $arr_tags;
 815                 }
 816         }
 817
 818         /**
 819          * Add a hasht sign to a string
 820          *
 821          * This method is used as callback function
 822          *
 823          * @param string $tag The pure tag name
 824          * @param int    $k   Counter for internal use
 825          * @return void
 826          */
 827         private static function arrAddHashes(&$tag, $k)
 828         {
 829                 $tag = "#" . $tag;
 830         }
 831
 832         /**
 833          * Add a scheme to an url
 834          *
 835          * The src attribute of some html elements (e.g. images)
 836          * can miss the scheme so we need to add the correct
 837          * scheme
 838          *
 839          * @param string $url    The url which possibly does have
 840          *                       a missing scheme (a link to an image)
 841          * @param string $scheme The url with a correct scheme
 842          *                       (e.g. the url from the webpage which does contain the image)
 843          *
 844          * @return string The url with a scheme
 845          */
 846         private static function completeUrl($url, $scheme)
 847         {
 848                 $urlarr = parse_url($url);
 849
 850                 // If the url does allready have an scheme
 851                 // we can stop the process here
 852                 if (isset($urlarr["scheme"])) {
 853                         return($url);
 854                 }
 855
 856                 $schemearr = parse_url($scheme);
 857
 858                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 859
 860                 if (!empty($schemearr["port"])) {
 861                         $complete .= ":".$schemearr["port"];
 862                 }
 863
 864                 if (!empty($urlarr["path"])) {
 865                         if (strpos($urlarr["path"], "/") !== 0) {
 866                                 $complete .= "/";
 867                         }
 868
 869                         $complete .= $urlarr["path"];
 870                 }
 871
 872                 if (!empty($urlarr["query"])) {
 873                         $complete .= "?".$urlarr["query"];
 874                 }
 875
 876                 if (!empty($urlarr["fragment"])) {
 877                         $complete .= "#".$urlarr["fragment"];
 878                 }
 879
 880                 return($complete);
 881         }
 882 }