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   $no_guessing If true the parse doens't search for
  79          *                            preview pictures
  80          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
  81          *                            to avoid endless loops
  82          *
  83          * @return array which contains needed data for embedding
  84          *    string 'url'      => The url of the parsed page
  85          *    string 'type'     => Content type
  86          *    string 'title'    => (optional) The title of the content
  87          *    string 'text'     => (optional) The description for the content
  88          *    string 'image'    => (optional) A preview image of the content (only available if $no_geuessing = false)
  89          *    array  'images'   => (optional) Array of preview pictures
  90          *    string 'keywords' => (optional) The tags which belong to the content
  91          *
  92          * @throws HTTPException\InternalServerErrorException
  93          * @see   ParseUrl::getSiteinfo() for more information about scraping
  94          * embeddable content
  95          */
  96         public static function getSiteinfoCached($url, $no_guessing = false, $do_oembed = true): array
  97         {
  98                 if (empty($url)) {
  99                         return [
 100                                 'url' => '',
 101                                 'type' => 'error',
 102                         ];
 103                 }
 104
 105                 $urlHash = hash('sha256', $url);
 106
 107                 $parsed_url = DBA::selectFirst('parsed_url', ['content'],
 108                         ['url_hash' => $urlHash, 'guessing' => !$no_guessing, 'oembed' => $do_oembed]
 109                 );
 110                 if (!empty($parsed_url['content'])) {
 111                         $data = unserialize($parsed_url['content']);
 112                         return $data;
 113                 }
 114
 115                 $data = self::getSiteinfo($url, $no_guessing, $do_oembed);
 116
 117                 $expires = $data['expires'];
 118
 119                 unset($data['expires']);
 120
 121                 DI::dba()->insert(
 122                         'parsed_url',
 123                         [
 124                                 'url_hash' => $urlHash,
 125                                 'guessing' => !$no_guessing,
 126                                 'oembed'   => $do_oembed,
 127                                 'url'      => $url,
 128                                 'content'  => serialize($data),
 129                                 'created'  => DateTimeFormat::utcNow(),
 130                                 'expires'  => $expires,
 131                         ],
 132                         Database::INSERT_UPDATE
 133                 );
 134
 135                 return $data;
 136         }
 137
 138         /**
 139          * Parse a page for embeddable content information
 140          *
 141          * This method parses to url for meta data which can be used to embed
 142          * the content. If available it prioritizes Open Graph meta tags.
 143          * If this is not available it uses the twitter cards meta tags.
 144          * As fallback it uses standard html elements with meta informations
 145          * like \<title\>Awesome Title\</title\> or
 146          * \<meta name="description" content="An awesome description"\>
 147          *
 148          * @param string $url         The url of the page which should be scraped
 149          * @param bool   $no_guessing If true the parse doens't search for
 150          *                            preview pictures
 151          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
 152          *                            to avoid endless loops
 153          * @param int    $count       Internal counter to avoid endless loops
 154          *
 155          * @return array which contains needed data for embedding
 156          *    string 'url'      => The url of the parsed page
 157          *    string 'type'     => Content type (error, link, photo, image, audio, video)
 158          *    string 'title'    => (optional) The title of the content
 159          *    string 'text'     => (optional) The description for the content
 160          *    string 'image'    => (optional) A preview image of the content (only available if $no_guessing = false)
 161          *    array  'images'   => (optional) Array of preview pictures
 162          *    string 'keywords' => (optional) The tags which belong to the content
 163          *
 164          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 165          * @todo  https://developers.google.com/+/plugins/snippet/
 166          * @verbatim
 167          * <meta itemprop="name" content="Awesome title">
 168          * <meta itemprop="description" content="An awesome description">
 169          * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
 170          *
 171          * <body itemscope itemtype="http://schema.org/Product">
 172          *   <h1 itemprop="name">Shiny Trinket</h1>
 173          *   <img itemprop="image" src="{image-url}" />
 174          *   <p itemprop="description">Shiny trinkets are shiny.</p>
 175          * </body>
 176          * @endverbatim
 177          */
 178         public static function getSiteinfo($url, $no_guessing = false, $do_oembed = true, $count = 1)
 179         {
 180                 if (empty($url)) {
 181                         return [
 182                                 'url' => '',
 183                                 'type' => 'error',
 184                         ];
 185                 }
 186
 187                 // Check if the URL does contain a scheme
 188                 $scheme = parse_url($url, PHP_URL_SCHEME);
 189
 190                 if ($scheme == '') {
 191                         $url = 'http://' . ltrim($url, '/');
 192                 }
 193
 194                 $url = trim($url, "'\"");
 195
 196                 $url = Network::stripTrackingQueryParams($url);
 197
 198                 $siteinfo = [
 199                         'url' => $url,
 200                         'type' => 'link',
 201                         'expires' => DateTimeFormat::utc(self::DEFAULT_EXPIRATION_FAILURE),
 202                 ];
 203
 204                 if ($count > 10) {
 205                         Logger::log('Endless loop detected for ' . $url, Logger::DEBUG);
 206                         return $siteinfo;
 207                 }
 208
 209                 $type = self::getContentType($url);
 210                 Logger::info('Got content-type', ['content-type' => $type, 'url' => $url]);
 211                 if (!empty($type) && in_array($type[0], ['image', 'video', 'audio'])) {
 212                         $siteinfo['type'] = $type[0];
 213                         return $siteinfo;
 214                 }
 215
 216                 if ((count($type) >= 2) && (($type[0] != 'text') || ($type[1] != 'html'))) {
 217                         Logger::info('Unparseable content-type, quitting here, ', ['content-type' => $type, 'url' => $url]);
 218                         return $siteinfo;
 219                 }
 220
 221                 $curlResult = DI::httpRequest()->get($url);
 222                 if (!$curlResult->isSuccess()) {
 223                         return $siteinfo;
 224                 }
 225
 226                 $siteinfo['expires'] = DateTimeFormat::utc(self::DEFAULT_EXPIRATION_SUCCESS);
 227
 228                 // If the file is too large then exit
 229                 if (($curlResult->getInfo()['download_content_length'] ?? 0) > 1000000) {
 230                         return $siteinfo;
 231                 }
 232
 233                 if ($cacheControlHeader = $curlResult->getHeader('Cache-Control')) {
 234                         if (preg_match('/max-age=([0-9]+)/i', $cacheControlHeader, $matches)) {
 235                                 $maxAge = max(86400, (int)array_pop($matches));
 236                                 $siteinfo['expires'] = DateTimeFormat::utc("now + $maxAge seconds");
 237                         }
 238                 }
 239
 240                 $header = $curlResult->getHeader();
 241                 $body = $curlResult->getBody();
 242
 243                 if ($do_oembed) {
 244                         $oembed_data = OEmbed::fetchURL($url, false, false);
 245
 246                         if (!empty($oembed_data->type)) {
 247                                 if (!in_array($oembed_data->type, ['error', 'rich', 'image', 'video', 'audio', ''])) {
 248                                         $siteinfo['type'] = $oembed_data->type;
 249                                 }
 250
 251                                 // See https://github.com/friendica/friendica/pull/5763#discussion_r217913178
 252                                 if ($siteinfo['type'] != 'photo') {
 253                                         if (!empty($oembed_data->title)) {
 254                                                 $siteinfo['title'] = trim($oembed_data->title);
 255                                         }
 256                                         if (!empty($oembed_data->description)) {
 257                                                 $siteinfo['text'] = trim($oembed_data->description);
 258                                         }
 259                                         if (!empty($oembed_data->author_name)) {
 260                                                 $siteinfo['author_name'] = trim($oembed_data->author_name);
 261                                         }
 262                                         if (!empty($oembed_data->author_url)) {
 263                                                 $siteinfo['author_url'] = trim($oembed_data->author_url);
 264                                         }
 265                                         if (!empty($oembed_data->provider_name)) {
 266                                                 $siteinfo['publisher_name'] = trim($oembed_data->provider_name);
 267                                         }
 268                                         if (!empty($oembed_data->provider_url)) {
 269                                                 $siteinfo['publisher_url'] = trim($oembed_data->provider_url);
 270                                         }
 271                                         if (!empty($oembed_data->thumbnail_url)) {
 272                                                 $siteinfo['image'] = $oembed_data->thumbnail_url;
 273                                         }
 274                                 }
 275                         }
 276                 }
 277
 278                 $charset = '';
 279                 // Look for a charset, first in headers
 280                 // Expected form: Content-Type: text/html; charset=ISO-8859-4
 281                 if (preg_match('/charset=([a-z0-9-_.\/]+)/i', $header, $matches)) {
 282                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 283                 }
 284
 285                 // Then in body that gets precedence
 286                 // Expected forms:
 287                 // - <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
 288                 // - <meta charset="utf-8">
 289                 // - <meta charset=utf-8>
 290                 // - <meta charSet="utf-8">
 291                 // We escape <style> and <script> tags since they can contain irrelevant charset information
 292                 // (see https://github.com/friendica/friendica/issues/9251#issuecomment-698636806)
 293                 Strings::performWithEscapedBlocks($body, '#<(?:style|script).*?</(?:style|script)>#ism', function ($body) use (&$charset) {
 294                         if (preg_match('/charset=["\']?([a-z0-9-_.\/]+)/i', $body, $matches)) {
 295                                 $charset = trim(trim(trim(array_pop($matches)), ';,'));
 296                         }
 297                 });
 298
 299                 $siteinfo['charset'] = $charset;
 300
 301                 if ($charset && strtoupper($charset) != 'UTF-8') {
 302                         // See https://github.com/friendica/friendica/issues/5470#issuecomment-418351211
 303                         $charset = str_ireplace('latin-1', 'latin1', $charset);
 304
 305                         Logger::log('detected charset ' . $charset, Logger::DEBUG);
 306                         $body = iconv($charset, 'UTF-8//TRANSLIT', $body);
 307                 }
 308
 309                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', 'UTF-8');
 310
 311                 $doc = new DOMDocument();
 312                 @$doc->loadHTML($body);
 313
 314                 XML::deleteNode($doc, 'style');
 315                 XML::deleteNode($doc, 'option');
 316                 XML::deleteNode($doc, 'h1');
 317                 XML::deleteNode($doc, 'h2');
 318                 XML::deleteNode($doc, 'h3');
 319                 XML::deleteNode($doc, 'h4');
 320                 XML::deleteNode($doc, 'h5');
 321                 XML::deleteNode($doc, 'h6');
 322                 XML::deleteNode($doc, 'ol');
 323                 XML::deleteNode($doc, 'ul');
 324
 325                 $xpath = new DOMXPath($doc);
 326
 327                 $list = $xpath->query('//meta[@content]');
 328                 foreach ($list as $node) {
 329                         $meta_tag = [];
 330                         if ($node->attributes->length) {
 331                                 foreach ($node->attributes as $attribute) {
 332                                         $meta_tag[$attribute->name] = $attribute->value;
 333                                 }
 334                         }
 335
 336                         if (@$meta_tag['http-equiv'] == 'refresh') {
 337                                 $path = $meta_tag['content'];
 338                                 $pathinfo = explode(';', $path);
 339                                 $content = '';
 340                                 foreach ($pathinfo as $value) {
 341                                         if (substr(strtolower($value), 0, 4) == 'url=') {
 342                                                 $content = substr($value, 4);
 343                                         }
 344                                 }
 345                                 if ($content != '') {
 346                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 347                                         return $siteinfo;
 348                                 }
 349                         }
 350                 }
 351
 352                 $list = $xpath->query('//title');
 353                 if ($list->length > 0) {
 354                         $siteinfo['title'] = trim($list->item(0)->nodeValue);
 355                 }
 356
 357                 $list = $xpath->query('//meta[@name]');
 358                 foreach ($list as $node) {
 359                         $meta_tag = [];
 360                         if ($node->attributes->length) {
 361                                 foreach ($node->attributes as $attribute) {
 362                                         $meta_tag[$attribute->name] = $attribute->value;
 363                                 }
 364                         }
 365
 366                         if (empty($meta_tag['content'])) {
 367                                 continue;
 368                         }
 369
 370                         $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 371
 372                         switch (strtolower($meta_tag['name'])) {
 373                                 case 'fulltitle':
 374                                         $siteinfo['title'] = trim($meta_tag['content']);
 375                                         break;
 376                                 case 'description':
 377                                         $siteinfo['text'] = trim($meta_tag['content']);
 378                                         break;
 379                                 case 'thumbnail':
 380                                         $siteinfo['image'] = $meta_tag['content'];
 381                                         break;
 382                                 case 'twitter:image':
 383                                         $siteinfo['image'] = $meta_tag['content'];
 384                                         break;
 385                                 case 'twitter:image:src':
 386                                         $siteinfo['image'] = $meta_tag['content'];
 387                                         break;
 388                                 case 'twitter:card':
 389                                         // Detect photo pages
 390                                         if ($meta_tag['content'] == 'summary_large_image') {
 391                                                 $siteinfo['type'] = 'photo';
 392                                         }
 393                                         break;
 394                                 case 'twitter:description':
 395                                         $siteinfo['text'] = trim($meta_tag['content']);
 396                                         break;
 397                                 case 'twitter:title':
 398                                         $siteinfo['title'] = trim($meta_tag['content']);
 399                                         break;
 400                                 case 'dc.title':
 401                                         $siteinfo['title'] = trim($meta_tag['content']);
 402                                         break;
 403                                 case 'dc.description':
 404                                         $siteinfo['text'] = trim($meta_tag['content']);
 405                                         break;
 406                                 case 'dc.creator':
 407                                         $siteinfo['publisher_name'] = trim($meta_tag['content']);
 408                                         break;
 409                                 case 'keywords':
 410                                         $keywords = explode(',', $meta_tag['content']);
 411                                         break;
 412                                 case 'news_keywords':
 413                                         $keywords = explode(',', $meta_tag['content']);
 414                                         break;
 415                         }
 416                 }
 417
 418                 if (isset($keywords)) {
 419                         $siteinfo['keywords'] = [];
 420                         foreach ($keywords as $keyword) {
 421                                 if (!in_array(trim($keyword), $siteinfo['keywords'])) {
 422                                         $siteinfo['keywords'][] = trim($keyword);
 423                                 }
 424                         }
 425                 }
 426
 427                 $list = $xpath->query('//meta[@property]');
 428                 foreach ($list as $node) {
 429                         $meta_tag = [];
 430                         if ($node->attributes->length) {
 431                                 foreach ($node->attributes as $attribute) {
 432                                         $meta_tag[$attribute->name] = $attribute->value;
 433                                 }
 434                         }
 435
 436                         if (!empty($meta_tag['content'])) {
 437                                 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 438
 439                                 switch (strtolower($meta_tag['property'])) {
 440                                         case 'og:image':
 441                                                 $siteinfo['image'] = $meta_tag['content'];
 442                                                 break;
 443                                         case 'og:image:url':
 444                                                 $siteinfo['image'] = $meta_tag['content'];
 445                                                 break;
 446                                         case 'og:image:secure_url':
 447                                                 $siteinfo['image'] = $meta_tag['content'];
 448                                                 break;
 449                                         case 'og:title':
 450                                                 $siteinfo['title'] = trim($meta_tag['content']);
 451                                                 break;
 452                                         case 'og:description':
 453                                                 $siteinfo['text'] = trim($meta_tag['content']);
 454                                                 break;
 455                                         case 'og:site_name':
 456                                                 $siteinfo['publisher_name'] = trim($meta_tag['content']);
 457                                                 break;
 458                                         case 'twitter:description':
 459                                                 $siteinfo['text'] = trim($meta_tag['content']);
 460                                                 break;
 461                                         case 'twitter:title':
 462                                                 $siteinfo['title'] = trim($meta_tag['content']);
 463                                                 break;
 464                                         case 'twitter:image':
 465                                                 $siteinfo['image'] = $meta_tag['content'];
 466                                                 break;
 467                                 }
 468                         }
 469                 }
 470
 471                 $list = $xpath->query("//script[@type='application/ld+json']");
 472                 foreach ($list as $node) {
 473                         if (!empty($node->nodeValue)) {
 474                                 $nodevalue = html_entity_decode($node->nodeValue, ENT_COMPAT, 'UTF-8');
 475                                 if ($jsonld = json_decode($nodevalue, true)) {
 476                                         if (!empty($jsonld['@graph']) && is_array($jsonld['@graph'])) {
 477                                                 foreach ($jsonld['@graph'] as $part) {
 478                                                         $siteinfo = self::parseJsonLd($siteinfo, $part);
 479                                                 }
 480                                         } else {
 481                                                 $siteinfo = self::parseJsonLd($siteinfo, $jsonld);
 482                                         }
 483                                 }
 484                         }
 485                 }
 486
 487                 // Prevent to have a photo type without an image
 488                 if ((empty($siteinfo['image']) || !empty($siteinfo['text'])) && ($siteinfo['type'] == 'photo')) {
 489                         $siteinfo['type'] = 'link';
 490                 }
 491
 492                 if (!empty($siteinfo['image'])) {
 493                         $src = self::completeUrl($siteinfo['image'], $url);
 494
 495                         unset($siteinfo['image']);
 496
 497                         $photodata = Images::getInfoFromURLCached($src);
 498
 499                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 500                                 $siteinfo['images'][] = ['src' => $src,
 501                                         'width' => $photodata[0],
 502                                         'height' => $photodata[1]];
 503                         }
 504                 }
 505
 506                 if (!empty($siteinfo['text']) && mb_strlen($siteinfo['text']) > self::MAX_DESC_COUNT) {
 507                         $siteinfo['text'] = mb_substr($siteinfo['text'], 0, self::MAX_DESC_COUNT) . '…';
 508                         $pos = mb_strrpos($siteinfo['text'], '.');
 509                         if ($pos > self::MIN_DESC_COUNT) {
 510                                 $siteinfo['text'] = mb_substr($siteinfo['text'], 0, $pos + 1);
 511                         }
 512                 }
 513
 514                 Logger::info('Siteinfo fetched', ['url' => $url, 'siteinfo' => $siteinfo]);
 515
 516                 Hook::callAll('getsiteinfo', $siteinfo);
 517
 518                 return $siteinfo;
 519         }
 520
 521         /**
 522          * Improve the siteinfo with information from the provided JSON-LD information
 523          * @see https://jsonld.com/
 524          *
 525          * @param array $siteinfo
 526          * @param array $jsonld
 527          * @return array siteinfo
 528          */
 529         private static function parseJsonLd(array $siteinfo, array $jsonld)
 530         {
 531                 $type = JsonLD::fetchElement($jsonld, '@type');
 532
 533                 switch ($type) {
 534                         case 'Article':
 535                         case 'NewsArticle':
 536                                 return self::parseJsonLdArticle($siteinfo, $jsonld);
 537                         case 'WebPage':
 538                                 return self::parseJsonLdWebPage($siteinfo, $jsonld);
 539                         case 'WebSite':
 540                                 return self::parseJsonLdWebSite($siteinfo, $jsonld);
 541                         case 'Organization':
 542                                 return self::parseJsonLdWebOrganization($siteinfo, $jsonld);
 543                         case 'Person':
 544                                 return self::parseJsonLdWebPerson($siteinfo, $jsonld);
 545                         case 'BreadcrumbList':
 546                         case 'Audio': /// @todo Can contain direct media links to audio - can be interesting in the future
 547                         case 'VideoObject':
 548                         case 'ImageObject':
 549                         case 'LiveBlogPosting':
 550                         case 'SocialMediaPosting':
 551                                         // quit silently
 552                                 return $siteinfo;
 553                         default:
 554                                 Logger::info('Unsupported or unknown type', ['type' => $type, 'url' => $siteinfo['url']]);
 555                                 return $siteinfo;
 556                 }
 557         }
 558
 559         /**
 560          * Improve the siteinfo with information from the provided JSON-LD information concerning authors and publishers
 561          *
 562          * @param array $siteinfo
 563          * @param array $jsonld
 564          * @return array siteinfo
 565          */
 566         private static function parseJsonLdAuthor(array $siteinfo, array $jsonld)
 567         {
 568                 $jsonldinfo = [];
 569
 570                 if (!empty($jsonld['publisher']) && is_array($jsonld['publisher'])) {
 571                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'name', '@type', 'Organization');
 572                         if (!empty($content) && is_string($content)) {
 573                                 $jsonldinfo['publisher_name'] = trim($content);
 574                         }
 575
 576                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'url', '@type', 'Organization');
 577                         if (!empty($content) && is_string($content)) {
 578                                 $jsonldinfo['publisher_url'] = trim($content);
 579                         }
 580
 581                         $brand = JsonLD::fetchElement($jsonld, 'publisher', 'brand', '@type', 'Organization');
 582                         if (!empty($brand)) {
 583                                 $content = JsonLD::fetchElement($brand, 'name', '@type', 'brand');
 584                                 if (!empty($content) && is_string($content)) {
 585                                         $jsonldinfo['publisher_name'] = trim($content);
 586                                 }
 587                         }
 588                 }
 589
 590                 if (!empty($jsonld['author']) && is_array($jsonld['author'])) {
 591                         $content = JsonLD::fetchElement($jsonld, 'author', 'name', '@type', 'Organization');
 592                         if (!empty($content) && is_string($content)) {
 593                                 $jsonldinfo['publisher_name'] = trim($content);
 594                         }
 595
 596                         $content = JsonLD::fetchElement($jsonld, 'author', 'url', '@type', 'Organization');
 597                         if (!empty($content) && is_string($content)) {
 598                                 $jsonldinfo['publisher_url'] = trim($content);
 599                         }
 600
 601                         $content = JsonLD::fetchElement($jsonld, 'author', 'name', '@type', 'Person');
 602                         if (!empty($content) && is_string($content)) {
 603                                 $jsonldinfo['author_name'] = trim($content);
 604                         }
 605
 606                         $content = JsonLD::fetchElement($jsonld, 'author', 'url', '@type', 'Person');
 607                         if (!empty($content) && is_string($content)) {
 608                                 $jsonldinfo['author_url'] = trim($content);
 609                         }
 610                 }
 611
 612                 Logger::info('Fetched author information', ['fetched' => $jsonldinfo]);
 613
 614                 return array_merge($siteinfo, $jsonldinfo);
 615         }
 616
 617         /**
 618          * Improve the siteinfo with information from the provided JSON-LD Article information
 619          *
 620          * @param array $siteinfo
 621          * @param array $jsonld
 622          * @return array siteinfo
 623          */
 624         private static function parseJsonLdArticle(array $siteinfo, array $jsonld)
 625         {
 626                 $jsonldinfo = [];
 627
 628                 $content = JsonLD::fetchElement($jsonld, 'headline');
 629                 if (!empty($content) && is_string($content)) {
 630                         $jsonldinfo['title'] = trim($content);
 631                 }
 632
 633                 $content = JsonLD::fetchElement($jsonld, 'alternativeHeadline');
 634                 if (!empty($content) && is_string($content)) {
 635                         $jsonldinfo['alternative_title'] = trim($content);
 636                 }
 637
 638                 $content = JsonLD::fetchElement($jsonld, 'description');
 639                 if (!empty($content) && is_string($content)) {
 640                         $jsonldinfo['text'] = trim($content);
 641                 }
 642
 643                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
 644                 if (!empty($content) && is_string($content)) {
 645                         $jsonldinfo['image'] = trim($content);
 646                 }
 647
 648 /// @todo Check for the correct separator, also check for dpublicates before adding
 649 //              $content = JsonLD::fetchElement($jsonld, 'keywords');
 650 //              if (!empty($content) && is_string($content)) {
 651 //                      $jsonldinfo['keywords'] = trim($content);
 652 //              }
 653
 654                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 655
 656                 Logger::info('Fetched article information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 657
 658                 return array_merge($siteinfo, $jsonldinfo);
 659         }
 660
 661         /**
 662          * Improve the siteinfo with information from the provided JSON-LD WebPage information
 663          *
 664          * @param array $siteinfo
 665          * @param array $jsonld
 666          * @return array siteinfo
 667          */
 668         private static function parseJsonLdWebPage(array $siteinfo, array $jsonld)
 669         {
 670                 $jsonldinfo = [];
 671
 672                 $content = JsonLD::fetchElement($jsonld, 'name');
 673                 if (!empty($content)) {
 674                         $jsonldinfo['title'] = trim($content);
 675                 }
 676
 677                 $content = JsonLD::fetchElement($jsonld, 'description');
 678                 if (!empty($content)) {
 679                         $jsonldinfo['text'] = trim($content);
 680                 }
 681
 682                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 683                 if (!empty($content)) {
 684                         $jsonldinfo['image'] = trim($content);
 685                 }
 686
 687                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 688
 689                 Logger::info('Fetched webpage information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 690
 691                 return array_merge($siteinfo, $jsonldinfo);
 692         }
 693
 694         /**
 695          * Improve the siteinfo with information from the provided JSON-LD WebSite information
 696          *
 697          * @param array $siteinfo
 698          * @param array $jsonld
 699          * @return array siteinfo
 700          */
 701         private static function parseJsonLdWebSite(array $siteinfo, array $jsonld)
 702         {
 703                 $jsonldinfo = [];
 704
 705                 $content = JsonLD::fetchElement($jsonld, 'name');
 706                 if (!empty($content)) {
 707                         $jsonldinfo['publisher_name'] = trim($content);
 708                 }
 709
 710                 $content = JsonLD::fetchElement($jsonld, 'description');
 711                 if (!empty($content)) {
 712                         $jsonldinfo['publisher_description'] = trim($content);
 713                 }
 714
 715                 $content = JsonLD::fetchElement($jsonld, 'url');
 716                 if (!empty($content)) {
 717                         $jsonldinfo['publisher_url'] = trim($content);
 718                 }
 719
 720                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 721                 if (!empty($content)) {
 722                         $jsonldinfo['image'] = trim($content);
 723                 }
 724
 725                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 726
 727                 Logger::info('Fetched WebSite information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 728                 return array_merge($siteinfo, $jsonldinfo);
 729         }
 730
 731         /**
 732          * Improve the siteinfo with information from the provided JSON-LD Organization information
 733          *
 734          * @param array $siteinfo
 735          * @param array $jsonld
 736          * @return array siteinfo
 737          */
 738         private static function parseJsonLdWebOrganization(array $siteinfo, array $jsonld)
 739         {
 740                 $jsonldinfo = [];
 741
 742                 $content = JsonLD::fetchElement($jsonld, 'name');
 743                 if (!empty($content)) {
 744                         $jsonldinfo['publisher_name'] = trim($content);
 745                 }
 746
 747                 $content = JsonLD::fetchElement($jsonld, 'url');
 748                 if (!empty($content)) {
 749                         $jsonldinfo['publisher_url'] = trim($content);
 750                 }
 751
 752                 $content = JsonLD::fetchElement($jsonld, 'logo', 'url', '@type', 'ImageObject');
 753                 if (!empty($content)) {
 754                         $jsonldinfo['publisher_img'] = trim($content);
 755                 }
 756
 757                 Logger::info('Fetched Organization information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 758                 return array_merge($siteinfo, $jsonldinfo);
 759         }
 760
 761         /**
 762          * Improve the siteinfo with information from the provided JSON-LD Person information
 763          *
 764          * @param array $siteinfo
 765          * @param array $jsonld
 766          * @return array siteinfo
 767          */
 768         private static function parseJsonLdWebPerson(array $siteinfo, array $jsonld)
 769         {
 770                 $jsonldinfo = [];
 771
 772                 $content = JsonLD::fetchElement($jsonld, 'name');
 773                 if (!empty($content)) {
 774                         $jsonldinfo['author_name'] = trim($content);
 775                 }
 776
 777                 $content = JsonLD::fetchElement($jsonld, 'description');
 778                 if (!empty($content)) {
 779                         $jsonldinfo['author_description'] = trim($content);
 780                 }
 781
 782                 $content = JsonLD::fetchElement($jsonld, 'url');
 783                 if (!empty($content)) {
 784                         $jsonldinfo['author_url'] = trim($content);
 785                 }
 786
 787                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
 788                 if (!empty($content)) {
 789                         $jsonldinfo['author_img'] = trim($content);
 790                 }
 791
 792                 Logger::info('Fetched Person information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 793                 return array_merge($siteinfo, $jsonldinfo);
 794         }
 795
 796         /**
 797          * Convert tags from CSV to an array
 798          *
 799          * @param string $string Tags
 800          * @return array with formatted Hashtags
 801          */
 802         public static function convertTagsToArray($string)
 803         {
 804                 $arr_tags = str_getcsv($string);
 805                 if (count($arr_tags)) {
 806                         // add the # sign to every tag
 807                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 808
 809                         return $arr_tags;
 810                 }
 811         }
 812
 813         /**
 814          * Add a hasht sign to a string
 815          *
 816          * This method is used as callback function
 817          *
 818          * @param string $tag The pure tag name
 819          * @param int    $k   Counter for internal use
 820          * @return void
 821          */
 822         private static function arrAddHashes(&$tag, $k)
 823         {
 824                 $tag = "#" . $tag;
 825         }
 826
 827         /**
 828          * Add a scheme to an url
 829          *
 830          * The src attribute of some html elements (e.g. images)
 831          * can miss the scheme so we need to add the correct
 832          * scheme
 833          *
 834          * @param string $url    The url which possibly does have
 835          *                       a missing scheme (a link to an image)
 836          * @param string $scheme The url with a correct scheme
 837          *                       (e.g. the url from the webpage which does contain the image)
 838          *
 839          * @return string The url with a scheme
 840          */
 841         private static function completeUrl($url, $scheme)
 842         {
 843                 $urlarr = parse_url($url);
 844
 845                 // If the url does allready have an scheme
 846                 // we can stop the process here
 847                 if (isset($urlarr["scheme"])) {
 848                         return($url);
 849                 }
 850
 851                 $schemearr = parse_url($scheme);
 852
 853                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 854
 855                 if (!empty($schemearr["port"])) {
 856                         $complete .= ":".$schemearr["port"];
 857                 }
 858
 859                 if (!empty($urlarr["path"])) {
 860                         if (strpos($urlarr["path"], "/") !== 0) {
 861                                 $complete .= "/";
 862                         }
 863
 864                         $complete .= $urlarr["path"];
 865                 }
 866
 867                 if (!empty($urlarr["query"])) {
 868                         $complete .= "?".$urlarr["query"];
 869                 }
 870
 871                 if (!empty($urlarr["fragment"])) {
 872                         $complete .= "#".$urlarr["fragment"];
 873                 }
 874
 875                 return($complete);
 876         }
 877 }