src/Util/ParseUrl.php

   1 <?php
   2 /**
   3  * @copyright Copyright (C) 2010-2021, the Friendica project
   4  *
   5  * @license GNU AGPL version 3 or any later version
   6  *
   7  * This program is free software: you can redistribute it and/or modify
   8  * it under the terms of the GNU Affero General Public License as
   9  * published by the Free Software Foundation, either version 3 of the
  10  * License, or (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15  * GNU Affero General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU Affero General Public License
  18  * along with this program.  If not, see <https://www.gnu.org/licenses/>.
  19  *
  20  */
  21
  22 namespace Friendica\Util;
  23
  24 use DOMDocument;
  25 use DOMXPath;
  26 use Friendica\Content\OEmbed;
  27 use Friendica\Core\Hook;
  28 use Friendica\Core\Logger;
  29 use Friendica\Database\Database;
  30 use Friendica\Database\DBA;
  31 use Friendica\DI;
  32 use Friendica\Network\HTTPException;
  33 use Friendica\Network\HTTPClient\Client\HttpClientOptions;
  34
  35 /**
  36  * Get information about a given URL
  37  *
  38  * Class with methods for extracting certain content from an url
  39  */
  40 class ParseUrl
  41 {
  42         const DEFAULT_EXPIRATION_FAILURE = 'now + 1 day';
  43         const DEFAULT_EXPIRATION_SUCCESS = 'now + 3 months';
  44
  45         /**
  46          * Maximum number of characters for the description
  47          */
  48         const MAX_DESC_COUNT = 250;
  49
  50         /**
  51          * Minimum number of characters for the description
  52          */
  53         const MIN_DESC_COUNT = 100;
  54
  55         /**
  56          * Fetch the content type of the given url
  57          * @param string $url URL of the page
  58          * @return array content type
  59          */
  60         public static function getContentType(string $url)
  61         {
  62                 $curlResult = DI::httpClient()->head($url);
  63                 if (!$curlResult->isSuccess()) {
  64                         return [];
  65                 }
  66
  67                 $contenttype =  $curlResult->getHeader('Content-Type')[0] ?? '';
  68                 if (empty($contenttype)) {
  69                         return [];
  70                 }
  71
  72                 return explode('/', current(explode(';', $contenttype)));
  73         }
  74
  75         /**
  76          * Search for chached embeddable data of an url otherwise fetch it
  77          *
  78          * @param string $url         The url of the page which should be scraped
  79          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
  80          *                            to avoid endless loops
  81          *
  82          * @return array which contains needed data for embedding
  83          *    string 'url'      => The url of the parsed page
  84          *    string 'type'     => Content type
  85          *    string 'title'    => (optional) The title of the content
  86          *    string 'text'     => (optional) The description for the content
  87          *    string 'image'    => (optional) A preview image of the content
  88          *    array  'images'   => (optional) Array of preview pictures
  89          *    string 'keywords' => (optional) The tags which belong to the content
  90          *
  91          * @throws HTTPException\InternalServerErrorException
  92          * @see   ParseUrl::getSiteinfo() for more information about scraping
  93          * embeddable content
  94          */
  95         public static function getSiteinfoCached($url, $do_oembed = true): array
  96         {
  97                 if (empty($url)) {
  98                         return [
  99                                 'url' => '',
 100                                 'type' => 'error',
 101                         ];
 102                 }
 103
 104                 $urlHash = hash('sha256', $url);
 105
 106                 $parsed_url = DBA::selectFirst('parsed_url', ['content'],
 107                         ['url_hash' => $urlHash, 'oembed' => $do_oembed]
 108                 );
 109                 if (!empty($parsed_url['content'])) {
 110                         $data = unserialize($parsed_url['content']);
 111                         return $data;
 112                 }
 113
 114                 $data = self::getSiteinfo($url, $do_oembed);
 115
 116                 $expires = $data['expires'];
 117
 118                 unset($data['expires']);
 119
 120                 DI::dba()->insert(
 121                         'parsed_url',
 122                         [
 123                                 'url_hash' => $urlHash,
 124                                 'oembed'   => $do_oembed,
 125                                 'url'      => $url,
 126                                 'content'  => serialize($data),
 127                                 'created'  => DateTimeFormat::utcNow(),
 128                                 'expires'  => $expires,
 129                         ],
 130                         Database::INSERT_UPDATE
 131                 );
 132
 133                 return $data;
 134         }
 135
 136         /**
 137          * Parse a page for embeddable content information
 138          *
 139          * This method parses to url for meta data which can be used to embed
 140          * the content. If available it prioritizes Open Graph meta tags.
 141          * If this is not available it uses the twitter cards meta tags.
 142          * As fallback it uses standard html elements with meta informations
 143          * like \<title\>Awesome Title\</title\> or
 144          * \<meta name="description" content="An awesome description"\>
 145          *
 146          * @param string $url         The url of the page which should be scraped
 147          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
 148          *                            to avoid endless loops
 149          * @param int    $count       Internal counter to avoid endless loops
 150          *
 151          * @return array which contains needed data for embedding
 152          *    string 'url'      => The url of the parsed page
 153          *    string 'type'     => Content type (error, link, photo, image, audio, video)
 154          *    string 'title'    => (optional) The title of the content
 155          *    string 'text'     => (optional) The description for the content
 156          *    string 'image'    => (optional) A preview image of the content
 157          *    array  'images'   => (optional) Array of preview pictures
 158          *    string 'keywords' => (optional) The tags which belong to the content
 159          *
 160          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 161          * @todo  https://developers.google.com/+/plugins/snippet/
 162          * @verbatim
 163          * <meta itemprop="name" content="Awesome title">
 164          * <meta itemprop="description" content="An awesome description">
 165          * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
 166          *
 167          * <body itemscope itemtype="http://schema.org/Product">
 168          *   <h1 itemprop="name">Shiny Trinket</h1>
 169          *   <img itemprop="image" src="{image-url}" />
 170          *   <p itemprop="description">Shiny trinkets are shiny.</p>
 171          * </body>
 172          * @endverbatim
 173          */
 174         public static function getSiteinfo($url, $do_oembed = true, $count = 1)
 175         {
 176                 if (empty($url)) {
 177                         return [
 178                                 'url' => '',
 179                                 'type' => 'error',
 180                         ];
 181                 }
 182
 183                 // Check if the URL does contain a scheme
 184                 $scheme = parse_url($url, PHP_URL_SCHEME);
 185
 186                 if ($scheme == '') {
 187                         $url = 'http://' . ltrim($url, '/');
 188                 }
 189
 190                 $url = trim($url, "'\"");
 191
 192                 $url = Network::stripTrackingQueryParams($url);
 193
 194                 $siteinfo = [
 195                         'url' => $url,
 196                         'type' => 'link',
 197                         'expires' => DateTimeFormat::utc(self::DEFAULT_EXPIRATION_FAILURE),
 198                 ];
 199
 200                 if ($count > 10) {
 201                         Logger::notice('Endless loop detected', ['url' => $url]);
 202                         return $siteinfo;
 203                 }
 204
 205                 $type = self::getContentType($url);
 206                 Logger::info('Got content-type', ['content-type' => $type, 'url' => $url]);
 207                 if (!empty($type) && in_array($type[0], ['image', 'video', 'audio'])) {
 208                         $siteinfo['type'] = $type[0];
 209                         return $siteinfo;
 210                 }
 211
 212                 if ((count($type) >= 2) && (($type[0] != 'text') || ($type[1] != 'html'))) {
 213                         Logger::info('Unparseable content-type, quitting here, ', ['content-type' => $type, 'url' => $url]);
 214                         return $siteinfo;
 215                 }
 216
 217                 $curlResult = DI::httpClient()->get($url, [HttpClientOptions::CONTENT_LENGTH => 1000000]);
 218                 if (!$curlResult->isSuccess() || empty($curlResult->getBody())) {
 219                         return $siteinfo;
 220                 }
 221
 222                 $siteinfo['expires'] = DateTimeFormat::utc(self::DEFAULT_EXPIRATION_SUCCESS);
 223
 224                 if ($cacheControlHeader = $curlResult->getHeader('Cache-Control')[0] ?? '') {
 225                         if (preg_match('/max-age=([0-9]+)/i', $cacheControlHeader, $matches)) {
 226                                 $maxAge = max(86400, (int)array_pop($matches));
 227                                 $siteinfo['expires'] = DateTimeFormat::utc("now + $maxAge seconds");
 228                         }
 229                 }
 230
 231                 $body = $curlResult->getBody();
 232
 233                 if ($do_oembed) {
 234                         $oembed_data = OEmbed::fetchURL($url, false, false);
 235
 236                         if (!empty($oembed_data->type)) {
 237                                 if (!in_array($oembed_data->type, ['error', 'rich', 'image', 'video', 'audio', ''])) {
 238                                         $siteinfo['type'] = $oembed_data->type;
 239                                 }
 240
 241                                 // See https://github.com/friendica/friendica/pull/5763#discussion_r217913178
 242                                 if ($siteinfo['type'] != 'photo') {
 243                                         if (!empty($oembed_data->title)) {
 244                                                 $siteinfo['title'] = trim($oembed_data->title);
 245                                         }
 246                                         if (!empty($oembed_data->description)) {
 247                                                 $siteinfo['text'] = trim($oembed_data->description);
 248                                         }
 249                                         if (!empty($oembed_data->author_name)) {
 250                                                 $siteinfo['author_name'] = trim($oembed_data->author_name);
 251                                         }
 252                                         if (!empty($oembed_data->author_url)) {
 253                                                 $siteinfo['author_url'] = trim($oembed_data->author_url);
 254                                         }
 255                                         if (!empty($oembed_data->provider_name)) {
 256                                                 $siteinfo['publisher_name'] = trim($oembed_data->provider_name);
 257                                         }
 258                                         if (!empty($oembed_data->provider_url)) {
 259                                                 $siteinfo['publisher_url'] = trim($oembed_data->provider_url);
 260                                         }
 261                                         if (!empty($oembed_data->thumbnail_url)) {
 262                                                 $siteinfo['image'] = $oembed_data->thumbnail_url;
 263                                         }
 264                                 }
 265                         }
 266                 }
 267
 268                 $charset = '';
 269                 // Look for a charset, first in headers
 270                 // Expected form: Content-Type: text/html; charset=ISO-8859-4
 271                 if (preg_match('/charset=([a-z0-9-_.\/]+)/i', $curlResult->getContentType(), $matches)) {
 272                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 273                 }
 274
 275                 // Then in body that gets precedence
 276                 // Expected forms:
 277                 // - <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
 278                 // - <meta charset="utf-8">
 279                 // - <meta charset=utf-8>
 280                 // - <meta charSet="utf-8">
 281                 // We escape <style> and <script> tags since they can contain irrelevant charset information
 282                 // (see https://github.com/friendica/friendica/issues/9251#issuecomment-698636806)
 283                 Strings::performWithEscapedBlocks($body, '#<(?:style|script).*?</(?:style|script)>#ism', function ($body) use (&$charset) {
 284                         if (preg_match('/charset=["\']?([a-z0-9-_.\/]+)/i', $body, $matches)) {
 285                                 $charset = trim(trim(trim(array_pop($matches)), ';,'));
 286                         }
 287                 });
 288
 289                 $siteinfo['charset'] = $charset;
 290
 291                 if ($charset && strtoupper($charset) != 'UTF-8') {
 292                         // See https://github.com/friendica/friendica/issues/5470#issuecomment-418351211
 293                         $charset = str_ireplace('latin-1', 'latin1', $charset);
 294
 295                         Logger::info('detected charset', ['charset' => $charset]);
 296                         $body = iconv($charset, 'UTF-8//TRANSLIT', $body);
 297                 }
 298
 299                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', 'UTF-8');
 300
 301                 $doc = new DOMDocument();
 302                 @$doc->loadHTML($body);
 303
 304                 XML::deleteNode($doc, 'style');
 305                 XML::deleteNode($doc, 'option');
 306                 XML::deleteNode($doc, 'h1');
 307                 XML::deleteNode($doc, 'h2');
 308                 XML::deleteNode($doc, 'h3');
 309                 XML::deleteNode($doc, 'h4');
 310                 XML::deleteNode($doc, 'h5');
 311                 XML::deleteNode($doc, 'h6');
 312                 XML::deleteNode($doc, 'ol');
 313                 XML::deleteNode($doc, 'ul');
 314
 315                 $xpath = new DOMXPath($doc);
 316
 317                 $list = $xpath->query('//meta[@content]');
 318                 foreach ($list as $node) {
 319                         $meta_tag = [];
 320                         if ($node->attributes->length) {
 321                                 foreach ($node->attributes as $attribute) {
 322                                         $meta_tag[$attribute->name] = $attribute->value;
 323                                 }
 324                         }
 325
 326                         if (@$meta_tag['http-equiv'] == 'refresh') {
 327                                 $path = $meta_tag['content'];
 328                                 $pathinfo = explode(';', $path);
 329                                 $content = '';
 330                                 foreach ($pathinfo as $value) {
 331                                         if (substr(strtolower($value), 0, 4) == 'url=') {
 332                                                 $content = substr($value, 4);
 333                                         }
 334                                 }
 335                                 if ($content != '') {
 336                                         $siteinfo = self::getSiteinfo($content, $do_oembed, ++$count);
 337                                         return $siteinfo;
 338                                 }
 339                         }
 340                 }
 341
 342                 $list = $xpath->query('//title');
 343                 if ($list->length > 0) {
 344                         $siteinfo['title'] = trim($list->item(0)->nodeValue);
 345                 }
 346
 347                 $list = $xpath->query('//meta[@name]');
 348                 foreach ($list as $node) {
 349                         $meta_tag = [];
 350                         if ($node->attributes->length) {
 351                                 foreach ($node->attributes as $attribute) {
 352                                         $meta_tag[$attribute->name] = $attribute->value;
 353                                 }
 354                         }
 355
 356                         if (empty($meta_tag['content'])) {
 357                                 continue;
 358                         }
 359
 360                         $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 361
 362                         switch (strtolower($meta_tag['name'])) {
 363                                 case 'fulltitle':
 364                                         $siteinfo['title'] = trim($meta_tag['content']);
 365                                         break;
 366                                 case 'description':
 367                                         $siteinfo['text'] = trim($meta_tag['content']);
 368                                         break;
 369                                 case 'thumbnail':
 370                                         $siteinfo['image'] = $meta_tag['content'];
 371                                         break;
 372                                 case 'twitter:image':
 373                                         $siteinfo['image'] = $meta_tag['content'];
 374                                         break;
 375                                 case 'twitter:image:src':
 376                                         $siteinfo['image'] = $meta_tag['content'];
 377                                         break;
 378                                 case 'twitter:description':
 379                                         $siteinfo['text'] = trim($meta_tag['content']);
 380                                         break;
 381                                 case 'twitter:title':
 382                                         $siteinfo['title'] = trim($meta_tag['content']);
 383                                         break;
 384                                 case 'twitter:player':
 385                                         $siteinfo['player']['embed'] = trim($meta_tag['content']);
 386                                         break;
 387                                 case 'twitter:player:stream':
 388                                         $siteinfo['player']['stream'] = trim($meta_tag['content']);
 389                                         break;
 390                                 case 'twitter:player:width':
 391                                         $siteinfo['player']['width'] = intval($meta_tag['content']);
 392                                         break;
 393                                 case 'twitter:player:height':
 394                                         $siteinfo['player']['height'] = intval($meta_tag['content']);
 395                                         break;
 396                                 case 'dc.title':
 397                                         $siteinfo['title'] = trim($meta_tag['content']);
 398                                         break;
 399                                 case 'dc.description':
 400                                         $siteinfo['text'] = trim($meta_tag['content']);
 401                                         break;
 402                                 case 'dc.creator':
 403                                         $siteinfo['publisher_name'] = trim($meta_tag['content']);
 404                                         break;
 405                                 case 'keywords':
 406                                         $keywords = explode(',', $meta_tag['content']);
 407                                         break;
 408                                 case 'news_keywords':
 409                                         $keywords = explode(',', $meta_tag['content']);
 410                                         break;
 411                         }
 412                 }
 413
 414                 if (isset($keywords)) {
 415                         $siteinfo['keywords'] = [];
 416                         foreach ($keywords as $keyword) {
 417                                 if (!in_array(trim($keyword), $siteinfo['keywords'])) {
 418                                         $siteinfo['keywords'][] = trim($keyword);
 419                                 }
 420                         }
 421                 }
 422
 423                 $list = $xpath->query('//meta[@property]');
 424                 foreach ($list as $node) {
 425                         $meta_tag = [];
 426                         if ($node->attributes->length) {
 427                                 foreach ($node->attributes as $attribute) {
 428                                         $meta_tag[$attribute->name] = $attribute->value;
 429                                 }
 430                         }
 431
 432                         if (!empty($meta_tag['content'])) {
 433                                 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 434
 435                                 switch (strtolower($meta_tag['property'])) {
 436                                         case 'og:image':
 437                                                 $siteinfo['image'] = $meta_tag['content'];
 438                                                 break;
 439                                         case 'og:image:url':
 440                                                 $siteinfo['image'] = $meta_tag['content'];
 441                                                 break;
 442                                         case 'og:image:secure_url':
 443                                                 $siteinfo['image'] = $meta_tag['content'];
 444                                                 break;
 445                                         case 'og:title':
 446                                                 $siteinfo['title'] = trim($meta_tag['content']);
 447                                                 break;
 448                                         case 'og:description':
 449                                                 $siteinfo['text'] = trim($meta_tag['content']);
 450                                                 break;
 451                                         case 'og:site_name':
 452                                                 $siteinfo['publisher_name'] = trim($meta_tag['content']);
 453                                                 break;
 454                                         case 'twitter:description':
 455                                                 $siteinfo['text'] = trim($meta_tag['content']);
 456                                                 break;
 457                                         case 'twitter:title':
 458                                                 $siteinfo['title'] = trim($meta_tag['content']);
 459                                                 break;
 460                                         case 'twitter:image':
 461                                                 $siteinfo['image'] = $meta_tag['content'];
 462                                                 break;
 463                                 }
 464                         }
 465                 }
 466
 467                 $list = $xpath->query("//script[@type='application/ld+json']");
 468                 foreach ($list as $node) {
 469                         if (!empty($node->nodeValue)) {
 470                                 if ($jsonld = json_decode($node->nodeValue, true)) {
 471                                         $siteinfo = self::parseParts($siteinfo, $jsonld);
 472                                 }
 473                         }
 474                 }
 475
 476                 if (!empty($siteinfo['player']['stream'])) {
 477                         // Only add player data to media arrays if there is no duplicate
 478                         $content_urls = array_merge(array_column($siteinfo['audio'] ?? [], 'content'), array_column($siteinfo['video'] ?? [], 'content'));
 479                         if (!in_array($siteinfo['player']['stream'], $content_urls)) {
 480                                 $contenttype = self::getContentType($siteinfo['player']['stream']);
 481                                 if (!empty($contenttype[0]) && in_array($contenttype[0], ['audio', 'video'])) {
 482                                         $media = ['content' => $siteinfo['player']['stream']];
 483
 484                                         if (!empty($siteinfo['player']['embed'])) {
 485                                                 $media['embed'] = $siteinfo['player']['embed'];
 486                                         }
 487
 488                                         $siteinfo[$contenttype[0]][] = $media;
 489                                 }
 490                         }
 491                 }
 492
 493                 if (!empty($siteinfo['image'])) {
 494                         $siteinfo['images'] = $siteinfo['images'] ?? [];
 495                         array_unshift($siteinfo['images'], ['url' => $siteinfo['image']]);
 496                         unset($siteinfo['image']);
 497                 }
 498
 499                 $siteinfo = self::checkMedia($url, $siteinfo);
 500
 501                 if (!empty($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', $siteinfo);
 512
 513                 ksort($siteinfo);
 514
 515                 return $siteinfo;
 516         }
 517
 518         /**
 519          * Check the attached media elements.
 520          * Fix existing data and add missing data.
 521          *
 522          * @param string $page_url
 523          * @param array $siteinfo
 524          * @return void
 525          */
 526         private static function checkMedia(string $page_url, array $siteinfo) : array
 527         {
 528                 if (!empty($siteinfo['images'])) {
 529                         array_walk($siteinfo['images'], function (&$image) use ($page_url) {
 530                                 // According to the specifications someone could place a picture url into the content field as well.
 531                                 // But this doesn't seem to happen in the wild, so we don't cover it here.
 532                                 if (!empty($image['url'])) {
 533                                         $image['url'] = self::completeUrl($image['url'], $page_url);
 534                                         $photodata = Images::getInfoFromURLCached($image['url']);
 535                                         if (!empty($photodata) && ($photodata[0] > 50) && ($photodata[1] > 50)) {
 536                                                 $image['src'] = $image['url'];
 537                                                 $image['width'] = $photodata[0];
 538                                                 $image['height'] = $photodata[1];
 539                                                 $image['contenttype'] = $photodata['mime'];
 540                                                 unset($image['url']);
 541                                                 ksort($image);
 542                                         } else {
 543                                                 $image = [];
 544                                         }
 545                                 } else {
 546                                         $image = [];
 547                                 }
 548                         });
 549
 550                         $siteinfo['images'] = array_values(array_filter($siteinfo['images']));
 551                 }
 552
 553                 foreach (['audio', 'video'] as $element) {
 554                         if (!empty($siteinfo[$element])) {
 555                                 array_walk($siteinfo[$element], function (&$media) use ($page_url, &$siteinfo) {
 556                                         $url = '';
 557                                         $embed = '';
 558                                         $content = '';
 559                                         $contenttype = '';
 560                                         foreach (['embed', 'content', 'url'] as $field) {
 561                                                 if (!empty($media[$field])) {
 562                                                         $media[$field] = self::completeUrl($media[$field], $page_url);
 563                                                         $type = self::getContentType($media[$field]);
 564                                                         if (($type[0] ?? '') == 'text') {
 565                                                                 if ($field == 'embed') {
 566                                                                         $embed = $media[$field];
 567                                                                 } else {
 568                                                                         $url = $media[$field];
 569                                                                 }
 570                                                         } elseif (!empty($type[0])) {
 571                                                                 $content = $media[$field];
 572                                                                 $contenttype = implode('/', $type);
 573                                                         }
 574                                                 }
 575                                                 unset($media[$field]);
 576                                         }
 577
 578                                         foreach (['image', 'preview'] as $field) {
 579                                                 if (!empty($media[$field])) {
 580                                                         $media[$field] = self::completeUrl($media[$field], $page_url);
 581                                                 }
 582                                         }
 583
 584                                         if (!empty($url)) {
 585                                                 $media['url'] = $url;
 586                                         }
 587                                         if (!empty($embed)) {
 588                                                 $media['embed'] = $embed;
 589                                                 if (empty($siteinfo['player']['embed'])) {
 590                                                         $siteinfo['player']['embed'] = $embed;
 591                                                 }
 592                                         }
 593                                         if (!empty($content)) {
 594                                                 $media['src'] = $content;
 595                                         }
 596                                         if (!empty($contenttype)) {
 597                                                 $media['contenttype'] = $contenttype;
 598                                         }
 599                                         if (empty($url) && empty($content) && empty($embed)) {
 600                                                 $media = [];
 601                                         }
 602                                         ksort($media);
 603                                 });
 604
 605                                 $siteinfo[$element] = array_values(array_filter($siteinfo[$element]));
 606                         }
 607                         if (empty($siteinfo[$element])) {
 608                                 unset($siteinfo[$element]);
 609                         }
 610                 }
 611                 return $siteinfo;
 612         }
 613
 614         /**
 615          * Convert tags from CSV to an array
 616          *
 617          * @param string $string Tags
 618          * @return array with formatted Hashtags
 619          */
 620         public static function convertTagsToArray($string)
 621         {
 622                 $arr_tags = str_getcsv($string);
 623                 if (count($arr_tags)) {
 624                         // add the # sign to every tag
 625                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 626
 627                         return $arr_tags;
 628                 }
 629         }
 630
 631         /**
 632          * Add a hasht sign to a string
 633          *
 634          * This method is used as callback function
 635          *
 636          * @param string $tag The pure tag name
 637          * @param int    $k   Counter for internal use
 638          * @return void
 639          */
 640         private static function arrAddHashes(&$tag, $k)
 641         {
 642                 $tag = "#" . $tag;
 643         }
 644
 645         /**
 646          * Add a scheme to an url
 647          *
 648          * The src attribute of some html elements (e.g. images)
 649          * can miss the scheme so we need to add the correct
 650          * scheme
 651          *
 652          * @param string $url    The url which possibly does have
 653          *                       a missing scheme (a link to an image)
 654          * @param string $scheme The url with a correct scheme
 655          *                       (e.g. the url from the webpage which does contain the image)
 656          *
 657          * @return string The url with a scheme
 658          */
 659         private static function completeUrl($url, $scheme)
 660         {
 661                 $urlarr = parse_url($url);
 662
 663                 // If the url does allready have an scheme
 664                 // we can stop the process here
 665                 if (isset($urlarr["scheme"])) {
 666                         return($url);
 667                 }
 668
 669                 $schemearr = parse_url($scheme);
 670
 671                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 672
 673                 if (!empty($schemearr["port"])) {
 674                         $complete .= ":".$schemearr["port"];
 675                 }
 676
 677                 if (!empty($urlarr["path"])) {
 678                         if (strpos($urlarr["path"], "/") !== 0) {
 679                                 $complete .= "/";
 680                         }
 681
 682                         $complete .= $urlarr["path"];
 683                 }
 684
 685                 if (!empty($urlarr["query"])) {
 686                         $complete .= "?".$urlarr["query"];
 687                 }
 688
 689                 if (!empty($urlarr["fragment"])) {
 690                         $complete .= "#".$urlarr["fragment"];
 691                 }
 692
 693                 return($complete);
 694         }
 695
 696         /**
 697          * Parse the Json-Ld parts of a web page
 698          *
 699          * @param array $siteinfo
 700          * @param array $jsonld
 701          * @return array siteinfo
 702          */
 703         private static function parseParts(array $siteinfo, array $jsonld)
 704         {
 705                 if (!empty($jsonld['@graph']) && is_array($jsonld['@graph'])) {
 706                         foreach ($jsonld['@graph'] as $part) {
 707                                 if (!empty($part) && is_array($part)) {
 708                                         $siteinfo = self::parseParts($siteinfo, $part);
 709                                 }
 710                         }
 711                 } elseif (!empty($jsonld['@type'])) {
 712                         $siteinfo = self::parseJsonLd($siteinfo, $jsonld);
 713                 } elseif (!empty($jsonld)) {
 714                         $keys = array_keys($jsonld);
 715                         $numeric_keys = true;
 716                         foreach ($keys as $key) {
 717                                 if (!is_int($key)) {
 718                                         $numeric_keys = false;
 719                                 }
 720                         }
 721                         if ($numeric_keys) {
 722                                 foreach ($jsonld as $part) {
 723                                         if (!empty($part) && is_array($part)) {
 724                                                 $siteinfo = self::parseParts($siteinfo, $part);
 725                                         }
 726                                 }
 727                         }
 728                 }
 729
 730                 array_walk_recursive($siteinfo, function (&$element) {
 731                         if (is_string($element)) {
 732                                 $element = trim(strip_tags(html_entity_decode($element, ENT_COMPAT, 'UTF-8')));
 733                         }
 734                 });
 735
 736                 return $siteinfo;
 737         }
 738
 739         /**
 740          * Improve the siteinfo with information from the provided JSON-LD information
 741          * @see https://jsonld.com/
 742          * @see https://schema.org/
 743          *
 744          * @param array $siteinfo
 745          * @param array $jsonld
 746          * @return array siteinfo
 747          */
 748         private static function parseJsonLd(array $siteinfo, array $jsonld)
 749         {
 750                 $type = JsonLD::fetchElement($jsonld, '@type');
 751                 if (empty($type)) {
 752                         Logger::info('Empty type', ['url' => $siteinfo['url']]);
 753                         return $siteinfo;
 754                 }
 755
 756                 // Silently ignore some types that aren't processed
 757                 if (in_array($type, ['SiteNavigationElement', 'JobPosting', 'CreativeWork', 'MusicAlbum',
 758                         'WPHeader', 'WPSideBar', 'WPFooter', 'LegalService', 'MusicRecording',
 759                         'ItemList', 'BreadcrumbList', 'Blog', 'Dataset', 'Product'])) {
 760                         return $siteinfo;
 761                 }
 762
 763                 switch ($type) {
 764                         case 'Article':
 765                         case 'AdvertiserContentArticle':
 766                         case 'NewsArticle':
 767                         case 'Report':
 768                         case 'SatiricalArticle':
 769                         case 'ScholarlyArticle':
 770                         case 'SocialMediaPosting':
 771                         case 'TechArticle':
 772                         case 'ReportageNewsArticle':
 773                         case 'SocialMediaPosting':
 774                         case 'BlogPosting':
 775                         case 'LiveBlogPosting':
 776                         case 'DiscussionForumPosting':
 777                                 return self::parseJsonLdArticle($siteinfo, $jsonld);
 778                         case 'WebPage':
 779                         case 'AboutPage':
 780                         case 'CheckoutPage':
 781                         case 'CollectionPage':
 782                         case 'ContactPage':
 783                         case 'FAQPage':
 784                         case 'ItemPage':
 785                         case 'MedicalWebPage':
 786                         case 'ProfilePage':
 787                         case 'QAPage':
 788                         case 'RealEstateListing':
 789                         case 'SearchResultsPage':
 790                         case 'MediaGallery':
 791                         case 'ImageGallery':
 792                         case 'VideoGallery':
 793                         case 'RadioEpisode':
 794                         case 'Event':
 795                                 return self::parseJsonLdWebPage($siteinfo, $jsonld);
 796                         case 'WebSite':
 797                                 return self::parseJsonLdWebSite($siteinfo, $jsonld);
 798                         case 'Organization':
 799                         case 'Airline':
 800                         case 'Consortium':
 801                         case 'Corporation':
 802                         case 'EducationalOrganization':
 803                         case 'FundingScheme':
 804                         case 'GovernmentOrganization':
 805                         case 'LibrarySystem':
 806                         case 'LocalBusiness':
 807                         case 'MedicalOrganization':
 808                         case 'NGO':
 809                         case 'NewsMediaOrganization':
 810                         case 'Project':
 811                         case 'SportsOrganization':
 812                         case 'WorkersUnion':
 813                                 return self::parseJsonLdWebOrganization($siteinfo, $jsonld);
 814                         case 'Person':
 815                         case 'Patient':
 816                         case 'PerformingGroup':
 817                         case 'DanceGroup';
 818                         case 'MusicGroup':
 819                         case 'TheaterGroup':
 820                                 return self::parseJsonLdWebPerson($siteinfo, $jsonld);
 821                         case 'AudioObject':
 822                         case 'Audio':
 823                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'audio');
 824                         case 'VideoObject':
 825                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'video');
 826                         case 'ImageObject':
 827                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'images');
 828                         default:
 829                                 Logger::info('Unknown type', ['type' => $type, 'url' => $siteinfo['url']]);
 830                                 return $siteinfo;
 831                 }
 832         }
 833
 834         /**
 835          * Fetch author and publisher data
 836          *
 837          * @param array $siteinfo
 838          * @param array $jsonld
 839          * @return array siteinfo
 840          */
 841         private static function parseJsonLdAuthor(array $siteinfo, array $jsonld)
 842         {
 843                 $jsonldinfo = [];
 844
 845                 if (!empty($jsonld['publisher']) && is_array($jsonld['publisher'])) {
 846                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'name');
 847                         if (!empty($content) && is_string($content)) {
 848                                 $jsonldinfo['publisher_name'] = trim($content);
 849                         }
 850
 851                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'url');
 852                         if (!empty($content) && is_string($content)) {
 853                                 $jsonldinfo['publisher_url'] = trim($content);
 854                         }
 855
 856                         $brand = JsonLD::fetchElement($jsonld, 'publisher', 'brand', '@type', 'Organization');
 857                         if (!empty($brand) && is_array($brand)) {
 858                                 $content = JsonLD::fetchElement($brand, 'name');
 859                                 if (!empty($content) && is_string($content)) {
 860                                         $jsonldinfo['publisher_name'] = trim($content);
 861                                 }
 862
 863                                 $content = JsonLD::fetchElement($brand, 'url');
 864                                 if (!empty($content) && is_string($content)) {
 865                                         $jsonldinfo['publisher_url'] = trim($content);
 866                                 }
 867
 868                                 $content = JsonLD::fetchElement($brand, 'logo', 'url');
 869                                 if (!empty($content) && is_string($content)) {
 870                                         $jsonldinfo['publisher_img'] = trim($content);
 871                                 }
 872                         }
 873
 874                         $logo = JsonLD::fetchElement($jsonld, 'publisher', 'logo');
 875                         if (!empty($logo) && is_array($logo)) {
 876                                 $content = JsonLD::fetchElement($logo, 'url');
 877                                 if (!empty($content) && is_string($content)) {
 878                                         $jsonldinfo['publisher_img'] = trim($content);
 879                                 }
 880                         }
 881                 } elseif (!empty($jsonld['publisher']) && is_string($jsonld['publisher'])) {
 882                         $jsonldinfo['publisher_name'] = trim($jsonld['publisher']);
 883                 }
 884
 885                 if (!empty($jsonld['author']) && is_array($jsonld['author'])) {
 886                         $content = JsonLD::fetchElement($jsonld, 'author', 'name');
 887                         if (!empty($content) && is_string($content)) {
 888                                 $jsonldinfo['author_name'] = trim($content);
 889                         }
 890
 891                         $content = JsonLD::fetchElement($jsonld, 'author', 'sameAs');
 892                         if (!empty($content) && is_string($content)) {
 893                                 $jsonldinfo['author_url'] = trim($content);
 894                         }
 895
 896                         $content = JsonLD::fetchElement($jsonld, 'author', 'url');
 897                         if (!empty($content) && is_string($content)) {
 898                                 $jsonldinfo['author_url'] = trim($content);
 899                         }
 900
 901                         $logo = JsonLD::fetchElement($jsonld, 'author', 'logo');
 902                         if (!empty($logo) && is_array($logo)) {
 903                                 $content = JsonLD::fetchElement($logo, 'url');
 904                                 if (!empty($content) && is_string($content)) {
 905                                         $jsonldinfo['author_img'] = trim($content);
 906                                 }
 907                         }
 908                 } elseif (!empty($jsonld['author']) && is_string($jsonld['author'])) {
 909                         $jsonldinfo['author_name'] = trim($jsonld['author']);
 910                 }
 911
 912                 Logger::info('Fetched Author information', ['fetched' => $jsonldinfo]);
 913
 914                 return array_merge($siteinfo, $jsonldinfo);
 915         }
 916
 917         /**
 918          * Fetch data from the provided JSON-LD Article type
 919          * @see https://schema.org/Article
 920          *
 921          * @param array $siteinfo
 922          * @param array $jsonld
 923          * @return array siteinfo
 924          */
 925         private static function parseJsonLdArticle(array $siteinfo, array $jsonld)
 926         {
 927                 $jsonldinfo = [];
 928
 929                 $content = JsonLD::fetchElement($jsonld, 'headline');
 930                 if (!empty($content) && is_string($content)) {
 931                         $jsonldinfo['title'] = trim($content);
 932                 }
 933
 934                 $content = JsonLD::fetchElement($jsonld, 'alternativeHeadline');
 935                 if (!empty($content) && is_string($content) && (($jsonldinfo['title'] ?? '') != trim($content))) {
 936                         $jsonldinfo['alternative_title'] = trim($content);
 937                 }
 938
 939                 $content = JsonLD::fetchElement($jsonld, 'description');
 940                 if (!empty($content) && is_string($content)) {
 941                         $jsonldinfo['text'] = trim($content);
 942                 }
 943
 944                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 945                 if (!empty($content)) {
 946                         $jsonldinfo['image'] = trim($content);
 947                 }
 948
 949                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
 950                 if (!empty($content) && is_string($content)) {
 951                         $jsonldinfo['image'] = trim($content);
 952                 }
 953
 954                 if (!empty($jsonld['keywords']) && !is_array($jsonld['keywords'])) {
 955                         $content = JsonLD::fetchElement($jsonld, 'keywords');
 956                         if (!empty($content)) {
 957                                 $siteinfo['keywords'] = [];
 958                                 $keywords = explode(',', $content);
 959                                 foreach ($keywords as $keyword) {
 960                                         $siteinfo['keywords'][] = trim($keyword);
 961                                 }
 962                         }
 963                 } elseif (!empty($jsonld['keywords'])) {
 964                         $content = JsonLD::fetchElementArray($jsonld, 'keywords');
 965                         if (!empty($content) && is_array($content)) {
 966                                 $jsonldinfo['keywords'] = $content;
 967                         }
 968                 }
 969
 970                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 971
 972                 Logger::info('Fetched article information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 973
 974                 return array_merge($siteinfo, $jsonldinfo);
 975         }
 976
 977         /**
 978          * Fetch data from the provided JSON-LD WebPage type
 979          * @see https://schema.org/WebPage
 980          *
 981          * @param array $siteinfo
 982          * @param array $jsonld
 983          * @return array siteinfo
 984          */
 985         private static function parseJsonLdWebPage(array $siteinfo, array $jsonld)
 986         {
 987                 $jsonldinfo = [];
 988
 989                 $content = JsonLD::fetchElement($jsonld, 'name');
 990                 if (!empty($content)) {
 991                         $jsonldinfo['title'] = trim($content);
 992                 }
 993
 994                 $content = JsonLD::fetchElement($jsonld, 'description');
 995                 if (!empty($content)) {
 996                         $jsonldinfo['text'] = trim($content);
 997                 }
 998
 999                 $content = JsonLD::fetchElement($jsonld, 'image');
1000                 if (!empty($content)) {
1001                         $jsonldinfo['image'] = trim($content);
1002                 }
1003
1004                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
1005                 if (!empty($content)) {
1006                         $jsonldinfo['image'] = trim($content);
1007                 }
1008
1009                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
1010
1011                 Logger::info('Fetched WebPage information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1012
1013                 return array_merge($siteinfo, $jsonldinfo);
1014         }
1015
1016         /**
1017          * Fetch data from the provided JSON-LD WebSite type
1018          * @see https://schema.org/WebSite
1019          *
1020          * @param array $siteinfo
1021          * @param array $jsonld
1022          * @return array siteinfo
1023          */
1024         private static function parseJsonLdWebSite(array $siteinfo, array $jsonld)
1025         {
1026                 $jsonldinfo = [];
1027
1028                 $content = JsonLD::fetchElement($jsonld, 'name');
1029                 if (!empty($content)) {
1030                         $jsonldinfo['publisher_name'] = trim($content);
1031                 }
1032
1033                 $content = JsonLD::fetchElement($jsonld, 'description');
1034                 if (!empty($content)) {
1035                         $jsonldinfo['publisher_description'] = trim($content);
1036                 }
1037
1038                 $content = JsonLD::fetchElement($jsonld, 'url');
1039                 if (!empty($content)) {
1040                         $jsonldinfo['publisher_url'] = trim($content);
1041                 }
1042
1043                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
1044                 if (!empty($content)) {
1045                         $jsonldinfo['image'] = trim($content);
1046                 }
1047
1048                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
1049
1050                 Logger::info('Fetched WebSite information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1051                 return array_merge($siteinfo, $jsonldinfo);
1052         }
1053
1054         /**
1055          * Fetch data from the provided JSON-LD Organization type
1056          * @see https://schema.org/Organization
1057          *
1058          * @param array $siteinfo
1059          * @param array $jsonld
1060          * @return array siteinfo
1061          */
1062         private static function parseJsonLdWebOrganization(array $siteinfo, array $jsonld)
1063         {
1064                 $jsonldinfo = [];
1065
1066                 $content = JsonLD::fetchElement($jsonld, 'name');
1067                 if (!empty($content)) {
1068                         $jsonldinfo['publisher_name'] = trim($content);
1069                 }
1070
1071                 $content = JsonLD::fetchElement($jsonld, 'description');
1072                 if (!empty($content)) {
1073                         $jsonldinfo['publisher_description'] = trim($content);
1074                 }
1075
1076                 $content = JsonLD::fetchElement($jsonld, 'url');
1077                 if (!empty($content)) {
1078                         $jsonldinfo['publisher_url'] = trim($content);
1079                 }
1080
1081                 $content = JsonLD::fetchElement($jsonld, 'logo', 'url', '@type', 'ImageObject');
1082                 if (!empty($content)) {
1083                         $jsonldinfo['publisher_img'] = trim($content);
1084                 }
1085
1086                 $content = JsonLD::fetchElement($jsonld, 'brand', 'name', '@type', 'Organization');
1087                 if (!empty($content)) {
1088                         $jsonldinfo['publisher_name'] = trim($content);
1089                 }
1090
1091                 $content = JsonLD::fetchElement($jsonld, 'brand', 'url', '@type', 'Organization');
1092                 if (!empty($content)) {
1093                         $jsonldinfo['publisher_url'] = trim($content);
1094                 }
1095
1096                 Logger::info('Fetched Organization information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1097                 return array_merge($siteinfo, $jsonldinfo);
1098         }
1099
1100         /**
1101          * Fetch data from the provided JSON-LD Person type
1102          * @see https://schema.org/Person
1103          *
1104          * @param array $siteinfo
1105          * @param array $jsonld
1106          * @return array siteinfo
1107          */
1108         private static function parseJsonLdWebPerson(array $siteinfo, array $jsonld)
1109         {
1110                 $jsonldinfo = [];
1111
1112                 $content = JsonLD::fetchElement($jsonld, 'name');
1113                 if (!empty($content)) {
1114                         $jsonldinfo['author_name'] = trim($content);
1115                 }
1116
1117                 $content = JsonLD::fetchElement($jsonld, 'description');
1118                 if (!empty($content)) {
1119                         $jsonldinfo['author_description'] = trim($content);
1120                 }
1121
1122                 $content = JsonLD::fetchElement($jsonld, 'sameAs');
1123                 if (!empty($content) && is_string($content)) {
1124                         $jsonldinfo['author_url'] = trim($content);
1125                 }
1126
1127                 $content = JsonLD::fetchElement($jsonld, 'url');
1128                 if (!empty($content)) {
1129                         $jsonldinfo['author_url'] = trim($content);
1130                 }
1131
1132                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
1133                 if (!empty($content) && !is_string($content)) {
1134                         Logger::notice('Unexpected return value for the author image', ['content' => $content]);
1135                 }
1136
1137                 if (!empty($content) && is_string($content)) {
1138                         $jsonldinfo['author_img'] = trim($content);
1139                 }
1140
1141                 Logger::info('Fetched Person information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1142                 return array_merge($siteinfo, $jsonldinfo);
1143         }
1144
1145         /**
1146          * Fetch data from the provided JSON-LD MediaObject type
1147          * @see https://schema.org/MediaObject
1148          *
1149          * @param array $siteinfo
1150          * @param array $jsonld
1151          * @return array siteinfo
1152          */
1153         private static function parseJsonLdMediaObject(array $siteinfo, array $jsonld, string $name)
1154         {
1155                 $media = [];
1156
1157                 $content = JsonLD::fetchElement($jsonld, 'caption');
1158                 if (!empty($content)) {
1159                         $media['caption'] = trim($content);
1160                 }
1161
1162                 $content = JsonLD::fetchElement($jsonld, 'url');
1163                 if (!empty($content)) {
1164                         $media['url'] = trim($content);
1165                 }
1166
1167                 $content = JsonLD::fetchElement($jsonld, 'mainEntityOfPage');
1168                 if (!empty($content)) {
1169                         $media['main'] = Strings::compareLink($content, $siteinfo['url']);
1170                 }
1171
1172                 $content = JsonLD::fetchElement($jsonld, 'description');
1173                 if (!empty($content)) {
1174                         $media['description'] = trim($content);
1175                 }
1176
1177                 $content = JsonLD::fetchElement($jsonld, 'name');
1178                 if (!empty($content) && (($media['description'] ?? '') != trim($content))) {
1179                         $media['name'] = trim($content);
1180                 }
1181
1182                 $content = JsonLD::fetchElement($jsonld, 'contentUrl');
1183                 if (!empty($content)) {
1184                         $media['content'] = trim($content);
1185                 }
1186
1187                 $content = JsonLD::fetchElement($jsonld, 'embedUrl');
1188                 if (!empty($content)) {
1189                         $media['embed'] = trim($content);
1190                 }
1191
1192                 $content = JsonLD::fetchElement($jsonld, 'height');
1193                 if (!empty($content)) {
1194                         $media['height'] = trim($content);
1195                 }
1196
1197                 $content = JsonLD::fetchElement($jsonld, 'width');
1198                 if (!empty($content)) {
1199                         $media['width'] = trim($content);
1200                 }
1201
1202                 $content = JsonLD::fetchElement($jsonld, 'image');
1203                 if (!empty($content)) {
1204                         $media['image'] = trim($content);
1205                 }
1206
1207                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
1208                 if (!empty($content) && (($media['image'] ?? '') != trim($content))) {
1209                         if (!empty($media['image'])) {
1210                                 $media['preview'] = trim($content);
1211                         } else {
1212                                 $media['image'] = trim($content);
1213                         }
1214                 }
1215
1216                 Logger::info('Fetched Media information', ['url' => $siteinfo['url'], 'fetched' => $media]);
1217                 $siteinfo[$name][] = $media;
1218                 return $siteinfo;
1219         }
1220 }