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                 $twitter_card = false;
 348                 $twitter_image = false;
 349
 350                 $list = $xpath->query('//meta[@name]');
 351                 foreach ($list as $node) {
 352                         $meta_tag = [];
 353                         if ($node->attributes->length) {
 354                                 foreach ($node->attributes as $attribute) {
 355                                         $meta_tag[$attribute->name] = $attribute->value;
 356                                 }
 357                         }
 358
 359                         if (empty($meta_tag['content'])) {
 360                                 continue;
 361                         }
 362
 363                         $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 364
 365                         switch (strtolower($meta_tag['name'])) {
 366                                 case 'fulltitle':
 367                                         $siteinfo['title'] = trim($meta_tag['content']);
 368                                         break;
 369                                 case 'description':
 370                                         $siteinfo['text'] = trim($meta_tag['content']);
 371                                         break;
 372                                 case 'thumbnail':
 373                                         $siteinfo['image'] = $meta_tag['content'];
 374                                         break;
 375                                 case 'twitter:image':
 376                                         $siteinfo['image'] = $meta_tag['content'];
 377                                         $twitter_image = true;
 378                                         break;
 379                                 case 'twitter:image:src':
 380                                         $siteinfo['image'] = $meta_tag['content'];
 381                                         break;
 382                                 case 'twitter:card':
 383                                         // Detect photo pages
 384                                         if ($meta_tag['content'] == 'summary_large_image') {
 385                                                 $twitter_card = true;
 386                                         }
 387                                         break;
 388                                 case 'twitter:description':
 389                                         $siteinfo['text'] = trim($meta_tag['content']);
 390                                         break;
 391                                 case 'twitter:title':
 392                                         $siteinfo['title'] = trim($meta_tag['content']);
 393                                         break;
 394                                 case 'dc.title':
 395                                         $siteinfo['title'] = trim($meta_tag['content']);
 396                                         break;
 397                                 case 'dc.description':
 398                                         $siteinfo['text'] = trim($meta_tag['content']);
 399                                         break;
 400                                 case 'dc.creator':
 401                                         $siteinfo['publisher_name'] = trim($meta_tag['content']);
 402                                         break;
 403                                 case 'keywords':
 404                                         $keywords = explode(',', $meta_tag['content']);
 405                                         break;
 406                                 case 'news_keywords':
 407                                         $keywords = explode(',', $meta_tag['content']);
 408                                         break;
 409                         }
 410                 }
 411
 412                 if (isset($keywords)) {
 413                         $siteinfo['keywords'] = [];
 414                         foreach ($keywords as $keyword) {
 415                                 if (!in_array(trim($keyword), $siteinfo['keywords'])) {
 416                                         $siteinfo['keywords'][] = trim($keyword);
 417                                 }
 418                         }
 419                 }
 420
 421                 $list = $xpath->query('//meta[@property]');
 422                 foreach ($list as $node) {
 423                         $meta_tag = [];
 424                         if ($node->attributes->length) {
 425                                 foreach ($node->attributes as $attribute) {
 426                                         $meta_tag[$attribute->name] = $attribute->value;
 427                                 }
 428                         }
 429
 430                         if (!empty($meta_tag['content'])) {
 431                                 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 432
 433                                 switch (strtolower($meta_tag['property'])) {
 434                                         case 'og:image':
 435                                                 $siteinfo['image'] = $meta_tag['content'];
 436                                                 break;
 437                                         case 'og:image:url':
 438                                                 $siteinfo['image'] = $meta_tag['content'];
 439                                                 break;
 440                                         case 'og:image:secure_url':
 441                                                 $siteinfo['image'] = $meta_tag['content'];
 442                                                 break;
 443                                         case 'og:title':
 444                                                 $siteinfo['title'] = trim($meta_tag['content']);
 445                                                 break;
 446                                         case 'og:description':
 447                                                 $siteinfo['text'] = trim($meta_tag['content']);
 448                                                 break;
 449                                         case 'og:site_name':
 450                                                 $siteinfo['publisher_name'] = trim($meta_tag['content']);
 451                                                 break;
 452                                         case 'twitter:description':
 453                                                 $siteinfo['text'] = trim($meta_tag['content']);
 454                                                 break;
 455                                         case 'twitter:title':
 456                                                 $siteinfo['title'] = trim($meta_tag['content']);
 457                                                 break;
 458                                         case 'twitter:image':
 459                                                 $siteinfo['image'] = $meta_tag['content'];
 460                                                 $twitter_image = true;
 461                                                 break;
 462                                 }
 463                         }
 464                 }
 465
 466                 $list = $xpath->query("//script[@type='application/ld+json']");
 467                 foreach ($list as $node) {
 468                         if (!empty($node->nodeValue)) {
 469                                 if ($jsonld = json_decode($node->nodeValue, true)) {
 470                                         $siteinfo = self::parseParts($siteinfo, $jsonld);
 471                                 }
 472                         }
 473                 }
 474
 475 // Currently deactivated, see https://github.com/friendica/friendica/pull/10148#issuecomment-821512658
 476                 // Prevent to have a photo type without an image
 477 //              if ($twitter_card && $twitter_image && !empty($siteinfo['image'])) {
 478 //                      $siteinfo['type'] = 'photo';
 479 //              }
 480
 481                 if (!empty($siteinfo['image'])) {
 482                         $siteinfo['images'] = $siteinfo['images'] ?? [];
 483                         array_unshift($siteinfo['images'], ['url' => $siteinfo['image']]);
 484                         unset($siteinfo['image']);
 485                 }
 486
 487                 $siteinfo = self::checkMedia($url, $siteinfo);
 488
 489                 if (!empty($siteinfo['text']) && mb_strlen($siteinfo['text']) > self::MAX_DESC_COUNT) {
 490                         $siteinfo['text'] = mb_substr($siteinfo['text'], 0, self::MAX_DESC_COUNT) . '…';
 491                         $pos = mb_strrpos($siteinfo['text'], '.');
 492                         if ($pos > self::MIN_DESC_COUNT) {
 493                                 $siteinfo['text'] = mb_substr($siteinfo['text'], 0, $pos + 1);
 494                         }
 495                 }
 496
 497                 Logger::info('Siteinfo fetched', ['url' => $url, 'siteinfo' => $siteinfo]);
 498
 499                 Hook::callAll('getsiteinfo', $siteinfo);
 500
 501                 return $siteinfo;
 502         }
 503
 504         /**
 505          * Check the attached media elements.
 506          * Fix existing data and add missing data.
 507          *
 508          * @param string $page_url
 509          * @param array $siteinfo
 510          * @return void
 511          */
 512         private static function checkMedia(string $page_url, array $siteinfo)
 513         {
 514                 if (!empty($siteinfo['images'])) {
 515                         array_walk($siteinfo['images'], function (&$image) use ($page_url) {
 516                                 // According to the specifications someone could place a picture url into the content field as well.
 517                                 // But this doesn't seem to happen in the wild, so we don't cover it here.
 518                                 if (!empty($image['url'])) {
 519                                         $image['url'] = self::completeUrl($image['url'], $page_url);
 520                                         $photodata = Images::getInfoFromURLCached($image['url']);
 521                                         if (!empty($photodata) && ($photodata[0] > 50) && ($photodata[1] > 50)) {
 522                                                 $image['src'] = $image['url'];
 523                                                 $image['width'] = $photodata[0];
 524                                                 $image['height'] = $photodata[1];
 525                                                 $image['contenttype'] = $photodata['mime'];
 526                                                 unset($image['url']);
 527                                                 ksort($image);
 528                                         } else {
 529                                                 $image = [];
 530                                         }
 531                                 } else {
 532                                         $image = [];
 533                                 }
 534                         });
 535
 536                         $siteinfo['images'] = array_values(array_filter($siteinfo['images']));
 537                 }
 538
 539                 foreach (['audio', 'video'] as $element) {
 540                         if (!empty($siteinfo[$element])) {
 541                                 array_walk($siteinfo[$element], function (&$media) use ($page_url, &$siteinfo) {
 542                                         $url = '';
 543                                         $embed = '';
 544                                         $content = '';
 545                                         $contenttype = '';
 546                                         foreach (['embed', 'content', 'url'] as $field) {
 547                                                 if (!empty($media[$field])) {
 548                                                         $media[$field] = self::completeUrl($media[$field], $page_url);
 549                                                         $type = self::getContentType($media[$field]);
 550                                                         if (($type[0] ?? '') == 'text') {
 551                                                                 if ($field == 'embed') {
 552                                                                         $embed = $media[$field];
 553                                                                 } else {
 554                                                                         $url = $media[$field];
 555                                                                 }
 556                                                         } elseif (!empty($type[0])) {
 557                                                                 $content = $media[$field];
 558                                                                 $contenttype = implode('/', $type);
 559                                                         }
 560                                                 }
 561                                                 unset($media[$field]);
 562                                         }
 563
 564                                         foreach (['image', 'preview'] as $field) {
 565                                                 if (!empty($media[$field])) {
 566                                                         $media[$field] = self::completeUrl($media[$field], $page_url);
 567                                                 }
 568                                         }
 569
 570                                         if (!empty($url)) {
 571                                                 $media['url'] = $url;
 572                                         }
 573                                         if (!empty($embed)) {
 574                                                 $media['embed'] = $embed;
 575                                                 if (!empty($media['main'])) {
 576                                                         $siteinfo['embed'] = $embed;
 577                                                 }
 578                                         }
 579                                         if (!empty($content)) {
 580                                                 $media['src'] = $content;
 581                                         }
 582                                         if (!empty($contenttype)) {
 583                                                 $media['contenttype'] = $contenttype;
 584                                         }
 585                                         if (empty($url) && empty($content) && empty($embed)) {
 586                                                 $media = [];
 587                                         }
 588                                         ksort($media);
 589                                 });
 590
 591                                 $siteinfo[$element] = array_values(array_filter($siteinfo[$element]));
 592                         }
 593                         if (empty($siteinfo[$element])) {
 594                                 unset($siteinfo[$element]);
 595                         }
 596                 }
 597                 return $siteinfo;
 598         }
 599
 600         /**
 601          * Convert tags from CSV to an array
 602          *
 603          * @param string $string Tags
 604          * @return array with formatted Hashtags
 605          */
 606         public static function convertTagsToArray($string)
 607         {
 608                 $arr_tags = str_getcsv($string);
 609                 if (count($arr_tags)) {
 610                         // add the # sign to every tag
 611                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 612
 613                         return $arr_tags;
 614                 }
 615         }
 616
 617         /**
 618          * Add a hasht sign to a string
 619          *
 620          * This method is used as callback function
 621          *
 622          * @param string $tag The pure tag name
 623          * @param int    $k   Counter for internal use
 624          * @return void
 625          */
 626         private static function arrAddHashes(&$tag, $k)
 627         {
 628                 $tag = "#" . $tag;
 629         }
 630
 631         /**
 632          * Add a scheme to an url
 633          *
 634          * The src attribute of some html elements (e.g. images)
 635          * can miss the scheme so we need to add the correct
 636          * scheme
 637          *
 638          * @param string $url    The url which possibly does have
 639          *                       a missing scheme (a link to an image)
 640          * @param string $scheme The url with a correct scheme
 641          *                       (e.g. the url from the webpage which does contain the image)
 642          *
 643          * @return string The url with a scheme
 644          */
 645         private static function completeUrl($url, $scheme)
 646         {
 647                 $urlarr = parse_url($url);
 648
 649                 // If the url does allready have an scheme
 650                 // we can stop the process here
 651                 if (isset($urlarr["scheme"])) {
 652                         return($url);
 653                 }
 654
 655                 $schemearr = parse_url($scheme);
 656
 657                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 658
 659                 if (!empty($schemearr["port"])) {
 660                         $complete .= ":".$schemearr["port"];
 661                 }
 662
 663                 if (!empty($urlarr["path"])) {
 664                         if (strpos($urlarr["path"], "/") !== 0) {
 665                                 $complete .= "/";
 666                         }
 667
 668                         $complete .= $urlarr["path"];
 669                 }
 670
 671                 if (!empty($urlarr["query"])) {
 672                         $complete .= "?".$urlarr["query"];
 673                 }
 674
 675                 if (!empty($urlarr["fragment"])) {
 676                         $complete .= "#".$urlarr["fragment"];
 677                 }
 678
 679                 return($complete);
 680         }
 681
 682         /**
 683          * Parse the Json-Ld parts of a web page
 684          *
 685          * @param array $siteinfo
 686          * @param array $jsonld
 687          * @return array siteinfo
 688          */
 689         private static function parseParts(array $siteinfo, array $jsonld)
 690         {
 691                 if (!empty($jsonld['@graph']) && is_array($jsonld['@graph'])) {
 692                         foreach ($jsonld['@graph'] as $part) {
 693                                 if (!empty($part) && is_array($part)) {
 694                                         $siteinfo = self::parseParts($siteinfo, $part);
 695                                 }
 696                         }
 697                 } elseif (!empty($jsonld['@type'])) {
 698                         $siteinfo = self::parseJsonLd($siteinfo, $jsonld);
 699                 } elseif (!empty($jsonld)) {
 700                         $keys = array_keys($jsonld);
 701                         $numeric_keys = true;
 702                         foreach ($keys as $key) {
 703                                 if (!is_int($key)) {
 704                                         $numeric_keys = false;
 705                                 }
 706                         }
 707                         if ($numeric_keys) {
 708                                 foreach ($jsonld as $part) {
 709                                         if (!empty($part) && is_array($part)) {
 710                                                 $siteinfo = self::parseParts($siteinfo, $part);
 711                                         }
 712                                 }
 713                         }
 714                 }
 715
 716                 array_walk_recursive($siteinfo, function (&$element) {
 717                         if (is_string($element)) {
 718                                 $element = trim(strip_tags(html_entity_decode($element, ENT_COMPAT, 'UTF-8')));
 719                         }
 720                 });
 721
 722                 return $siteinfo;
 723         }
 724
 725         /**
 726          * Improve the siteinfo with information from the provided JSON-LD information
 727          * @see https://jsonld.com/
 728          * @see https://schema.org/
 729          *
 730          * @param array $siteinfo
 731          * @param array $jsonld
 732          * @return array siteinfo
 733          */
 734         private static function parseJsonLd(array $siteinfo, array $jsonld)
 735         {
 736                 $type = JsonLD::fetchElement($jsonld, '@type');
 737                 if (empty($type)) {
 738                         Logger::info('Empty type', ['url' => $siteinfo['url']]);
 739                         return $siteinfo;
 740                 }
 741
 742                 // Silently ignore some types that aren't processed
 743                 if (in_array($type, ['SiteNavigationElement', 'JobPosting', 'CreativeWork', 'MusicAlbum',
 744                         'WPHeader', 'WPSideBar', 'WPFooter', 'LegalService', 'MusicRecording',
 745                         'ItemList', 'BreadcrumbList', 'Blog', 'Dataset', 'Product'])) {
 746                         return $siteinfo;
 747                 }
 748
 749                 switch ($type) {
 750                         case 'Article':
 751                         case 'AdvertiserContentArticle':
 752                         case 'NewsArticle':
 753                         case 'Report':
 754                         case 'SatiricalArticle':
 755                         case 'ScholarlyArticle':
 756                         case 'SocialMediaPosting':
 757                         case 'TechArticle':
 758                         case 'ReportageNewsArticle':
 759                         case 'SocialMediaPosting':
 760                         case 'BlogPosting':
 761                         case 'LiveBlogPosting':
 762                         case 'DiscussionForumPosting':
 763                                 return self::parseJsonLdArticle($siteinfo, $jsonld);
 764                         case 'WebPage':
 765                         case 'AboutPage':
 766                         case 'CheckoutPage':
 767                         case 'CollectionPage':
 768                         case 'ContactPage':
 769                         case 'FAQPage':
 770                         case 'ItemPage':
 771                         case 'MedicalWebPage':
 772                         case 'ProfilePage':
 773                         case 'QAPage':
 774                         case 'RealEstateListing':
 775                         case 'SearchResultsPage':
 776                         case 'MediaGallery':
 777                         case 'ImageGallery':
 778                         case 'VideoGallery':
 779                         case 'RadioEpisode':
 780                         case 'Event':
 781                                 return self::parseJsonLdWebPage($siteinfo, $jsonld);
 782                         case 'WebSite':
 783                                 return self::parseJsonLdWebSite($siteinfo, $jsonld);
 784                         case 'Organization':
 785                         case 'Airline':
 786                         case 'Consortium':
 787                         case 'Corporation':
 788                         case 'EducationalOrganization':
 789                         case 'FundingScheme':
 790                         case 'GovernmentOrganization':
 791                         case 'LibrarySystem':
 792                         case 'LocalBusiness':
 793                         case 'MedicalOrganization':
 794                         case 'NGO':
 795                         case 'NewsMediaOrganization':
 796                         case 'Project':
 797                         case 'SportsOrganization':
 798                         case 'WorkersUnion':
 799                                 return self::parseJsonLdWebOrganization($siteinfo, $jsonld);
 800                         case 'Person':
 801                         case 'Patient':
 802                         case 'PerformingGroup':
 803                         case 'DanceGroup';
 804                         case 'MusicGroup':
 805                         case 'TheaterGroup':
 806                                 return self::parseJsonLdWebPerson($siteinfo, $jsonld);
 807                         case 'AudioObject':
 808                         case 'Audio':
 809                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'audio');
 810                         case 'VideoObject':
 811                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'video');
 812                         case 'ImageObject':
 813                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'images');
 814                         default:
 815                                 Logger::info('Unknown type', ['type' => $type, 'url' => $siteinfo['url']]);
 816                                 return $siteinfo;
 817                 }
 818         }
 819
 820         /**
 821          * Fetch author and publisher data
 822          *
 823          * @param array $siteinfo
 824          * @param array $jsonld
 825          * @return array siteinfo
 826          */
 827         private static function parseJsonLdAuthor(array $siteinfo, array $jsonld)
 828         {
 829                 $jsonldinfo = [];
 830
 831                 if (!empty($jsonld['publisher']) && is_array($jsonld['publisher'])) {
 832                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'name');
 833                         if (!empty($content) && is_string($content)) {
 834                                 $jsonldinfo['publisher_name'] = trim($content);
 835                         }
 836
 837                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'url');
 838                         if (!empty($content) && is_string($content)) {
 839                                 $jsonldinfo['publisher_url'] = trim($content);
 840                         }
 841
 842                         $brand = JsonLD::fetchElement($jsonld, 'publisher', 'brand', '@type', 'Organization');
 843                         if (!empty($brand) && is_array($brand)) {
 844                                 $content = JsonLD::fetchElement($brand, 'name');
 845                                 if (!empty($content) && is_string($content)) {
 846                                         $jsonldinfo['publisher_name'] = trim($content);
 847                                 }
 848
 849                                 $content = JsonLD::fetchElement($brand, 'url');
 850                                 if (!empty($content) && is_string($content)) {
 851                                         $jsonldinfo['publisher_url'] = trim($content);
 852                                 }
 853
 854                                 $content = JsonLD::fetchElement($brand, 'logo', 'url');
 855                                 if (!empty($content) && is_string($content)) {
 856                                         $jsonldinfo['publisher_img'] = trim($content);
 857                                 }
 858                         }
 859
 860                         $logo = JsonLD::fetchElement($jsonld, 'publisher', 'logo');
 861                         if (!empty($logo) && is_array($logo)) {
 862                                 $content = JsonLD::fetchElement($logo, 'url');
 863                                 if (!empty($content) && is_string($content)) {
 864                                         $jsonldinfo['publisher_img'] = trim($content);
 865                                 }
 866                         }
 867                 } elseif (!empty($jsonld['publisher']) && is_string($jsonld['publisher'])) {
 868                         $jsonldinfo['publisher_name'] = trim($jsonld['publisher']);
 869                 }
 870
 871                 if (!empty($jsonld['author']) && is_array($jsonld['author'])) {
 872                         $content = JsonLD::fetchElement($jsonld, 'author', 'name');
 873                         if (!empty($content) && is_string($content)) {
 874                                 $jsonldinfo['author_name'] = trim($content);
 875                         }
 876
 877                         $content = JsonLD::fetchElement($jsonld, 'author', 'sameAs');
 878                         if (!empty($content) && is_string($content)) {
 879                                 $jsonldinfo['author_url'] = trim($content);
 880                         }
 881
 882                         $content = JsonLD::fetchElement($jsonld, 'author', 'url');
 883                         if (!empty($content) && is_string($content)) {
 884                                 $jsonldinfo['author_url'] = trim($content);
 885                         }
 886
 887                         $logo = JsonLD::fetchElement($jsonld, 'author', 'logo');
 888                         if (!empty($logo) && is_array($logo)) {
 889                                 $content = JsonLD::fetchElement($logo, 'url');
 890                                 if (!empty($content) && is_string($content)) {
 891                                         $jsonldinfo['author_img'] = trim($content);
 892                                 }
 893                         }
 894                 } elseif (!empty($jsonld['author']) && is_string($jsonld['author'])) {
 895                         $jsonldinfo['author_name'] = trim($jsonld['author']);
 896                 }
 897
 898                 Logger::info('Fetched Author information', ['fetched' => $jsonldinfo]);
 899
 900                 return array_merge($siteinfo, $jsonldinfo);
 901         }
 902
 903         /**
 904          * Fetch data from the provided JSON-LD Article type
 905          * @see https://schema.org/Article
 906          *
 907          * @param array $siteinfo
 908          * @param array $jsonld
 909          * @return array siteinfo
 910          */
 911         private static function parseJsonLdArticle(array $siteinfo, array $jsonld)
 912         {
 913                 $jsonldinfo = [];
 914
 915                 $content = JsonLD::fetchElement($jsonld, 'headline');
 916                 if (!empty($content) && is_string($content)) {
 917                         $jsonldinfo['title'] = trim($content);
 918                 }
 919
 920                 $content = JsonLD::fetchElement($jsonld, 'alternativeHeadline');
 921                 if (!empty($content) && is_string($content) && (($jsonldinfo['title'] ?? '') != trim($content))) {
 922                         $jsonldinfo['alternative_title'] = trim($content);
 923                 }
 924
 925                 $content = JsonLD::fetchElement($jsonld, 'description');
 926                 if (!empty($content) && is_string($content)) {
 927                         $jsonldinfo['text'] = trim($content);
 928                 }
 929
 930                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 931                 if (!empty($content)) {
 932                         $jsonldinfo['image'] = trim($content);
 933                 }
 934
 935                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
 936                 if (!empty($content) && is_string($content)) {
 937                         $jsonldinfo['image'] = trim($content);
 938                 }
 939
 940                 if (!empty($jsonld['keywords']) && !is_array($jsonld['keywords'])) {
 941                         $content = JsonLD::fetchElement($jsonld, 'keywords');
 942                         if (!empty($content)) {
 943                                 $siteinfo['keywords'] = [];
 944                                 $keywords = explode(',', $content);
 945                                 foreach ($keywords as $keyword) {
 946                                         $siteinfo['keywords'][] = trim($keyword);
 947                                 }
 948                         }
 949                 } elseif (!empty($jsonld['keywords'])) {
 950                         $content = JsonLD::fetchElementArray($jsonld, 'keywords');
 951                         if (!empty($content) && is_array($content)) {
 952                                 $jsonldinfo['keywords'] = $content;
 953                         }
 954                 }
 955
 956                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 957
 958                 Logger::info('Fetched article information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 959
 960                 return array_merge($siteinfo, $jsonldinfo);
 961         }
 962
 963         /**
 964          * Fetch data from the provided JSON-LD WebPage type
 965          * @see https://schema.org/WebPage
 966          *
 967          * @param array $siteinfo
 968          * @param array $jsonld
 969          * @return array siteinfo
 970          */
 971         private static function parseJsonLdWebPage(array $siteinfo, array $jsonld)
 972         {
 973                 $jsonldinfo = [];
 974
 975                 $content = JsonLD::fetchElement($jsonld, 'name');
 976                 if (!empty($content)) {
 977                         $jsonldinfo['title'] = trim($content);
 978                 }
 979
 980                 $content = JsonLD::fetchElement($jsonld, 'description');
 981                 if (!empty($content)) {
 982                         $jsonldinfo['text'] = trim($content);
 983                 }
 984
 985                 $content = JsonLD::fetchElement($jsonld, 'image');
 986                 if (!empty($content)) {
 987                         $jsonldinfo['image'] = trim($content);
 988                 }
 989
 990                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 991                 if (!empty($content)) {
 992                         $jsonldinfo['image'] = trim($content);
 993                 }
 994
 995                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 996
 997                 Logger::info('Fetched WebPage information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 998
 999                 return array_merge($siteinfo, $jsonldinfo);
1000         }
1001
1002         /**
1003          * Fetch data from the provided JSON-LD WebSite type
1004          * @see https://schema.org/WebSite
1005          *
1006          * @param array $siteinfo
1007          * @param array $jsonld
1008          * @return array siteinfo
1009          */
1010         private static function parseJsonLdWebSite(array $siteinfo, array $jsonld)
1011         {
1012                 $jsonldinfo = [];
1013
1014                 $content = JsonLD::fetchElement($jsonld, 'name');
1015                 if (!empty($content)) {
1016                         $jsonldinfo['publisher_name'] = trim($content);
1017                 }
1018
1019                 $content = JsonLD::fetchElement($jsonld, 'description');
1020                 if (!empty($content)) {
1021                         $jsonldinfo['publisher_description'] = trim($content);
1022                 }
1023
1024                 $content = JsonLD::fetchElement($jsonld, 'url');
1025                 if (!empty($content)) {
1026                         $jsonldinfo['publisher_url'] = trim($content);
1027                 }
1028
1029                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
1030                 if (!empty($content)) {
1031                         $jsonldinfo['image'] = trim($content);
1032                 }
1033
1034                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
1035
1036                 Logger::info('Fetched WebSite information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1037                 return array_merge($siteinfo, $jsonldinfo);
1038         }
1039
1040         /**
1041          * Fetch data from the provided JSON-LD Organization type
1042          * @see https://schema.org/Organization
1043          *
1044          * @param array $siteinfo
1045          * @param array $jsonld
1046          * @return array siteinfo
1047          */
1048         private static function parseJsonLdWebOrganization(array $siteinfo, array $jsonld)
1049         {
1050                 $jsonldinfo = [];
1051
1052                 $content = JsonLD::fetchElement($jsonld, 'name');
1053                 if (!empty($content)) {
1054                         $jsonldinfo['publisher_name'] = trim($content);
1055                 }
1056
1057                 $content = JsonLD::fetchElement($jsonld, 'description');
1058                 if (!empty($content)) {
1059                         $jsonldinfo['publisher_description'] = trim($content);
1060                 }
1061
1062                 $content = JsonLD::fetchElement($jsonld, 'url');
1063                 if (!empty($content)) {
1064                         $jsonldinfo['publisher_url'] = trim($content);
1065                 }
1066
1067                 $content = JsonLD::fetchElement($jsonld, 'logo', 'url', '@type', 'ImageObject');
1068                 if (!empty($content)) {
1069                         $jsonldinfo['publisher_img'] = trim($content);
1070                 }
1071
1072                 $content = JsonLD::fetchElement($jsonld, 'brand', 'name', '@type', 'Organization');
1073                 if (!empty($content)) {
1074                         $jsonldinfo['publisher_name'] = trim($content);
1075                 }
1076
1077                 $content = JsonLD::fetchElement($jsonld, 'brand', 'url', '@type', 'Organization');
1078                 if (!empty($content)) {
1079                         $jsonldinfo['publisher_url'] = trim($content);
1080                 }
1081
1082                 Logger::info('Fetched Organization information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1083                 return array_merge($siteinfo, $jsonldinfo);
1084         }
1085
1086         /**
1087          * Fetch data from the provided JSON-LD Person type
1088          * @see https://schema.org/Person
1089          *
1090          * @param array $siteinfo
1091          * @param array $jsonld
1092          * @return array siteinfo
1093          */
1094         private static function parseJsonLdWebPerson(array $siteinfo, array $jsonld)
1095         {
1096                 $jsonldinfo = [];
1097
1098                 $content = JsonLD::fetchElement($jsonld, 'name');
1099                 if (!empty($content)) {
1100                         $jsonldinfo['author_name'] = trim($content);
1101                 }
1102
1103                 $content = JsonLD::fetchElement($jsonld, 'description');
1104                 if (!empty($content)) {
1105                         $jsonldinfo['author_description'] = trim($content);
1106                 }
1107
1108                 $content = JsonLD::fetchElement($jsonld, 'sameAs');
1109                 if (!empty($content) && is_string($content)) {
1110                         $jsonldinfo['author_url'] = trim($content);
1111                 }
1112
1113                 $content = JsonLD::fetchElement($jsonld, 'url');
1114                 if (!empty($content)) {
1115                         $jsonldinfo['author_url'] = trim($content);
1116                 }
1117
1118                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
1119                 if (!empty($content) && !is_string($content)) {
1120                         Logger::notice('Unexpected return value for the author image', ['content' => $content]);
1121                 }
1122
1123                 if (!empty($content) && is_string($content)) {
1124                         $jsonldinfo['author_img'] = trim($content);
1125                 }
1126
1127                 Logger::info('Fetched Person information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1128                 return array_merge($siteinfo, $jsonldinfo);
1129         }
1130
1131         /**
1132          * Fetch data from the provided JSON-LD MediaObject type
1133          * @see https://schema.org/MediaObject
1134          *
1135          * @param array $siteinfo
1136          * @param array $jsonld
1137          * @return array siteinfo
1138          */
1139         private static function parseJsonLdMediaObject(array $siteinfo, array $jsonld, string $name)
1140         {
1141                 $media = [];
1142
1143                 $content = JsonLD::fetchElement($jsonld, 'caption');
1144                 if (!empty($content)) {
1145                         $media['caption'] = trim($content);
1146                 }
1147
1148                 $content = JsonLD::fetchElement($jsonld, 'url');
1149                 if (!empty($content)) {
1150                         $media['url'] = trim($content);
1151                 }
1152
1153                 $content = JsonLD::fetchElement($jsonld, 'mainEntityOfPage');
1154                 if (!empty($content)) {
1155                         $media['main'] = Strings::compareLink($content, $siteinfo['url']);
1156                 }
1157
1158                 $content = JsonLD::fetchElement($jsonld, 'description');
1159                 if (!empty($content)) {
1160                         $media['description'] = trim($content);
1161                 }
1162
1163                 $content = JsonLD::fetchElement($jsonld, 'name');
1164                 if (!empty($content) && (($media['description'] ?? '') != trim($content))) {
1165                         $media['name'] = trim($content);
1166                 }
1167
1168                 $content = JsonLD::fetchElement($jsonld, 'contentUrl');
1169                 if (!empty($content)) {
1170                         $media['content'] = trim($content);
1171                 }
1172
1173                 $content = JsonLD::fetchElement($jsonld, 'embedUrl');
1174                 if (!empty($content)) {
1175                         $media['embed'] = trim($content);
1176                 }
1177
1178                 $content = JsonLD::fetchElement($jsonld, 'height');
1179                 if (!empty($content)) {
1180                         $media['height'] = trim($content);
1181                 }
1182
1183                 $content = JsonLD::fetchElement($jsonld, 'width');
1184                 if (!empty($content)) {
1185                         $media['width'] = trim($content);
1186                 }
1187
1188                 $content = JsonLD::fetchElement($jsonld, 'image');
1189                 if (!empty($content)) {
1190                         $media['image'] = trim($content);
1191                 }
1192
1193                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
1194                 if (!empty($content) && (($media['image'] ?? '') != trim($content))) {
1195                         if (!empty($media['image'])) {
1196                                 $media['preview'] = trim($content);
1197                         } else {
1198                                 $media['image'] = trim($content);
1199                         }
1200                 }
1201
1202                 Logger::info('Fetched Media information', ['url' => $siteinfo['url'], 'fetched' => $media]);
1203                 $siteinfo[$name][] = $media;
1204                 return $siteinfo;
1205         }
1206 }