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
  34 /**
  35  * Get information about a given URL
  36  *
  37  * Class with methods for extracting certain content from an url
  38  */
  39 class ParseUrl
  40 {
  41         const DEFAULT_EXPIRATION_FAILURE = 'now + 1 day';
  42         const DEFAULT_EXPIRATION_SUCCESS = 'now + 3 months';
  43
  44         /**
  45          * Maximum number of characters for the description
  46          */
  47         const MAX_DESC_COUNT = 250;
  48
  49         /**
  50          * Minimum number of characters for the description
  51          */
  52         const MIN_DESC_COUNT = 100;
  53
  54         /**
  55          * Fetch the content type of the given url
  56          * @param string $url URL of the page
  57          * @return array content type
  58          */
  59         public static function getContentType(string $url)
  60         {
  61                 $curlResult = DI::httpRequest()->head($url);
  62                 if (!$curlResult->isSuccess()) {
  63                         return [];
  64                 }
  65
  66                 $contenttype =  $curlResult->getHeader('Content-Type');
  67                 if (empty($contenttype)) {
  68                         return [];
  69                 }
  70
  71                 return explode('/', current(explode(';', $contenttype)));
  72         }
  73
  74         /**
  75          * Search for chached embeddable data of an url otherwise fetch it
  76          *
  77          * @param string $url         The url of the page which should be scraped
  78          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
  79          *                            to avoid endless loops
  80          *
  81          * @return array which contains needed data for embedding
  82          *    string 'url'      => The url of the parsed page
  83          *    string 'type'     => Content type
  84          *    string 'title'    => (optional) The title of the content
  85          *    string 'text'     => (optional) The description for the content
  86          *    string 'image'    => (optional) A preview image of the content
  87          *    array  'images'   => (optional) Array of preview pictures
  88          *    string 'keywords' => (optional) The tags which belong to the content
  89          *
  90          * @throws HTTPException\InternalServerErrorException
  91          * @see   ParseUrl::getSiteinfo() for more information about scraping
  92          * embeddable content
  93          */
  94         public static function getSiteinfoCached($url, $do_oembed = true): array
  95         {
  96                 if (empty($url)) {
  97                         return [
  98                                 'url' => '',
  99                                 'type' => 'error',
 100                         ];
 101                 }
 102
 103                 $urlHash = hash('sha256', $url);
 104
 105                 $parsed_url = DBA::selectFirst('parsed_url', ['content'],
 106                         ['url_hash' => $urlHash, 'oembed' => $do_oembed]
 107                 );
 108                 if (!empty($parsed_url['content'])) {
 109                         $data = unserialize($parsed_url['content']);
 110                         return $data;
 111                 }
 112
 113                 $data = self::getSiteinfo($url, $do_oembed);
 114
 115                 $expires = $data['expires'];
 116
 117                 unset($data['expires']);
 118
 119                 DI::dba()->insert(
 120                         'parsed_url',
 121                         [
 122                                 'url_hash' => $urlHash,
 123                                 'oembed'   => $do_oembed,
 124                                 'url'      => $url,
 125                                 'content'  => serialize($data),
 126                                 'created'  => DateTimeFormat::utcNow(),
 127                                 'expires'  => $expires,
 128                         ],
 129                         Database::INSERT_UPDATE
 130                 );
 131
 132                 return $data;
 133         }
 134
 135         /**
 136          * Parse a page for embeddable content information
 137          *
 138          * This method parses to url for meta data which can be used to embed
 139          * the content. If available it prioritizes Open Graph meta tags.
 140          * If this is not available it uses the twitter cards meta tags.
 141          * As fallback it uses standard html elements with meta informations
 142          * like \<title\>Awesome Title\</title\> or
 143          * \<meta name="description" content="An awesome description"\>
 144          *
 145          * @param string $url         The url of the page which should be scraped
 146          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
 147          *                            to avoid endless loops
 148          * @param int    $count       Internal counter to avoid endless loops
 149          *
 150          * @return array which contains needed data for embedding
 151          *    string 'url'      => The url of the parsed page
 152          *    string 'type'     => Content type (error, link, photo, image, audio, video)
 153          *    string 'title'    => (optional) The title of the content
 154          *    string 'text'     => (optional) The description for the content
 155          *    string 'image'    => (optional) A preview image of the content
 156          *    array  'images'   => (optional) Array of preview pictures
 157          *    string 'keywords' => (optional) The tags which belong to the content
 158          *
 159          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 160          * @todo  https://developers.google.com/+/plugins/snippet/
 161          * @verbatim
 162          * <meta itemprop="name" content="Awesome title">
 163          * <meta itemprop="description" content="An awesome description">
 164          * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
 165          *
 166          * <body itemscope itemtype="http://schema.org/Product">
 167          *   <h1 itemprop="name">Shiny Trinket</h1>
 168          *   <img itemprop="image" src="{image-url}" />
 169          *   <p itemprop="description">Shiny trinkets are shiny.</p>
 170          * </body>
 171          * @endverbatim
 172          */
 173         public static function getSiteinfo($url, $do_oembed = true, $count = 1)
 174         {
 175                 if (empty($url)) {
 176                         return [
 177                                 'url' => '',
 178                                 'type' => 'error',
 179                         ];
 180                 }
 181
 182                 // Check if the URL does contain a scheme
 183                 $scheme = parse_url($url, PHP_URL_SCHEME);
 184
 185                 if ($scheme == '') {
 186                         $url = 'http://' . ltrim($url, '/');
 187                 }
 188
 189                 $url = trim($url, "'\"");
 190
 191                 $url = Network::stripTrackingQueryParams($url);
 192
 193                 $siteinfo = [
 194                         'url' => $url,
 195                         'type' => 'link',
 196                         'expires' => DateTimeFormat::utc(self::DEFAULT_EXPIRATION_FAILURE),
 197                 ];
 198
 199                 if ($count > 10) {
 200                         Logger::log('Endless loop detected for ' . $url, Logger::DEBUG);
 201                         return $siteinfo;
 202                 }
 203
 204                 $type = self::getContentType($url);
 205                 Logger::info('Got content-type', ['content-type' => $type, 'url' => $url]);
 206                 if (!empty($type) && in_array($type[0], ['image', 'video', 'audio'])) {
 207                         $siteinfo['type'] = $type[0];
 208                         return $siteinfo;
 209                 }
 210
 211                 if ((count($type) >= 2) && (($type[0] != 'text') || ($type[1] != 'html'))) {
 212                         Logger::info('Unparseable content-type, quitting here, ', ['content-type' => $type, 'url' => $url]);
 213                         return $siteinfo;
 214                 }
 215
 216                 $curlResult = DI::httpRequest()->get($url);
 217                 if (!$curlResult->isSuccess() || empty($curlResult->getBody())) {
 218                         return $siteinfo;
 219                 }
 220
 221                 $siteinfo['expires'] = DateTimeFormat::utc(self::DEFAULT_EXPIRATION_SUCCESS);
 222
 223                 // If the file is too large then exit
 224                 if (($curlResult->getInfo()['download_content_length'] ?? 0) > 1000000) {
 225                         return $siteinfo;
 226                 }
 227
 228                 if ($cacheControlHeader = $curlResult->getHeader('Cache-Control')) {
 229                         if (preg_match('/max-age=([0-9]+)/i', $cacheControlHeader, $matches)) {
 230                                 $maxAge = max(86400, (int)array_pop($matches));
 231                                 $siteinfo['expires'] = DateTimeFormat::utc("now + $maxAge seconds");
 232                         }
 233                 }
 234
 235                 $header = $curlResult->getHeader();
 236                 $body = $curlResult->getBody();
 237
 238                 if ($do_oembed) {
 239                         $oembed_data = OEmbed::fetchURL($url, false, false);
 240
 241                         if (!empty($oembed_data->type)) {
 242                                 if (!in_array($oembed_data->type, ['error', 'rich', 'image', 'video', 'audio', ''])) {
 243                                         $siteinfo['type'] = $oembed_data->type;
 244                                 }
 245
 246                                 // See https://github.com/friendica/friendica/pull/5763#discussion_r217913178
 247                                 if ($siteinfo['type'] != 'photo') {
 248                                         if (!empty($oembed_data->title)) {
 249                                                 $siteinfo['title'] = trim($oembed_data->title);
 250                                         }
 251                                         if (!empty($oembed_data->description)) {
 252                                                 $siteinfo['text'] = trim($oembed_data->description);
 253                                         }
 254                                         if (!empty($oembed_data->author_name)) {
 255                                                 $siteinfo['author_name'] = trim($oembed_data->author_name);
 256                                         }
 257                                         if (!empty($oembed_data->author_url)) {
 258                                                 $siteinfo['author_url'] = trim($oembed_data->author_url);
 259                                         }
 260                                         if (!empty($oembed_data->provider_name)) {
 261                                                 $siteinfo['publisher_name'] = trim($oembed_data->provider_name);
 262                                         }
 263                                         if (!empty($oembed_data->provider_url)) {
 264                                                 $siteinfo['publisher_url'] = trim($oembed_data->provider_url);
 265                                         }
 266                                         if (!empty($oembed_data->thumbnail_url)) {
 267                                                 $siteinfo['image'] = $oembed_data->thumbnail_url;
 268                                         }
 269                                 }
 270                         }
 271                 }
 272
 273                 $charset = '';
 274                 // Look for a charset, first in headers
 275                 // Expected form: Content-Type: text/html; charset=ISO-8859-4
 276                 if (preg_match('/charset=([a-z0-9-_.\/]+)/i', $header, $matches)) {
 277                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 278                 }
 279
 280                 // Then in body that gets precedence
 281                 // Expected forms:
 282                 // - <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
 283                 // - <meta charset="utf-8">
 284                 // - <meta charset=utf-8>
 285                 // - <meta charSet="utf-8">
 286                 // We escape <style> and <script> tags since they can contain irrelevant charset information
 287                 // (see https://github.com/friendica/friendica/issues/9251#issuecomment-698636806)
 288                 Strings::performWithEscapedBlocks($body, '#<(?:style|script).*?</(?:style|script)>#ism', function ($body) use (&$charset) {
 289                         if (preg_match('/charset=["\']?([a-z0-9-_.\/]+)/i', $body, $matches)) {
 290                                 $charset = trim(trim(trim(array_pop($matches)), ';,'));
 291                         }
 292                 });
 293
 294                 $siteinfo['charset'] = $charset;
 295
 296                 if ($charset && strtoupper($charset) != 'UTF-8') {
 297                         // See https://github.com/friendica/friendica/issues/5470#issuecomment-418351211
 298                         $charset = str_ireplace('latin-1', 'latin1', $charset);
 299
 300                         Logger::log('detected charset ' . $charset, Logger::DEBUG);
 301                         $body = iconv($charset, 'UTF-8//TRANSLIT', $body);
 302                 }
 303
 304                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', 'UTF-8');
 305
 306                 $doc = new DOMDocument();
 307                 @$doc->loadHTML($body);
 308
 309                 XML::deleteNode($doc, 'style');
 310                 XML::deleteNode($doc, 'option');
 311                 XML::deleteNode($doc, 'h1');
 312                 XML::deleteNode($doc, 'h2');
 313                 XML::deleteNode($doc, 'h3');
 314                 XML::deleteNode($doc, 'h4');
 315                 XML::deleteNode($doc, 'h5');
 316                 XML::deleteNode($doc, 'h6');
 317                 XML::deleteNode($doc, 'ol');
 318                 XML::deleteNode($doc, 'ul');
 319
 320                 $xpath = new DOMXPath($doc);
 321
 322                 $list = $xpath->query('//meta[@content]');
 323                 foreach ($list as $node) {
 324                         $meta_tag = [];
 325                         if ($node->attributes->length) {
 326                                 foreach ($node->attributes as $attribute) {
 327                                         $meta_tag[$attribute->name] = $attribute->value;
 328                                 }
 329                         }
 330
 331                         if (@$meta_tag['http-equiv'] == 'refresh') {
 332                                 $path = $meta_tag['content'];
 333                                 $pathinfo = explode(';', $path);
 334                                 $content = '';
 335                                 foreach ($pathinfo as $value) {
 336                                         if (substr(strtolower($value), 0, 4) == 'url=') {
 337                                                 $content = substr($value, 4);
 338                                         }
 339                                 }
 340                                 if ($content != '') {
 341                                         $siteinfo = self::getSiteinfo($content, $do_oembed, ++$count);
 342                                         return $siteinfo;
 343                                 }
 344                         }
 345                 }
 346
 347                 $list = $xpath->query('//title');
 348                 if ($list->length > 0) {
 349                         $siteinfo['title'] = trim($list->item(0)->nodeValue);
 350                 }
 351
 352                 $twitter_card = false;
 353                 $twitter_image = false;
 354
 355                 $list = $xpath->query('//meta[@name]');
 356                 foreach ($list as $node) {
 357                         $meta_tag = [];
 358                         if ($node->attributes->length) {
 359                                 foreach ($node->attributes as $attribute) {
 360                                         $meta_tag[$attribute->name] = $attribute->value;
 361                                 }
 362                         }
 363
 364                         if (empty($meta_tag['content'])) {
 365                                 continue;
 366                         }
 367
 368                         $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 369
 370                         switch (strtolower($meta_tag['name'])) {
 371                                 case 'fulltitle':
 372                                         $siteinfo['title'] = trim($meta_tag['content']);
 373                                         break;
 374                                 case 'description':
 375                                         $siteinfo['text'] = trim($meta_tag['content']);
 376                                         break;
 377                                 case 'thumbnail':
 378                                         $siteinfo['image'] = $meta_tag['content'];
 379                                         break;
 380                                 case 'twitter:image':
 381                                         $siteinfo['image'] = $meta_tag['content'];
 382                                         $twitter_image = true;
 383                                         break;
 384                                 case 'twitter:image:src':
 385                                         $siteinfo['image'] = $meta_tag['content'];
 386                                         break;
 387                                 case 'twitter:card':
 388                                         // Detect photo pages
 389                                         if ($meta_tag['content'] == 'summary_large_image') {
 390                                                 $twitter_card = true;
 391                                         }
 392                                         break;
 393                                 case 'twitter:description':
 394                                         $siteinfo['text'] = trim($meta_tag['content']);
 395                                         break;
 396                                 case 'twitter:title':
 397                                         $siteinfo['title'] = trim($meta_tag['content']);
 398                                         break;
 399                                 case 'dc.title':
 400                                         $siteinfo['title'] = trim($meta_tag['content']);
 401                                         break;
 402                                 case 'dc.description':
 403                                         $siteinfo['text'] = trim($meta_tag['content']);
 404                                         break;
 405                                 case 'dc.creator':
 406                                         $siteinfo['publisher_name'] = trim($meta_tag['content']);
 407                                         break;
 408                                 case 'keywords':
 409                                         $keywords = explode(',', $meta_tag['content']);
 410                                         break;
 411                                 case 'news_keywords':
 412                                         $keywords = explode(',', $meta_tag['content']);
 413                                         break;
 414                         }
 415                 }
 416
 417                 if (isset($keywords)) {
 418                         $siteinfo['keywords'] = [];
 419                         foreach ($keywords as $keyword) {
 420                                 if (!in_array(trim($keyword), $siteinfo['keywords'])) {
 421                                         $siteinfo['keywords'][] = trim($keyword);
 422                                 }
 423                         }
 424                 }
 425
 426                 $list = $xpath->query('//meta[@property]');
 427                 foreach ($list as $node) {
 428                         $meta_tag = [];
 429                         if ($node->attributes->length) {
 430                                 foreach ($node->attributes as $attribute) {
 431                                         $meta_tag[$attribute->name] = $attribute->value;
 432                                 }
 433                         }
 434
 435                         if (!empty($meta_tag['content'])) {
 436                                 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 437
 438                                 switch (strtolower($meta_tag['property'])) {
 439                                         case 'og:image':
 440                                                 $siteinfo['image'] = $meta_tag['content'];
 441                                                 break;
 442                                         case 'og:image:url':
 443                                                 $siteinfo['image'] = $meta_tag['content'];
 444                                                 break;
 445                                         case 'og:image:secure_url':
 446                                                 $siteinfo['image'] = $meta_tag['content'];
 447                                                 break;
 448                                         case 'og:title':
 449                                                 $siteinfo['title'] = trim($meta_tag['content']);
 450                                                 break;
 451                                         case 'og:description':
 452                                                 $siteinfo['text'] = trim($meta_tag['content']);
 453                                                 break;
 454                                         case 'og:site_name':
 455                                                 $siteinfo['publisher_name'] = trim($meta_tag['content']);
 456                                                 break;
 457                                         case 'twitter:description':
 458                                                 $siteinfo['text'] = trim($meta_tag['content']);
 459                                                 break;
 460                                         case 'twitter:title':
 461                                                 $siteinfo['title'] = trim($meta_tag['content']);
 462                                                 break;
 463                                         case 'twitter:image':
 464                                                 $siteinfo['image'] = $meta_tag['content'];
 465                                                 $twitter_image = true;
 466                                                 break;
 467                                 }
 468                         }
 469                 }
 470
 471                 $list = $xpath->query("//script[@type='application/ld+json']");
 472                 foreach ($list as $node) {
 473                         if (!empty($node->nodeValue)) {
 474                                 if ($jsonld = json_decode($node->nodeValue, true)) {
 475                                         $siteinfo = self::parseParts($siteinfo, $jsonld);
 476                                 }
 477                         }
 478                 }
 479
 480 // Currently deactivated, see https://github.com/friendica/friendica/pull/10148#issuecomment-821512658
 481                 // Prevent to have a photo type without an image
 482 //              if ($twitter_card && $twitter_image && !empty($siteinfo['image'])) {
 483 //                      $siteinfo['type'] = 'photo';
 484 //              }
 485
 486                 if (!empty($siteinfo['image'])) {
 487                         $siteinfo['images'] = $siteinfo['images'] ?? [];
 488                         array_unshift($siteinfo['images'], ['url' => $siteinfo['image']]);
 489                         unset($siteinfo['image']);
 490                 }
 491
 492                 $siteinfo = self::checkMedia($url, $siteinfo);
 493
 494                 if (!empty($siteinfo['text']) && mb_strlen($siteinfo['text']) > self::MAX_DESC_COUNT) {
 495                         $siteinfo['text'] = mb_substr($siteinfo['text'], 0, self::MAX_DESC_COUNT) . '…';
 496                         $pos = mb_strrpos($siteinfo['text'], '.');
 497                         if ($pos > self::MIN_DESC_COUNT) {
 498                                 $siteinfo['text'] = mb_substr($siteinfo['text'], 0, $pos + 1);
 499                         }
 500                 }
 501
 502                 Logger::info('Siteinfo fetched', ['url' => $url, 'siteinfo' => $siteinfo]);
 503
 504                 Hook::callAll('getsiteinfo', $siteinfo);
 505
 506                 return $siteinfo;
 507         }
 508
 509         /**
 510          * Check the attached media elements.
 511          * Fix existing data and add missing data.
 512          *
 513          * @param string $page_url
 514          * @param array $siteinfo
 515          * @return void
 516          */
 517         private static function checkMedia(string $page_url, array $siteinfo)
 518         {
 519                 if (!empty($siteinfo['images'])) {
 520                         array_walk($siteinfo['images'], function (&$image) use ($page_url) {
 521                                 // According to the specifications someone could place a picture url into the content field as well.
 522                                 // But this doesn't seem to happen in the wild, so we don't cover it here.
 523                                 if (!empty($image['url'])) {
 524                                         $image['url'] = self::completeUrl($image['url'], $page_url);
 525                                         $photodata = Images::getInfoFromURLCached($image['url']);
 526                                         if (!empty($photodata) && ($photodata[0] > 50) && ($photodata[1] > 50)) {
 527                                                 $image['src'] = $image['url'];
 528                                                 $image['width'] = $photodata[0];
 529                                                 $image['height'] = $photodata[1];
 530                                                 $image['contenttype'] = $photodata['mime'];
 531                                                 unset($image['url']);
 532                                                 ksort($image);
 533                                         } else {
 534                                                 $image = [];
 535                                         }
 536                                 } else {
 537                                         $image = [];
 538                                 }
 539                         });
 540
 541                         $siteinfo['images'] = array_values(array_filter($siteinfo['images']));
 542                 }
 543
 544                 foreach (['audio', 'video'] as $element) {
 545                         if (!empty($siteinfo[$element])) {
 546                                 array_walk($siteinfo[$element], function (&$media) use ($page_url, &$siteinfo) {
 547                                         $url = '';
 548                                         $embed = '';
 549                                         $content = '';
 550                                         $contenttype = '';
 551                                         foreach (['embed', 'content', 'url'] as $field) {
 552                                                 if (!empty($media[$field])) {
 553                                                         $media[$field] = self::completeUrl($media[$field], $page_url);
 554                                                         $type = self::getContentType($media[$field]);
 555                                                         if (($type[0] ?? '') == 'text') {
 556                                                                 if ($field == 'embed') {
 557                                                                         $embed = $media[$field];
 558                                                                 } else {
 559                                                                         $url = $media[$field];
 560                                                                 }
 561                                                         } elseif (!empty($type[0])) {
 562                                                                 $content = $media[$field];
 563                                                                 $contenttype = implode('/', $type);
 564                                                         }
 565                                                 }
 566                                                 unset($media[$field]);
 567                                         }
 568
 569                                         foreach (['image', 'preview'] as $field) {
 570                                                 if (!empty($media[$field])) {
 571                                                         $media[$field] = self::completeUrl($media[$field], $page_url);
 572                                                 }
 573                                         }
 574
 575                                         if (!empty($url)) {
 576                                                 $media['url'] = $url;
 577                                         }
 578                                         if (!empty($embed)) {
 579                                                 $media['embed'] = $embed;
 580                                                 if (!empty($media['main'])) {
 581                                                         $siteinfo['embed'] = $embed;
 582                                                 }
 583                                         }
 584                                         if (!empty($content)) {
 585                                                 $media['src'] = $content;
 586                                         }
 587                                         if (!empty($contenttype)) {
 588                                                 $media['contenttype'] = $contenttype;
 589                                         }
 590                                         if (empty($url) && empty($content) && empty($embed)) {
 591                                                 $media = [];
 592                                         }
 593                                         ksort($media);
 594                                 });
 595
 596                                 $siteinfo[$element] = array_values(array_filter($siteinfo[$element]));
 597                         }
 598                         if (empty($siteinfo[$element])) {
 599                                 unset($siteinfo[$element]);
 600                         }
 601                 }
 602                 return $siteinfo;
 603         }
 604
 605         /**
 606          * Convert tags from CSV to an array
 607          *
 608          * @param string $string Tags
 609          * @return array with formatted Hashtags
 610          */
 611         public static function convertTagsToArray($string)
 612         {
 613                 $arr_tags = str_getcsv($string);
 614                 if (count($arr_tags)) {
 615                         // add the # sign to every tag
 616                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 617
 618                         return $arr_tags;
 619                 }
 620         }
 621
 622         /**
 623          * Add a hasht sign to a string
 624          *
 625          * This method is used as callback function
 626          *
 627          * @param string $tag The pure tag name
 628          * @param int    $k   Counter for internal use
 629          * @return void
 630          */
 631         private static function arrAddHashes(&$tag, $k)
 632         {
 633                 $tag = "#" . $tag;
 634         }
 635
 636         /**
 637          * Add a scheme to an url
 638          *
 639          * The src attribute of some html elements (e.g. images)
 640          * can miss the scheme so we need to add the correct
 641          * scheme
 642          *
 643          * @param string $url    The url which possibly does have
 644          *                       a missing scheme (a link to an image)
 645          * @param string $scheme The url with a correct scheme
 646          *                       (e.g. the url from the webpage which does contain the image)
 647          *
 648          * @return string The url with a scheme
 649          */
 650         private static function completeUrl($url, $scheme)
 651         {
 652                 $urlarr = parse_url($url);
 653
 654                 // If the url does allready have an scheme
 655                 // we can stop the process here
 656                 if (isset($urlarr["scheme"])) {
 657                         return($url);
 658                 }
 659
 660                 $schemearr = parse_url($scheme);
 661
 662                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 663
 664                 if (!empty($schemearr["port"])) {
 665                         $complete .= ":".$schemearr["port"];
 666                 }
 667
 668                 if (!empty($urlarr["path"])) {
 669                         if (strpos($urlarr["path"], "/") !== 0) {
 670                                 $complete .= "/";
 671                         }
 672
 673                         $complete .= $urlarr["path"];
 674                 }
 675
 676                 if (!empty($urlarr["query"])) {
 677                         $complete .= "?".$urlarr["query"];
 678                 }
 679
 680                 if (!empty($urlarr["fragment"])) {
 681                         $complete .= "#".$urlarr["fragment"];
 682                 }
 683
 684                 return($complete);
 685         }
 686
 687         /**
 688          * Parse the Json-Ld parts of a web page
 689          *
 690          * @param array $siteinfo
 691          * @param array $jsonld
 692          * @return array siteinfo
 693          */
 694         private static function parseParts(array $siteinfo, array $jsonld)
 695         {
 696                 if (!empty($jsonld['@graph']) && is_array($jsonld['@graph'])) {
 697                         foreach ($jsonld['@graph'] as $part) {
 698                                 if (!empty($part)) {
 699                                         $siteinfo = self::parseParts($siteinfo, $part);
 700                                 }
 701                         }
 702                 } elseif (!empty($jsonld['@type'])) {
 703                         $siteinfo = self::parseJsonLd($siteinfo, $jsonld);
 704                 } elseif (!empty($jsonld)) {
 705                         $keys = array_keys($jsonld);
 706                         $numeric_keys = true;
 707                         foreach ($keys as $key) {
 708                                 if (!is_int($key)) {
 709                                         $numeric_keys = false;
 710                                 }
 711                         }
 712                         if ($numeric_keys) {
 713                                 foreach ($jsonld as $part) {
 714                                         if (!empty($part)) {
 715                                                 $siteinfo = self::parseParts($siteinfo, $part);
 716                                         }
 717                                 }
 718                         }
 719                 }
 720
 721                 array_walk_recursive($siteinfo, function (&$element) {
 722                         if (is_string($element)) {
 723                                 $element = trim(strip_tags(html_entity_decode($element, ENT_COMPAT, 'UTF-8')));
 724                         }
 725                 });
 726
 727                 return $siteinfo;
 728         }
 729
 730         /**
 731          * Improve the siteinfo with information from the provided JSON-LD information
 732          * @see https://jsonld.com/
 733          * @see https://schema.org/
 734          *
 735          * @param array $siteinfo
 736          * @param array $jsonld
 737          * @return array siteinfo
 738          */
 739         private static function parseJsonLd(array $siteinfo, array $jsonld)
 740         {
 741                 $type = JsonLD::fetchElement($jsonld, '@type');
 742                 if (empty($type)) {
 743                         Logger::info('Empty type', ['url' => $siteinfo['url']]);
 744                         return $siteinfo;
 745                 }
 746
 747                 // Silently ignore some types that aren't processed
 748                 if (in_array($type, ['SiteNavigationElement', 'JobPosting', 'CreativeWork', 'MusicAlbum',
 749                         'WPHeader', 'WPSideBar', 'WPFooter', 'LegalService', 'MusicRecording',
 750                         'ItemList', 'BreadcrumbList', 'Blog', 'Dataset', 'Product'])) {
 751                         return $siteinfo;
 752                 }
 753
 754                 switch ($type) {
 755                         case 'Article':
 756                         case 'AdvertiserContentArticle':
 757                         case 'NewsArticle':
 758                         case 'Report':
 759                         case 'SatiricalArticle':
 760                         case 'ScholarlyArticle':
 761                         case 'SocialMediaPosting':
 762                         case 'TechArticle':
 763                         case 'ReportageNewsArticle':
 764                         case 'SocialMediaPosting':
 765                         case 'BlogPosting':
 766                         case 'LiveBlogPosting':
 767                         case 'DiscussionForumPosting':
 768                                 return self::parseJsonLdArticle($siteinfo, $jsonld);
 769                         case 'WebPage':
 770                         case 'AboutPage':
 771                         case 'CheckoutPage':
 772                         case 'CollectionPage':
 773                         case 'ContactPage':
 774                         case 'FAQPage':
 775                         case 'ItemPage':
 776                         case 'MedicalWebPage':
 777                         case 'ProfilePage':
 778                         case 'QAPage':
 779                         case 'RealEstateListing':
 780                         case 'SearchResultsPage':
 781                         case 'MediaGallery':
 782                         case 'ImageGallery':
 783                         case 'VideoGallery':
 784                         case 'RadioEpisode':
 785                         case 'Event':
 786                                 return self::parseJsonLdWebPage($siteinfo, $jsonld);
 787                         case 'WebSite':
 788                                 return self::parseJsonLdWebSite($siteinfo, $jsonld);
 789                         case 'Organization':
 790                         case 'Airline':
 791                         case 'Consortium':
 792                         case 'Corporation':
 793                         case 'EducationalOrganization':
 794                         case 'FundingScheme':
 795                         case 'GovernmentOrganization':
 796                         case 'LibrarySystem':
 797                         case 'LocalBusiness':
 798                         case 'MedicalOrganization':
 799                         case 'NGO':
 800                         case 'NewsMediaOrganization':
 801                         case 'Project':
 802                         case 'SportsOrganization':
 803                         case 'WorkersUnion':
 804                                 return self::parseJsonLdWebOrganization($siteinfo, $jsonld);
 805                         case 'Person':
 806                         case 'Patient':
 807                         case 'PerformingGroup':
 808                         case 'DanceGroup';
 809                         case 'MusicGroup':
 810                         case 'TheaterGroup':
 811                                 return self::parseJsonLdWebPerson($siteinfo, $jsonld);
 812                         case 'AudioObject':
 813                         case 'Audio':
 814                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'audio');
 815                         case 'VideoObject':
 816                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'video');
 817                         case 'ImageObject':
 818                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'images');
 819                         default:
 820                                 Logger::info('Unknown type', ['type' => $type, 'url' => $siteinfo['url']]);
 821                                 return $siteinfo;
 822                 }
 823         }
 824
 825         /**
 826          * Fetch author and publisher data
 827          *
 828          * @param array $siteinfo
 829          * @param array $jsonld
 830          * @return array siteinfo
 831          */
 832         private static function parseJsonLdAuthor(array $siteinfo, array $jsonld)
 833         {
 834                 $jsonldinfo = [];
 835
 836                 if (!empty($jsonld['publisher']) && is_array($jsonld['publisher'])) {
 837                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'name');
 838                         if (!empty($content) && is_string($content)) {
 839                                 $jsonldinfo['publisher_name'] = trim($content);
 840                         }
 841
 842                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'url');
 843                         if (!empty($content) && is_string($content)) {
 844                                 $jsonldinfo['publisher_url'] = trim($content);
 845                         }
 846
 847                         $brand = JsonLD::fetchElement($jsonld, 'publisher', 'brand', '@type', 'Organization');
 848                         if (!empty($brand) && is_array($brand)) {
 849                                 $content = JsonLD::fetchElement($brand, 'name');
 850                                 if (!empty($content) && is_string($content)) {
 851                                         $jsonldinfo['publisher_name'] = trim($content);
 852                                 }
 853
 854                                 $content = JsonLD::fetchElement($brand, 'url');
 855                                 if (!empty($content) && is_string($content)) {
 856                                         $jsonldinfo['publisher_url'] = trim($content);
 857                                 }
 858
 859                                 $content = JsonLD::fetchElement($brand, 'logo', 'url');
 860                                 if (!empty($content) && is_string($content)) {
 861                                         $jsonldinfo['publisher_img'] = trim($content);
 862                                 }
 863                         }
 864
 865                         $logo = JsonLD::fetchElement($jsonld, 'publisher', 'logo');
 866                         if (!empty($logo) && is_array($logo)) {
 867                                 $content = JsonLD::fetchElement($logo, 'url');
 868                                 if (!empty($content) && is_string($content)) {
 869                                         $jsonldinfo['publisher_img'] = trim($content);
 870                                 }
 871                         }
 872                 } elseif (!empty($jsonld['publisher']) && is_string($jsonld['publisher'])) {
 873                         $jsonldinfo['publisher_name'] = trim($jsonld['publisher']);
 874                 }
 875
 876                 if (!empty($jsonld['author']) && is_array($jsonld['author'])) {
 877                         $content = JsonLD::fetchElement($jsonld, 'author', 'name');
 878                         if (!empty($content) && is_string($content)) {
 879                                 $jsonldinfo['author_name'] = trim($content);
 880                         }
 881
 882                         $content = JsonLD::fetchElement($jsonld, 'author', 'sameAs');
 883                         if (!empty($content) && is_string($content)) {
 884                                 $jsonldinfo['author_url'] = trim($content);
 885                         }
 886
 887                         $content = JsonLD::fetchElement($jsonld, 'author', 'url');
 888                         if (!empty($content) && is_string($content)) {
 889                                 $jsonldinfo['author_url'] = trim($content);
 890                         }
 891
 892                         $logo = JsonLD::fetchElement($jsonld, 'author', 'logo');
 893                         if (!empty($logo) && is_array($logo)) {
 894                                 $content = JsonLD::fetchElement($logo, 'url');
 895                                 if (!empty($content) && is_string($content)) {
 896                                         $jsonldinfo['author_img'] = trim($content);
 897                                 }
 898                         }
 899                 } elseif (!empty($jsonld['author']) && is_string($jsonld['author'])) {
 900                         $jsonldinfo['author_name'] = trim($jsonld['author']);
 901                 }
 902
 903                 Logger::info('Fetched Author information', ['fetched' => $jsonldinfo]);
 904
 905                 return array_merge($siteinfo, $jsonldinfo);
 906         }
 907
 908         /**
 909          * Fetch data from the provided JSON-LD Article type
 910          * @see https://schema.org/Article
 911          *
 912          * @param array $siteinfo
 913          * @param array $jsonld
 914          * @return array siteinfo
 915          */
 916         private static function parseJsonLdArticle(array $siteinfo, array $jsonld)
 917         {
 918                 $jsonldinfo = [];
 919
 920                 $content = JsonLD::fetchElement($jsonld, 'headline');
 921                 if (!empty($content) && is_string($content)) {
 922                         $jsonldinfo['title'] = trim($content);
 923                 }
 924
 925                 $content = JsonLD::fetchElement($jsonld, 'alternativeHeadline');
 926                 if (!empty($content) && is_string($content) && (($jsonldinfo['title'] ?? '') != trim($content))) {
 927                         $jsonldinfo['alternative_title'] = trim($content);
 928                 }
 929
 930                 $content = JsonLD::fetchElement($jsonld, 'description');
 931                 if (!empty($content) && is_string($content)) {
 932                         $jsonldinfo['text'] = trim($content);
 933                 }
 934
 935                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 936                 if (!empty($content)) {
 937                         $jsonldinfo['image'] = trim($content);
 938                 }
 939
 940                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
 941                 if (!empty($content) && is_string($content)) {
 942                         $jsonldinfo['image'] = trim($content);
 943                 }
 944
 945                 if (!empty($jsonld['keywords']) && !is_array($jsonld['keywords'])) {
 946                         $content = JsonLD::fetchElement($jsonld, 'keywords');
 947                         if (!empty($content)) {
 948                                 $siteinfo['keywords'] = [];
 949                                 $keywords = explode(',', $content);
 950                                 foreach ($keywords as $keyword) {
 951                                         $siteinfo['keywords'][] = trim($keyword);
 952                                 }
 953                         }
 954                 } elseif (!empty($jsonld['keywords'])) {
 955                         $content = JsonLD::fetchElementArray($jsonld, 'keywords');
 956                         if (!empty($content) && is_array($content)) {
 957                                 $jsonldinfo['keywords'] = $content;
 958                         }
 959                 }
 960
 961                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 962
 963                 Logger::info('Fetched article information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 964
 965                 return array_merge($siteinfo, $jsonldinfo);
 966         }
 967
 968         /**
 969          * Fetch data from the provided JSON-LD WebPage type
 970          * @see https://schema.org/WebPage
 971          *
 972          * @param array $siteinfo
 973          * @param array $jsonld
 974          * @return array siteinfo
 975          */
 976         private static function parseJsonLdWebPage(array $siteinfo, array $jsonld)
 977         {
 978                 $jsonldinfo = [];
 979
 980                 $content = JsonLD::fetchElement($jsonld, 'name');
 981                 if (!empty($content)) {
 982                         $jsonldinfo['title'] = trim($content);
 983                 }
 984
 985                 $content = JsonLD::fetchElement($jsonld, 'description');
 986                 if (!empty($content)) {
 987                         $jsonldinfo['text'] = trim($content);
 988                 }
 989
 990                 $content = JsonLD::fetchElement($jsonld, 'image');
 991                 if (!empty($content)) {
 992                         $jsonldinfo['image'] = trim($content);
 993                 }
 994
 995                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 996                 if (!empty($content)) {
 997                         $jsonldinfo['image'] = trim($content);
 998                 }
 999
1000                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
1001
1002                 Logger::info('Fetched WebPage information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1003
1004                 return array_merge($siteinfo, $jsonldinfo);
1005         }
1006
1007         /**
1008          * Fetch data from the provided JSON-LD WebSite type
1009          * @see https://schema.org/WebSite
1010          *
1011          * @param array $siteinfo
1012          * @param array $jsonld
1013          * @return array siteinfo
1014          */
1015         private static function parseJsonLdWebSite(array $siteinfo, array $jsonld)
1016         {
1017                 $jsonldinfo = [];
1018
1019                 $content = JsonLD::fetchElement($jsonld, 'name');
1020                 if (!empty($content)) {
1021                         $jsonldinfo['publisher_name'] = trim($content);
1022                 }
1023
1024                 $content = JsonLD::fetchElement($jsonld, 'description');
1025                 if (!empty($content)) {
1026                         $jsonldinfo['publisher_description'] = trim($content);
1027                 }
1028
1029                 $content = JsonLD::fetchElement($jsonld, 'url');
1030                 if (!empty($content)) {
1031                         $jsonldinfo['publisher_url'] = trim($content);
1032                 }
1033
1034                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
1035                 if (!empty($content)) {
1036                         $jsonldinfo['image'] = trim($content);
1037                 }
1038
1039                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
1040
1041                 Logger::info('Fetched WebSite information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1042                 return array_merge($siteinfo, $jsonldinfo);
1043         }
1044
1045         /**
1046          * Fetch data from the provided JSON-LD Organization type
1047          * @see https://schema.org/Organization
1048          *
1049          * @param array $siteinfo
1050          * @param array $jsonld
1051          * @return array siteinfo
1052          */
1053         private static function parseJsonLdWebOrganization(array $siteinfo, array $jsonld)
1054         {
1055                 $jsonldinfo = [];
1056
1057                 $content = JsonLD::fetchElement($jsonld, 'name');
1058                 if (!empty($content)) {
1059                         $jsonldinfo['publisher_name'] = trim($content);
1060                 }
1061
1062                 $content = JsonLD::fetchElement($jsonld, 'description');
1063                 if (!empty($content)) {
1064                         $jsonldinfo['publisher_description'] = trim($content);
1065                 }
1066
1067                 $content = JsonLD::fetchElement($jsonld, 'url');
1068                 if (!empty($content)) {
1069                         $jsonldinfo['publisher_url'] = trim($content);
1070                 }
1071
1072                 $content = JsonLD::fetchElement($jsonld, 'logo', 'url', '@type', 'ImageObject');
1073                 if (!empty($content)) {
1074                         $jsonldinfo['publisher_img'] = trim($content);
1075                 }
1076
1077                 $content = JsonLD::fetchElement($jsonld, 'brand', 'name', '@type', 'Organization');
1078                 if (!empty($content)) {
1079                         $jsonldinfo['publisher_name'] = trim($content);
1080                 }
1081
1082                 $content = JsonLD::fetchElement($jsonld, 'brand', 'url', '@type', 'Organization');
1083                 if (!empty($content)) {
1084                         $jsonldinfo['publisher_url'] = trim($content);
1085                 }
1086
1087                 Logger::info('Fetched Organization information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1088                 return array_merge($siteinfo, $jsonldinfo);
1089         }
1090
1091         /**
1092          * Fetch data from the provided JSON-LD Person type
1093          * @see https://schema.org/Person
1094          *
1095          * @param array $siteinfo
1096          * @param array $jsonld
1097          * @return array siteinfo
1098          */
1099         private static function parseJsonLdWebPerson(array $siteinfo, array $jsonld)
1100         {
1101                 $jsonldinfo = [];
1102
1103                 $content = JsonLD::fetchElement($jsonld, 'name');
1104                 if (!empty($content)) {
1105                         $jsonldinfo['author_name'] = trim($content);
1106                 }
1107
1108                 $content = JsonLD::fetchElement($jsonld, 'description');
1109                 if (!empty($content)) {
1110                         $jsonldinfo['author_description'] = trim($content);
1111                 }
1112
1113                 $content = JsonLD::fetchElement($jsonld, 'sameAs');
1114                 if (!empty($content) && is_string($content)) {
1115                         $jsonldinfo['author_url'] = trim($content);
1116                 }
1117
1118                 $content = JsonLD::fetchElement($jsonld, 'url');
1119                 if (!empty($content)) {
1120                         $jsonldinfo['author_url'] = trim($content);
1121                 }
1122
1123                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
1124                 if (!empty($content)) {
1125                         $jsonldinfo['author_img'] = trim($content);
1126                 }
1127
1128                 Logger::info('Fetched Person information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1129                 return array_merge($siteinfo, $jsonldinfo);
1130         }
1131
1132         /**
1133          * Fetch data from the provided JSON-LD MediaObject type
1134          * @see https://schema.org/MediaObject
1135          *
1136          * @param array $siteinfo
1137          * @param array $jsonld
1138          * @return array siteinfo
1139          */
1140         private static function parseJsonLdMediaObject(array $siteinfo, array $jsonld, string $name)
1141         {
1142                 $media = [];
1143
1144                 $content = JsonLD::fetchElement($jsonld, 'caption');
1145                 if (!empty($content)) {
1146                         $media['caption'] = trim($content);
1147                 }
1148
1149                 $content = JsonLD::fetchElement($jsonld, 'url');
1150                 if (!empty($content)) {
1151                         $media['url'] = trim($content);
1152                 }
1153
1154                 $content = JsonLD::fetchElement($jsonld, 'mainEntityOfPage');
1155                 if (!empty($content)) {
1156                         $media['main'] = Strings::compareLink($content, $siteinfo['url']);
1157                 }
1158
1159                 $content = JsonLD::fetchElement($jsonld, 'description');
1160                 if (!empty($content)) {
1161                         $media['description'] = trim($content);
1162                 }
1163
1164                 $content = JsonLD::fetchElement($jsonld, 'name');
1165                 if (!empty($content) && (($media['description'] ?? '') != trim($content))) {
1166                         $media['name'] = trim($content);
1167                 }
1168
1169                 $content = JsonLD::fetchElement($jsonld, 'contentUrl');
1170                 if (!empty($content)) {
1171                         $media['content'] = trim($content);
1172                 }
1173
1174                 $content = JsonLD::fetchElement($jsonld, 'embedUrl');
1175                 if (!empty($content)) {
1176                         $media['embed'] = trim($content);
1177                 }
1178
1179                 $content = JsonLD::fetchElement($jsonld, 'height');
1180                 if (!empty($content)) {
1181                         $media['height'] = trim($content);
1182                 }
1183
1184                 $content = JsonLD::fetchElement($jsonld, 'width');
1185                 if (!empty($content)) {
1186                         $media['width'] = trim($content);
1187                 }
1188
1189                 $content = JsonLD::fetchElement($jsonld, 'image');
1190                 if (!empty($content)) {
1191                         $media['image'] = trim($content);
1192                 }
1193
1194                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
1195                 if (!empty($content) && (($media['image'] ?? '') != trim($content))) {
1196                         if (!empty($media['image'])) {
1197                                 $media['preview'] = trim($content);
1198                         } else {
1199                                 $media['image'] = trim($content);
1200                         }
1201                 }
1202
1203                 Logger::info('Fetched Media information', ['url' => $siteinfo['url'], 'fetched' => $media]);
1204                 $siteinfo[$name][] = $media;
1205                 return $siteinfo;
1206         }
1207 }