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, $jsonld);
 476                                 }
 477                         }
 478                 }
 479
 480                 // Prevent to have a photo type without an image
 481                 if ($twitter_card && $twitter_image && !empty($siteinfo['image'])) {
 482                         $siteinfo['type'] = 'photo';
 483                 }
 484
 485                 if (!empty($siteinfo['image'])) {
 486                         $siteinfo['images'] = $siteinfo['images'] ?? [];
 487                         array_unshift($siteinfo['images'], ['url' => $siteinfo['image']]);
 488                         unset($siteinfo['image']);
 489                 }
 490
 491                 $siteinfo = self::checkMedia($url, $siteinfo);
 492
 493                 if (!empty($siteinfo['text']) && mb_strlen($siteinfo['text']) > self::MAX_DESC_COUNT) {
 494                         $siteinfo['text'] = mb_substr($siteinfo['text'], 0, self::MAX_DESC_COUNT) . '…';
 495                         $pos = mb_strrpos($siteinfo['text'], '.');
 496                         if ($pos > self::MIN_DESC_COUNT) {
 497                                 $siteinfo['text'] = mb_substr($siteinfo['text'], 0, $pos + 1);
 498                         }
 499                 }
 500
 501                 Logger::info('Siteinfo fetched', ['url' => $url, 'siteinfo' => $siteinfo]);
 502
 503                 Hook::callAll('getsiteinfo', $siteinfo);
 504
 505                 return $siteinfo;
 506         }
 507
 508         /**
 509          * Check the attached media elements.
 510          * Fix existing data and add missing data.
 511          *
 512          * @param string $page_url
 513          * @param array $siteinfo
 514          * @return void
 515          */
 516         private static function checkMedia(string $page_url, array $siteinfo)
 517         {
 518                 if (!empty($siteinfo['images'])) {
 519                         array_walk($siteinfo['images'], function (&$image) use ($page_url) {
 520                                 // According to the specifications someone could place a picture url into the content field as well.
 521                                 // But this doesn't seem to happen in the wild, so we don't cover it here.
 522                                 if (!empty($image['url'])) {
 523                                         $image['url'] = self::completeUrl($image['url'], $page_url);
 524                                         $photodata = Images::getInfoFromURLCached($image['url']);
 525                                         if (!empty($photodata) && ($photodata[0] > 50) && ($photodata[1] > 50)) {
 526                                                 $image['src'] = $image['url'];
 527                                                 $image['width'] = $photodata[0];
 528                                                 $image['height'] = $photodata[1];
 529                                                 $image['contenttype'] = $photodata['mime'];
 530                                                 unset($image['url']);
 531                                                 ksort($image);
 532                                         } else {
 533                                                 $image = [];
 534                                         }
 535                                 } else {
 536                                         $image = [];
 537                                 }
 538                         });
 539
 540                         $siteinfo['images'] = array_values(array_filter($siteinfo['images']));
 541                 }
 542
 543                 foreach (['audio', 'video'] as $element) {
 544                         if (!empty($siteinfo[$element])) {
 545                                 array_walk($siteinfo[$element], function (&$media) use ($page_url, &$siteinfo) {
 546                                         $url = '';
 547                                         $embed = '';
 548                                         $content = '';
 549                                         $contenttype = '';
 550                                         foreach (['embed', 'content', 'url'] as $field) {
 551                                                 if (!empty($media[$field])) {
 552                                                         $media[$field] = self::completeUrl($media[$field], $page_url);
 553                                                         $type = self::getContentType($media[$field]);
 554                                                         if (($type[0] ?? '') == 'text') {
 555                                                                 if ($field == 'embed') {
 556                                                                         $embed = $media[$field];
 557                                                                 } else {
 558                                                                         $url = $media[$field];
 559                                                                 }
 560                                                         } elseif (!empty($type[0])) {
 561                                                                 $content = $media[$field];
 562                                                                 $contenttype = implode('/', $type);
 563                                                         }
 564                                                 }
 565                                                 unset($media[$field]);
 566                                         }
 567
 568                                         foreach (['image', 'preview'] as $field) {
 569                                                 if (!empty($media[$field])) {
 570                                                         $media[$field] = self::completeUrl($media[$field], $page_url);
 571                                                 }
 572                                         }
 573
 574                                         if (!empty($url)) {
 575                                                 $media['url'] = $url;
 576                                         }
 577                                         if (!empty($embed)) {
 578                                                 $media['embed'] = $embed;
 579                                                 if (!empty($media['main'])) {
 580                                                         $siteinfo['embed'] = $embed;
 581                                                 }
 582                                         }
 583                                         if (!empty($content)) {
 584                                                 $media['src'] = $content;
 585                                         }
 586                                         if (!empty($contenttype)) {
 587                                                 $media['contenttype'] = $contenttype;
 588                                         }
 589                                         if (empty($url) && empty($content) && empty($embed)) {
 590                                                 $media = [];
 591                                         }
 592                                         ksort($media);
 593                                 });
 594
 595                                 $siteinfo[$element] = array_values(array_filter($siteinfo[$element]));
 596                         }
 597                         if (empty($siteinfo[$element])) {
 598                                 unset($siteinfo[$element]);
 599                         }
 600                 }
 601                 return $siteinfo;
 602         }
 603
 604         /**
 605          * Convert tags from CSV to an array
 606          *
 607          * @param string $string Tags
 608          * @return array with formatted Hashtags
 609          */
 610         public static function convertTagsToArray($string)
 611         {
 612                 $arr_tags = str_getcsv($string);
 613                 if (count($arr_tags)) {
 614                         // add the # sign to every tag
 615                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 616
 617                         return $arr_tags;
 618                 }
 619         }
 620
 621         /**
 622          * Add a hasht sign to a string
 623          *
 624          * This method is used as callback function
 625          *
 626          * @param string $tag The pure tag name
 627          * @param int    $k   Counter for internal use
 628          * @return void
 629          */
 630         private static function arrAddHashes(&$tag, $k)
 631         {
 632                 $tag = "#" . $tag;
 633         }
 634
 635         /**
 636          * Add a scheme to an url
 637          *
 638          * The src attribute of some html elements (e.g. images)
 639          * can miss the scheme so we need to add the correct
 640          * scheme
 641          *
 642          * @param string $url    The url which possibly does have
 643          *                       a missing scheme (a link to an image)
 644          * @param string $scheme The url with a correct scheme
 645          *                       (e.g. the url from the webpage which does contain the image)
 646          *
 647          * @return string The url with a scheme
 648          */
 649         private static function completeUrl($url, $scheme)
 650         {
 651                 $urlarr = parse_url($url);
 652
 653                 // If the url does allready have an scheme
 654                 // we can stop the process here
 655                 if (isset($urlarr["scheme"])) {
 656                         return($url);
 657                 }
 658
 659                 $schemearr = parse_url($scheme);
 660
 661                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 662
 663                 if (!empty($schemearr["port"])) {
 664                         $complete .= ":".$schemearr["port"];
 665                 }
 666
 667                 if (!empty($urlarr["path"])) {
 668                         if (strpos($urlarr["path"], "/") !== 0) {
 669                                 $complete .= "/";
 670                         }
 671
 672                         $complete .= $urlarr["path"];
 673                 }
 674
 675                 if (!empty($urlarr["query"])) {
 676                         $complete .= "?".$urlarr["query"];
 677                 }
 678
 679                 if (!empty($urlarr["fragment"])) {
 680                         $complete .= "#".$urlarr["fragment"];
 681                 }
 682
 683                 return($complete);
 684         }
 685
 686         /**
 687          * Parse the Json-Ld parts of a web page
 688          *
 689          * @param array $siteinfo
 690          * @param array $jsonld
 691          * @return array siteinfo
 692          */
 693         private static function parseParts(array $siteinfo, array $jsonld)
 694         {
 695                 if (!empty($jsonld['@graph']) && is_array($jsonld['@graph'])) {
 696                         foreach ($jsonld['@graph'] as $part) {
 697                                 if (!empty($part)) {
 698                                         $siteinfo = self::parseParts($siteinfo, $part);
 699                                 }
 700                         }
 701                 } elseif (!empty($jsonld['@type'])) {
 702                         $siteinfo = self::parseJsonLd($siteinfo, $jsonld);
 703                 } elseif (!empty($jsonld)) {
 704                         $keys = array_keys($jsonld);
 705                         $numeric_keys = true;
 706                         foreach ($keys as $key) {
 707                                 if (!is_int($key)) {
 708                                         $numeric_keys = false;
 709                                 }
 710                         }
 711                         if ($numeric_keys) {
 712                                 foreach ($jsonld as $part) {
 713                                         if (!empty($part)) {
 714                                                 $siteinfo = self::parseParts($siteinfo, $part);
 715                                         }
 716                                 }
 717                         }
 718                 }
 719
 720                 array_walk_recursive($siteinfo, function (&$element) {
 721                         if (is_string($element)) {
 722                                 $element = trim(strip_tags(html_entity_decode($element, ENT_COMPAT, 'UTF-8')));
 723                         }
 724                 });
 725
 726                 return $siteinfo;
 727         }
 728
 729         /**
 730          * Improve the siteinfo with information from the provided JSON-LD information
 731          * @see https://jsonld.com/
 732          * @see https://schema.org/
 733          *
 734          * @param array $siteinfo
 735          * @param array $jsonld
 736          * @return array siteinfo
 737          */
 738         private static function parseJsonLd(array $siteinfo, array $jsonld)
 739         {
 740                 $type = JsonLD::fetchElement($jsonld, '@type');
 741                 if (empty($type)) {
 742                         Logger::info('Empty type', ['url' => $siteinfo['url']]);
 743                         return $siteinfo;
 744                 }
 745
 746                 // Silently ignore some types that aren't processed
 747                 if (in_array($type, ['SiteNavigationElement', 'JobPosting', 'CreativeWork', 'MusicAlbum',
 748                         'WPHeader', 'WPSideBar', 'WPFooter', 'LegalService', 'MusicRecording',
 749                         'ItemList', 'BreadcrumbList', 'Blog', 'Dataset', 'Product'])) {
 750                         return $siteinfo;
 751                 }
 752
 753                 switch ($type) {
 754                         case 'Article':
 755                         case 'AdvertiserContentArticle':
 756                         case 'NewsArticle':
 757                         case 'Report':
 758                         case 'SatiricalArticle':
 759                         case 'ScholarlyArticle':
 760                         case 'SocialMediaPosting':
 761                         case 'TechArticle':
 762                         case 'ReportageNewsArticle':
 763                         case 'SocialMediaPosting':
 764                         case 'BlogPosting':
 765                         case 'LiveBlogPosting':
 766                         case 'DiscussionForumPosting':
 767                                 return self::parseJsonLdArticle($siteinfo, $jsonld);
 768                         case 'WebPage':
 769                         case 'AboutPage':
 770                         case 'CheckoutPage':
 771                         case 'CollectionPage':
 772                         case 'ContactPage':
 773                         case 'FAQPage':
 774                         case 'ItemPage':
 775                         case 'MedicalWebPage':
 776                         case 'ProfilePage':
 777                         case 'QAPage':
 778                         case 'RealEstateListing':
 779                         case 'SearchResultsPage':
 780                         case 'MediaGallery':
 781                         case 'ImageGallery':
 782                         case 'VideoGallery':
 783                         case 'RadioEpisode':
 784                         case 'Event':
 785                                 return self::parseJsonLdWebPage($siteinfo, $jsonld);
 786                         case 'WebSite':
 787                                 return self::parseJsonLdWebSite($siteinfo, $jsonld);
 788                         case 'Organization':
 789                         case 'Airline':
 790                         case 'Consortium':
 791                         case 'Corporation':
 792                         case 'EducationalOrganization':
 793                         case 'FundingScheme':
 794                         case 'GovernmentOrganization':
 795                         case 'LibrarySystem':
 796                         case 'LocalBusiness':
 797                         case 'MedicalOrganization':
 798                         case 'NGO':
 799                         case 'NewsMediaOrganization':
 800                         case 'Project':
 801                         case 'SportsOrganization':
 802                         case 'WorkersUnion':
 803                                 return self::parseJsonLdWebOrganization($siteinfo, $jsonld);
 804                         case 'Person':
 805                         case 'Patient':
 806                         case 'PerformingGroup':
 807                         case 'DanceGroup';
 808                         case 'MusicGroup':
 809                         case 'TheaterGroup':
 810                                 return self::parseJsonLdWebPerson($siteinfo, $jsonld);
 811                         case 'AudioObject':
 812                         case 'Audio':
 813                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'audio');
 814                         case 'VideoObject':
 815                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'video');
 816                         case 'ImageObject':
 817                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'images');
 818                         default:
 819                                 Logger::info('Unknown type', ['type' => $type, 'url' => $siteinfo['url']]);
 820                                 return $siteinfo;
 821                 }
 822         }
 823
 824         /**
 825          * Fetch author and publisher data
 826          *
 827          * @param array $siteinfo
 828          * @param array $jsonld
 829          * @return array siteinfo
 830          */
 831         private static function parseJsonLdAuthor(array $siteinfo, array $jsonld)
 832         {
 833                 $jsonldinfo = [];
 834
 835                 if (!empty($jsonld['publisher']) && is_array($jsonld['publisher'])) {
 836                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'name');
 837                         if (!empty($content) && is_string($content)) {
 838                                 $jsonldinfo['publisher_name'] = trim($content);
 839                         }
 840
 841                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'url');
 842                         if (!empty($content) && is_string($content)) {
 843                                 $jsonldinfo['publisher_url'] = trim($content);
 844                         }
 845
 846                         $brand = JsonLD::fetchElement($jsonld, 'publisher', 'brand', '@type', 'Organization');
 847                         if (!empty($brand) && is_array($brand)) {
 848                                 $content = JsonLD::fetchElement($brand, 'name');
 849                                 if (!empty($content) && is_string($content)) {
 850                                         $jsonldinfo['publisher_name'] = trim($content);
 851                                 }
 852
 853                                 $content = JsonLD::fetchElement($brand, 'url');
 854                                 if (!empty($content) && is_string($content)) {
 855                                         $jsonldinfo['publisher_url'] = trim($content);
 856                                 }
 857
 858                                 $content = JsonLD::fetchElement($brand, 'logo', 'url');
 859                                 if (!empty($content) && is_string($content)) {
 860                                         $jsonldinfo['publisher_img'] = trim($content);
 861                                 }
 862                         }
 863
 864                         $logo = JsonLD::fetchElement($jsonld, 'publisher', 'logo');
 865                         if (!empty($logo) && is_array($logo)) {
 866                                 $content = JsonLD::fetchElement($logo, 'url');
 867                                 if (!empty($content) && is_string($content)) {
 868                                         $jsonldinfo['publisher_img'] = trim($content);
 869                                 }
 870                         }
 871                 } elseif (!empty($jsonld['publisher']) && is_string($jsonld['publisher'])) {
 872                         $jsonldinfo['publisher_name'] = trim($jsonld['publisher']);
 873                 }
 874
 875                 if (!empty($jsonld['author']) && is_array($jsonld['author'])) {
 876                         $content = JsonLD::fetchElement($jsonld, 'author', 'name');
 877                         if (!empty($content) && is_string($content)) {
 878                                 $jsonldinfo['author_name'] = trim($content);
 879                         }
 880
 881                         $content = JsonLD::fetchElement($jsonld, 'author', 'sameAs');
 882                         if (!empty($content) && is_string($content)) {
 883                                 $jsonldinfo['author_url'] = trim($content);
 884                         }
 885
 886                         $content = JsonLD::fetchElement($jsonld, 'author', 'url');
 887                         if (!empty($content) && is_string($content)) {
 888                                 $jsonldinfo['author_url'] = trim($content);
 889                         }
 890
 891                         $logo = JsonLD::fetchElement($jsonld, 'author', 'logo');
 892                         if (!empty($logo) && is_array($logo)) {
 893                                 $content = JsonLD::fetchElement($logo, 'url');
 894                                 if (!empty($content) && is_string($content)) {
 895                                         $jsonldinfo['author_img'] = trim($content);
 896                                 }
 897                         }
 898                 } elseif (!empty($jsonld['author']) && is_string($jsonld['author'])) {
 899                         $jsonldinfo['author_name'] = trim($jsonld['author']);
 900                 }
 901
 902                 Logger::info('Fetched Author information', ['fetched' => $jsonldinfo]);
 903
 904                 return array_merge($siteinfo, $jsonldinfo);
 905         }
 906
 907         /**
 908          * Fetch data from the provided JSON-LD Article type
 909          * @see https://schema.org/Article
 910          *
 911          * @param array $siteinfo
 912          * @param array $jsonld
 913          * @return array siteinfo
 914          */
 915         private static function parseJsonLdArticle(array $siteinfo, array $jsonld)
 916         {
 917                 $jsonldinfo = [];
 918
 919                 $content = JsonLD::fetchElement($jsonld, 'headline');
 920                 if (!empty($content) && is_string($content)) {
 921                         $jsonldinfo['title'] = trim($content);
 922                 }
 923
 924                 $content = JsonLD::fetchElement($jsonld, 'alternativeHeadline');
 925                 if (!empty($content) && is_string($content) && (($jsonldinfo['title'] ?? '') != trim($content))) {
 926                         $jsonldinfo['alternative_title'] = trim($content);
 927                 }
 928
 929                 $content = JsonLD::fetchElement($jsonld, 'description');
 930                 if (!empty($content) && is_string($content)) {
 931                         $jsonldinfo['text'] = trim($content);
 932                 }
 933
 934                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 935                 if (!empty($content)) {
 936                         $jsonldinfo['image'] = trim($content);
 937                 }
 938
 939                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
 940                 if (!empty($content) && is_string($content)) {
 941                         $jsonldinfo['image'] = trim($content);
 942                 }
 943
 944                 if (!empty($jsonld['keywords']) && !is_array($jsonld['keywords'])) {
 945                         $content = JsonLD::fetchElement($jsonld, 'keywords');
 946                         if (!empty($content)) {
 947                                 $siteinfo['keywords'] = [];
 948                                 $keywords = explode(',', $content);
 949                                 foreach ($keywords as $keyword) {
 950                                         $siteinfo['keywords'][] = trim($keyword);
 951                                 }
 952                         }
 953                 } elseif (!empty($jsonld['keywords'])) {
 954                         $content = JsonLD::fetchElementArray($jsonld, 'keywords');
 955                         if (!empty($content) && is_array($content)) {
 956                                 $jsonldinfo['keywords'] = $content;
 957                         }
 958                 }
 959
 960                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 961
 962                 Logger::info('Fetched article information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 963
 964                 return array_merge($siteinfo, $jsonldinfo);
 965         }
 966
 967         /**
 968          * Fetch data from the provided JSON-LD WebPage type
 969          * @see https://schema.org/WebPage
 970          *
 971          * @param array $siteinfo
 972          * @param array $jsonld
 973          * @return array siteinfo
 974          */
 975         private static function parseJsonLdWebPage(array $siteinfo, array $jsonld)
 976         {
 977                 $jsonldinfo = [];
 978
 979                 $content = JsonLD::fetchElement($jsonld, 'name');
 980                 if (!empty($content)) {
 981                         $jsonldinfo['title'] = trim($content);
 982                 }
 983
 984                 $content = JsonLD::fetchElement($jsonld, 'description');
 985                 if (!empty($content)) {
 986                         $jsonldinfo['text'] = trim($content);
 987                 }
 988
 989                 $content = JsonLD::fetchElement($jsonld, 'image');
 990                 if (!empty($content)) {
 991                         $jsonldinfo['image'] = trim($content);
 992                 }
 993
 994                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 995                 if (!empty($content)) {
 996                         $jsonldinfo['image'] = trim($content);
 997                 }
 998
 999                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
1000
1001                 Logger::info('Fetched WebPage information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1002
1003                 return array_merge($siteinfo, $jsonldinfo);
1004         }
1005
1006         /**
1007          * Fetch data from the provided JSON-LD WebSite type
1008          * @see https://schema.org/WebSite
1009          *
1010          * @param array $siteinfo
1011          * @param array $jsonld
1012          * @return array siteinfo
1013          */
1014         private static function parseJsonLdWebSite(array $siteinfo, array $jsonld)
1015         {
1016                 $jsonldinfo = [];
1017
1018                 $content = JsonLD::fetchElement($jsonld, 'name');
1019                 if (!empty($content)) {
1020                         $jsonldinfo['publisher_name'] = trim($content);
1021                 }
1022
1023                 $content = JsonLD::fetchElement($jsonld, 'description');
1024                 if (!empty($content)) {
1025                         $jsonldinfo['publisher_description'] = trim($content);
1026                 }
1027
1028                 $content = JsonLD::fetchElement($jsonld, 'url');
1029                 if (!empty($content)) {
1030                         $jsonldinfo['publisher_url'] = trim($content);
1031                 }
1032
1033                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
1034                 if (!empty($content)) {
1035                         $jsonldinfo['image'] = trim($content);
1036                 }
1037
1038                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
1039
1040                 Logger::info('Fetched WebSite information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1041                 return array_merge($siteinfo, $jsonldinfo);
1042         }
1043
1044         /**
1045          * Fetch data from the provided JSON-LD Organization type
1046          * @see https://schema.org/Organization
1047          *
1048          * @param array $siteinfo
1049          * @param array $jsonld
1050          * @return array siteinfo
1051          */
1052         private static function parseJsonLdWebOrganization(array $siteinfo, array $jsonld)
1053         {
1054                 $jsonldinfo = [];
1055
1056                 $content = JsonLD::fetchElement($jsonld, 'name');
1057                 if (!empty($content)) {
1058                         $jsonldinfo['publisher_name'] = trim($content);
1059                 }
1060
1061                 $content = JsonLD::fetchElement($jsonld, 'description');
1062                 if (!empty($content)) {
1063                         $jsonldinfo['publisher_description'] = trim($content);
1064                 }
1065
1066                 $content = JsonLD::fetchElement($jsonld, 'url');
1067                 if (!empty($content)) {
1068                         $jsonldinfo['publisher_url'] = trim($content);
1069                 }
1070
1071                 $content = JsonLD::fetchElement($jsonld, 'logo', 'url', '@type', 'ImageObject');
1072                 if (!empty($content)) {
1073                         $jsonldinfo['publisher_img'] = trim($content);
1074                 }
1075
1076                 $content = JsonLD::fetchElement($jsonld, 'brand', 'name', '@type', 'Organization');
1077                 if (!empty($content)) {
1078                         $jsonldinfo['publisher_name'] = trim($content);
1079                 }
1080
1081                 $content = JsonLD::fetchElement($jsonld, 'brand', 'url', '@type', 'Organization');
1082                 if (!empty($content)) {
1083                         $jsonldinfo['publisher_url'] = trim($content);
1084                 }
1085
1086                 Logger::info('Fetched Organization information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1087                 return array_merge($siteinfo, $jsonldinfo);
1088         }
1089
1090         /**
1091          * Fetch data from the provided JSON-LD Person type
1092          * @see https://schema.org/Person
1093          *
1094          * @param array $siteinfo
1095          * @param array $jsonld
1096          * @return array siteinfo
1097          */
1098         private static function parseJsonLdWebPerson(array $siteinfo, array $jsonld)
1099         {
1100                 $jsonldinfo = [];
1101
1102                 $content = JsonLD::fetchElement($jsonld, 'name');
1103                 if (!empty($content)) {
1104                         $jsonldinfo['author_name'] = trim($content);
1105                 }
1106
1107                 $content = JsonLD::fetchElement($jsonld, 'description');
1108                 if (!empty($content)) {
1109                         $jsonldinfo['author_description'] = trim($content);
1110                 }
1111
1112                 $content = JsonLD::fetchElement($jsonld, 'sameAs');
1113                 if (!empty($content) && is_string($content)) {
1114                         $jsonldinfo['author_url'] = trim($content);
1115                 }
1116
1117                 $content = JsonLD::fetchElement($jsonld, 'url');
1118                 if (!empty($content)) {
1119                         $jsonldinfo['author_url'] = trim($content);
1120                 }
1121
1122                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
1123                 if (!empty($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 }