src/Util/ParseUrl.php

   1 <?php
   2 /**
   3  * @copyright Copyright (C) 2020, Friendica
   4  *
   5  * @license GNU AGPL version 3 or any later version
   6  *
   7  * This program is free software: you can redistribute it and/or modify
   8  * it under the terms of the GNU Affero General Public License as
   9  * published by the Free Software Foundation, either version 3 of the
  10  * License, or (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15  * GNU Affero General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU Affero General Public License
  18  * along with this program.  If not, see <https://www.gnu.org/licenses/>.
  19  *
  20  */
  21
  22 namespace Friendica\Util;
  23
  24 use DOMDocument;
  25 use DOMXPath;
  26 use Friendica\Content\OEmbed;
  27 use Friendica\Core\Hook;
  28 use Friendica\Core\Logger;
  29 use Friendica\Database\Database;
  30 use Friendica\Database\DBA;
  31 use Friendica\DI;
  32 use Friendica\Network\HTTPException;
  33
  34 /**
  35  * Get information about a given URL
  36  *
  37  * Class with methods for extracting certain content from an url
  38  */
  39 class ParseUrl
  40 {
  41         const DEFAULT_EXPIRATION_FAILURE = 'now + 1 day';
  42         const DEFAULT_EXPIRATION_SUCCESS = 'now + 3 months';
  43
  44         /**
  45          * Maximum number of characters for the description
  46          */
  47         const MAX_DESC_COUNT = 250;
  48
  49         /**
  50          * Minimum number of characters for the description
  51          */
  52         const MIN_DESC_COUNT = 100;
  53
  54         /**
  55          * Fetch the content type of the given url
  56          * @param string $url URL of the page
  57          * @return array content type
  58          */
  59         public static function getContentType(string $url)
  60         {
  61                 $curlResult = DI::httpRequest()->head($url);
  62                 if (!$curlResult->isSuccess()) {
  63                         return [];
  64                 }
  65
  66                 $contenttype =  $curlResult->getHeader('Content-Type');
  67                 if (empty($contenttype)) {
  68                         return [];
  69                 }
  70
  71                 return explode('/', current(explode(';', $contenttype)));
  72         }
  73
  74         /**
  75          * Search for chached embeddable data of an url otherwise fetch it
  76          *
  77          * @param string $url         The url of the page which should be scraped
  78          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
  79          *                            to avoid endless loops
  80          *
  81          * @return array which contains needed data for embedding
  82          *    string 'url'      => The url of the parsed page
  83          *    string 'type'     => Content type
  84          *    string 'title'    => (optional) The title of the content
  85          *    string 'text'     => (optional) The description for the content
  86          *    string 'image'    => (optional) A preview image of the content
  87          *    array  'images'   => (optional) Array of preview pictures
  88          *    string 'keywords' => (optional) The tags which belong to the content
  89          *
  90          * @throws HTTPException\InternalServerErrorException
  91          * @see   ParseUrl::getSiteinfo() for more information about scraping
  92          * embeddable content
  93          */
  94         public static function getSiteinfoCached($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()) {
 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                 $list = $xpath->query('//meta[@name]');
 353                 foreach ($list as $node) {
 354                         $meta_tag = [];
 355                         if ($node->attributes->length) {
 356                                 foreach ($node->attributes as $attribute) {
 357                                         $meta_tag[$attribute->name] = $attribute->value;
 358                                 }
 359                         }
 360
 361                         if (empty($meta_tag['content'])) {
 362                                 continue;
 363                         }
 364
 365                         $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 366
 367                         switch (strtolower($meta_tag['name'])) {
 368                                 case 'fulltitle':
 369                                         $siteinfo['title'] = trim($meta_tag['content']);
 370                                         break;
 371                                 case 'description':
 372                                         $siteinfo['text'] = trim($meta_tag['content']);
 373                                         break;
 374                                 case 'thumbnail':
 375                                         $siteinfo['image'] = $meta_tag['content'];
 376                                         break;
 377                                 case 'twitter:image':
 378                                         $siteinfo['image'] = $meta_tag['content'];
 379                                         break;
 380                                 case 'twitter:image:src':
 381                                         $siteinfo['image'] = $meta_tag['content'];
 382                                         break;
 383                                 case 'twitter:card':
 384                                         // Detect photo pages
 385                                         if ($meta_tag['content'] == 'summary_large_image') {
 386                                                 $siteinfo['type'] = 'photo';
 387                                         }
 388                                         break;
 389                                 case 'twitter:description':
 390                                         $siteinfo['text'] = trim($meta_tag['content']);
 391                                         break;
 392                                 case 'twitter:title':
 393                                         $siteinfo['title'] = trim($meta_tag['content']);
 394                                         break;
 395                                 case 'dc.title':
 396                                         $siteinfo['title'] = trim($meta_tag['content']);
 397                                         break;
 398                                 case 'dc.description':
 399                                         $siteinfo['text'] = trim($meta_tag['content']);
 400                                         break;
 401                                 case 'dc.creator':
 402                                         $siteinfo['publisher_name'] = trim($meta_tag['content']);
 403                                         break;
 404                                 case 'keywords':
 405                                         $keywords = explode(',', $meta_tag['content']);
 406                                         break;
 407                                 case 'news_keywords':
 408                                         $keywords = explode(',', $meta_tag['content']);
 409                                         break;
 410                         }
 411                 }
 412
 413                 if (isset($keywords)) {
 414                         $siteinfo['keywords'] = [];
 415                         foreach ($keywords as $keyword) {
 416                                 if (!in_array(trim($keyword), $siteinfo['keywords'])) {
 417                                         $siteinfo['keywords'][] = trim($keyword);
 418                                 }
 419                         }
 420                 }
 421
 422                 $list = $xpath->query('//meta[@property]');
 423                 foreach ($list as $node) {
 424                         $meta_tag = [];
 425                         if ($node->attributes->length) {
 426                                 foreach ($node->attributes as $attribute) {
 427                                         $meta_tag[$attribute->name] = $attribute->value;
 428                                 }
 429                         }
 430
 431                         if (!empty($meta_tag['content'])) {
 432                                 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 433
 434                                 switch (strtolower($meta_tag['property'])) {
 435                                         case 'og:image':
 436                                                 $siteinfo['image'] = $meta_tag['content'];
 437                                                 break;
 438                                         case 'og:image:url':
 439                                                 $siteinfo['image'] = $meta_tag['content'];
 440                                                 break;
 441                                         case 'og:image:secure_url':
 442                                                 $siteinfo['image'] = $meta_tag['content'];
 443                                                 break;
 444                                         case 'og:title':
 445                                                 $siteinfo['title'] = trim($meta_tag['content']);
 446                                                 break;
 447                                         case 'og:description':
 448                                                 $siteinfo['text'] = trim($meta_tag['content']);
 449                                                 break;
 450                                         case 'og:site_name':
 451                                                 $siteinfo['publisher_name'] = trim($meta_tag['content']);
 452                                                 break;
 453                                         case 'twitter:description':
 454                                                 $siteinfo['text'] = trim($meta_tag['content']);
 455                                                 break;
 456                                         case 'twitter:title':
 457                                                 $siteinfo['title'] = trim($meta_tag['content']);
 458                                                 break;
 459                                         case 'twitter:image':
 460                                                 $siteinfo['image'] = $meta_tag['content'];
 461                                                 break;
 462                                 }
 463                         }
 464                 }
 465
 466                 $list = $xpath->query("//script[@type='application/ld+json']");
 467                 foreach ($list as $node) {
 468                         if (!empty($node->nodeValue)) {
 469                                 $nodevalue = html_entity_decode($node->nodeValue, ENT_COMPAT, 'UTF-8');
 470                                 if ($jsonld = json_decode($nodevalue, true)) {
 471                                         $siteinfo = self::parseParts($siteinfo, $jsonld);
 472                                 }
 473                         }
 474                 }
 475
 476                 // Prevent to have a photo type without an image
 477                 if ((empty($siteinfo['image']) || !empty($siteinfo['text'])) && ($siteinfo['type'] == 'photo')) {
 478                         $siteinfo['type'] = 'link';
 479                 }
 480
 481                 if (!empty($siteinfo['image']) && empty($siteinfo['images'])) {
 482                         $src = self::completeUrl($siteinfo['image'], $url);
 483
 484                         unset($siteinfo['image']);
 485
 486                         $photodata = Images::getInfoFromURLCached($src);
 487
 488                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 489                                 $siteinfo['images'][] = ['src' => $src,
 490                                         'width' => $photodata[0],
 491                                         'height' => $photodata[1]];
 492                         }
 493                 }
 494
 495                 if (!empty($siteinfo['text']) && mb_strlen($siteinfo['text']) > self::MAX_DESC_COUNT) {
 496                         $siteinfo['text'] = mb_substr($siteinfo['text'], 0, self::MAX_DESC_COUNT) . '…';
 497                         $pos = mb_strrpos($siteinfo['text'], '.');
 498                         if ($pos > self::MIN_DESC_COUNT) {
 499                                 $siteinfo['text'] = mb_substr($siteinfo['text'], 0, $pos + 1);
 500                         }
 501                 }
 502
 503                 Logger::info('Siteinfo fetched', ['url' => $url, 'siteinfo' => $siteinfo]);
 504
 505                 Hook::callAll('getsiteinfo', $siteinfo);
 506
 507                 return $siteinfo;
 508         }
 509
 510         /**
 511          * Convert tags from CSV to an array
 512          *
 513          * @param string $string Tags
 514          * @return array with formatted Hashtags
 515          */
 516         public static function convertTagsToArray($string)
 517         {
 518                 $arr_tags = str_getcsv($string);
 519                 if (count($arr_tags)) {
 520                         // add the # sign to every tag
 521                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 522
 523                         return $arr_tags;
 524                 }
 525         }
 526
 527         /**
 528          * Add a hasht sign to a string
 529          *
 530          * This method is used as callback function
 531          *
 532          * @param string $tag The pure tag name
 533          * @param int    $k   Counter for internal use
 534          * @return void
 535          */
 536         private static function arrAddHashes(&$tag, $k)
 537         {
 538                 $tag = "#" . $tag;
 539         }
 540
 541         /**
 542          * Add a scheme to an url
 543          *
 544          * The src attribute of some html elements (e.g. images)
 545          * can miss the scheme so we need to add the correct
 546          * scheme
 547          *
 548          * @param string $url    The url which possibly does have
 549          *                       a missing scheme (a link to an image)
 550          * @param string $scheme The url with a correct scheme
 551          *                       (e.g. the url from the webpage which does contain the image)
 552          *
 553          * @return string The url with a scheme
 554          */
 555         private static function completeUrl($url, $scheme)
 556         {
 557                 $urlarr = parse_url($url);
 558
 559                 // If the url does allready have an scheme
 560                 // we can stop the process here
 561                 if (isset($urlarr["scheme"])) {
 562                         return($url);
 563                 }
 564
 565                 $schemearr = parse_url($scheme);
 566
 567                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 568
 569                 if (!empty($schemearr["port"])) {
 570                         $complete .= ":".$schemearr["port"];
 571                 }
 572
 573                 if (!empty($urlarr["path"])) {
 574                         if (strpos($urlarr["path"], "/") !== 0) {
 575                                 $complete .= "/";
 576                         }
 577
 578                         $complete .= $urlarr["path"];
 579                 }
 580
 581                 if (!empty($urlarr["query"])) {
 582                         $complete .= "?".$urlarr["query"];
 583                 }
 584
 585                 if (!empty($urlarr["fragment"])) {
 586                         $complete .= "#".$urlarr["fragment"];
 587                 }
 588
 589                 return($complete);
 590         }
 591
 592         /**
 593          * Parse the Json-Ld parts of a web page
 594          *
 595          * @param array $siteinfo
 596          * @param array $jsonld
 597          * @return array siteinfo
 598          */
 599         private static function parseParts(array $siteinfo, array $jsonld)
 600         {
 601                 if (!empty($jsonld['@graph']) && is_array($jsonld['@graph'])) {
 602                         foreach ($jsonld['@graph'] as $part) {
 603                                 $siteinfo = self::parseParts($siteinfo, $part);
 604                         }
 605                 } elseif (!empty($jsonld['@type'])) {
 606                         $siteinfo = self::parseJsonLd($siteinfo, $jsonld);
 607                 } elseif (!empty($jsonld)) {
 608                         $keys = array_keys($jsonld);
 609                         $numeric_keys = true;
 610                         foreach ($keys as $key) {
 611                                 if (!is_int($key)) {
 612                                         $numeric_keys = false;
 613                                 }
 614                         }
 615                         if ($numeric_keys) {
 616                                 foreach ($jsonld as $part) {
 617                                         $siteinfo = self::parseParts($siteinfo, $part);
 618                                 }
 619                         }
 620                 }
 621
 622                 return $siteinfo;
 623         }
 624
 625         /**
 626          * Improve the siteinfo with information from the provided JSON-LD information
 627          * @see https://jsonld.com/
 628          * @see https://schema.org/
 629          *
 630          * @param array $siteinfo
 631          * @param array $jsonld
 632          * @return array siteinfo
 633          */
 634         private static function parseJsonLd(array $siteinfo, array $jsonld)
 635         {
 636                 $type = JsonLD::fetchElement($jsonld, '@type');
 637                 if (empty($type)) {
 638                         Logger::info('Empty type', ['url' => $siteinfo['url']]);
 639                         return $siteinfo;
 640                 }
 641
 642                 // Silently ignore some types that aren't processed
 643                 if (in_array($type, ['SiteNavigationElement', 'JobPosting', 'CreativeWork', 'MusicAlbum',
 644                         'WPHeader', 'WPSideBar', 'WPFooter', 'LegalService', 'MusicRecording',
 645                         'ItemList', 'BreadcrumbList', 'Blog', 'Dataset', 'Product'])) {
 646                         return $siteinfo;
 647                 }
 648
 649                 switch ($type) {
 650                         case 'Article':
 651                         case 'AdvertiserContentArticle':
 652                         case 'NewsArticle':
 653                         case 'Report':
 654                         case 'SatiricalArticle':
 655                         case 'ScholarlyArticle':
 656                         case 'SocialMediaPosting':
 657                         case 'TechArticle':
 658                         case 'ReportageNewsArticle':
 659                         case 'SocialMediaPosting':
 660                         case 'BlogPosting':
 661                         case 'LiveBlogPosting':
 662                         case 'DiscussionForumPosting':
 663                                 return self::parseJsonLdArticle($siteinfo, $jsonld);
 664                         case 'WebPage':
 665                         case 'AboutPage':
 666                         case 'CheckoutPage':
 667                         case 'CollectionPage':
 668                         case 'ContactPage':
 669                         case 'FAQPage':
 670                         case 'ItemPage':
 671                         case 'MedicalWebPage':
 672                         case 'ProfilePage':
 673                         case 'QAPage':
 674                         case 'RealEstateListing':
 675                         case 'SearchResultsPage':
 676                         case 'MediaGallery':
 677                         case 'ImageGallery':
 678                         case 'VideoGallery':
 679                         case 'RadioEpisode':
 680                         case 'Event':
 681                                 return self::parseJsonLdWebPage($siteinfo, $jsonld);
 682                         case 'WebSite':
 683                                 return self::parseJsonLdWebSite($siteinfo, $jsonld);
 684                         case 'Organization':
 685                         case 'Airline':
 686                         case 'Consortium':
 687                         case 'Corporation':
 688                         case 'EducationalOrganization':
 689                         case 'FundingScheme':
 690                         case 'GovernmentOrganization':
 691                         case 'LibrarySystem':
 692                         case 'LocalBusiness':
 693                         case 'MedicalOrganization':
 694                         case 'NGO':
 695                         case 'NewsMediaOrganization':
 696                         case 'Project':
 697                         case 'SportsOrganization':
 698                         case 'WorkersUnion':
 699                                 return self::parseJsonLdWebOrganization($siteinfo, $jsonld);
 700                         case 'Person':
 701                         case 'Patient':
 702                         case 'PerformingGroup':
 703                         case 'DanceGroup';
 704                         case 'MusicGroup':
 705                         case 'TheaterGroup':
 706                                 return self::parseJsonLdWebPerson($siteinfo, $jsonld);
 707                         case 'AudioObject':
 708                         case 'Audio':
 709                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'audio');
 710                         case 'VideoObject':
 711                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'video');
 712                         case 'ImageObject':
 713                                 return self::parseJsonLdMediaObject($siteinfo, $jsonld, 'images');
 714                         default:
 715                                 Logger::info('Unknown type', ['type' => $type, 'url' => $siteinfo['url']]);
 716                                 return $siteinfo;
 717                 }
 718         }
 719
 720         /**
 721          * Fetch author and publisher data
 722          *
 723          * @param array $siteinfo
 724          * @param array $jsonld
 725          * @return array siteinfo
 726          */
 727         private static function parseJsonLdAuthor(array $siteinfo, array $jsonld)
 728         {
 729                 $jsonldinfo = [];
 730
 731                 if (!empty($jsonld['publisher']) && is_array($jsonld['publisher'])) {
 732                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'name');
 733                         if (!empty($content) && is_string($content)) {
 734                                 $jsonldinfo['publisher_name'] = trim($content);
 735                         }
 736
 737                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'sameAs');
 738                         if (!empty($content) && is_string($content)) {
 739                                 $jsonldinfo['publisher_url'] = trim($content);
 740                         }
 741
 742                         $content = JsonLD::fetchElement($jsonld, 'publisher', 'url');
 743                         if (!empty($content) && is_string($content)) {
 744                                 $jsonldinfo['publisher_url'] = trim($content);
 745                         }
 746
 747                         $brand = JsonLD::fetchElement($jsonld, 'publisher', 'brand', '@type', 'Organization');
 748                         if (!empty($brand) && is_array($brand)) {
 749                                 $content = JsonLD::fetchElement($brand, 'name');
 750                                 if (!empty($content) && is_string($content)) {
 751                                         $jsonldinfo['publisher_name'] = trim($content);
 752                                 }
 753
 754                                 $content = JsonLD::fetchElement($brand, 'sameAs');
 755                                 if (!empty($content) && is_string($content)) {
 756                                         $jsonldinfo['publisher_url'] = trim($content);
 757                                 }
 758
 759                                 $content = JsonLD::fetchElement($brand, 'url');
 760                                 if (!empty($content) && is_string($content)) {
 761                                         $jsonldinfo['publisher_url'] = trim($content);
 762                                 }
 763
 764                                 $content = JsonLD::fetchElement($brand, 'logo', 'url');
 765                                 if (!empty($content) && is_string($content)) {
 766                                         $jsonldinfo['publisher_img'] = trim($content);
 767                                 }
 768                         }
 769
 770                         $logo = JsonLD::fetchElement($jsonld, 'publisher', 'logo');
 771                         if (!empty($logo) && is_array($logo)) {
 772                                 $content = JsonLD::fetchElement($logo, 'url');
 773                                 if (!empty($content) && is_string($content)) {
 774                                         $jsonldinfo['publisher_img'] = trim($content);
 775                                 }
 776                         }
 777                 } elseif (!empty($jsonld['publisher']) && is_string($jsonld['publisher'])) {
 778                         $jsonldinfo['publisher_name'] = trim($jsonld['publisher']);
 779                 }
 780
 781                 if (!empty($jsonld['author']) && is_array($jsonld['author'])) {
 782                         $content = JsonLD::fetchElement($jsonld, 'author', 'name');
 783                         if (!empty($content) && is_string($content)) {
 784                                 $jsonldinfo['author_name'] = trim($content);
 785                         }
 786
 787                         $content = JsonLD::fetchElement($jsonld, 'author', 'sameAs');
 788                         if (!empty($content) && is_string($content)) {
 789                                 $jsonldinfo['author_url'] = trim($content);
 790                         }
 791
 792                         $content = JsonLD::fetchElement($jsonld, 'author', 'url');
 793                         if (!empty($content) && is_string($content)) {
 794                                 $jsonldinfo['author_url'] = trim($content);
 795                         }
 796
 797                         $logo = JsonLD::fetchElement($jsonld, 'author', 'logo');
 798                         if (!empty($logo) && is_array($logo)) {
 799                                 $content = JsonLD::fetchElement($logo, 'url');
 800                                 if (!empty($content) && is_string($content)) {
 801                                         $jsonldinfo['author_img'] = trim($content);
 802                                 }
 803                         }
 804                 } elseif (!empty($jsonld['author']) && is_string($jsonld['author'])) {
 805                         $jsonldinfo['author_name'] = trim($jsonld['author']);
 806                 }
 807
 808                 Logger::info('Fetched Author information', ['fetched' => $jsonldinfo]);
 809
 810                 return array_merge($siteinfo, $jsonldinfo);
 811         }
 812
 813         /**
 814          * Fetch data from the provided JSON-LD Article type
 815          * @see https://schema.org/Article
 816          *
 817          * @param array $siteinfo
 818          * @param array $jsonld
 819          * @return array siteinfo
 820          */
 821         private static function parseJsonLdArticle(array $siteinfo, array $jsonld)
 822         {
 823                 $jsonldinfo = [];
 824
 825                 $content = JsonLD::fetchElement($jsonld, 'headline');
 826                 if (!empty($content) && is_string($content)) {
 827                         $jsonldinfo['title'] = trim($content);
 828                 }
 829
 830                 $content = JsonLD::fetchElement($jsonld, 'alternativeHeadline');
 831                 if (!empty($content) && is_string($content) && (($jsonldinfo['title'] ?? '') != trim($content))) {
 832                         $jsonldinfo['alternative_title'] = trim($content);
 833                 }
 834
 835                 $content = JsonLD::fetchElement($jsonld, 'description');
 836                 if (!empty($content) && is_string($content)) {
 837                         $jsonldinfo['text'] = trim($content);
 838                 }
 839
 840                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 841                 if (!empty($content)) {
 842                         $jsonldinfo['image'] = trim($content);
 843                 }
 844
 845                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
 846                 if (!empty($content) && is_string($content)) {
 847                         $jsonldinfo['image'] = trim($content);
 848                 }
 849
 850                 if (!empty($jsonld['keywords']) && !is_array($jsonld['keywords'])) {
 851                         $content = JsonLD::fetchElement($jsonld, 'keywords');
 852                         if (!empty($content)) {
 853                                 $siteinfo['keywords'] = [];
 854                                 $keywords = explode(',', $content);
 855                                 foreach ($keywords as $keyword) {
 856                                         $siteinfo['keywords'][] = trim($keyword);
 857                                 }
 858                         }
 859                 } else {
 860                         $content = JsonLD::fetchElementArray($jsonld, 'keywords');
 861                         if (!empty($content) && is_array($content)) {
 862                                 $jsonldinfo['keywords'] = $content;
 863                         }
 864                 }
 865
 866                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 867
 868                 Logger::info('Fetched article information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 869
 870                 return array_merge($siteinfo, $jsonldinfo);
 871         }
 872
 873         /**
 874          * Fetch data from the provided JSON-LD WebPage type
 875          * @see https://schema.org/WebPage
 876          *
 877          * @param array $siteinfo
 878          * @param array $jsonld
 879          * @return array siteinfo
 880          */
 881         private static function parseJsonLdWebPage(array $siteinfo, array $jsonld)
 882         {
 883                 $jsonldinfo = [];
 884
 885                 $content = JsonLD::fetchElement($jsonld, 'name');
 886                 if (!empty($content)) {
 887                         $jsonldinfo['title'] = trim($content);
 888                 }
 889
 890                 $content = JsonLD::fetchElement($jsonld, 'description');
 891                 if (!empty($content)) {
 892                         $jsonldinfo['text'] = trim($content);
 893                 }
 894
 895                 $content = JsonLD::fetchElement($jsonld, 'image');
 896                 if (!empty($content)) {
 897                         $jsonldinfo['image'] = trim($content);
 898                 }
 899
 900                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 901                 if (!empty($content)) {
 902                         $jsonldinfo['image'] = trim($content);
 903                 }
 904
 905                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 906
 907                 Logger::info('Fetched WebPage information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 908
 909                 return array_merge($siteinfo, $jsonldinfo);
 910         }
 911
 912         /**
 913          * Fetch data from the provided JSON-LD WebSite type
 914          * @see https://schema.org/WebSite
 915          *
 916          * @param array $siteinfo
 917          * @param array $jsonld
 918          * @return array siteinfo
 919          */
 920         private static function parseJsonLdWebSite(array $siteinfo, array $jsonld)
 921         {
 922                 $jsonldinfo = [];
 923
 924                 $content = JsonLD::fetchElement($jsonld, 'name');
 925                 if (!empty($content)) {
 926                         $jsonldinfo['publisher_name'] = trim($content);
 927                 }
 928
 929                 $content = JsonLD::fetchElement($jsonld, 'description');
 930                 if (!empty($content)) {
 931                         $jsonldinfo['publisher_description'] = trim($content);
 932                 }
 933
 934                 $content = JsonLD::fetchElement($jsonld, 'url');
 935                 if (!empty($content)) {
 936                         $jsonldinfo['publisher_url'] = trim($content);
 937                 }
 938
 939                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
 940                 if (!empty($content)) {
 941                         $jsonldinfo['image'] = trim($content);
 942                 }
 943
 944                 $jsonldinfo = self::parseJsonLdAuthor($jsonldinfo, $jsonld);
 945
 946                 Logger::info('Fetched WebSite information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 947                 return array_merge($siteinfo, $jsonldinfo);
 948         }
 949
 950         /**
 951          * Fetch data from the provided JSON-LD Organization type
 952          * @see https://schema.org/Organization
 953          *
 954          * @param array $siteinfo
 955          * @param array $jsonld
 956          * @return array siteinfo
 957          */
 958         private static function parseJsonLdWebOrganization(array $siteinfo, array $jsonld)
 959         {
 960                 $jsonldinfo = [];
 961
 962                 $content = JsonLD::fetchElement($jsonld, 'name');
 963                 if (!empty($content)) {
 964                         $jsonldinfo['publisher_name'] = trim($content);
 965                 }
 966
 967                 $content = JsonLD::fetchElement($jsonld, 'description');
 968                 if (!empty($content)) {
 969                         $jsonldinfo['publisher_description'] = trim($content);
 970                 }
 971
 972                 $content = JsonLD::fetchElement($jsonld, 'sameAs');
 973                 if (!empty($content) && is_string($content)) {
 974                         $jsonldinfo['publisher_url'] = trim($content);
 975                 }
 976
 977                 $content = JsonLD::fetchElement($jsonld, 'url');
 978                 if (!empty($content)) {
 979                         $jsonldinfo['publisher_url'] = trim($content);
 980                 }
 981
 982                 $content = JsonLD::fetchElement($jsonld, 'logo', 'url', '@type', 'ImageObject');
 983                 if (!empty($content)) {
 984                         $jsonldinfo['publisher_img'] = trim($content);
 985                 }
 986
 987                 $content = JsonLD::fetchElement($jsonld, 'brand', 'name', '@type', 'Organization');
 988                 if (!empty($content)) {
 989                         $jsonldinfo['publisher_name'] = trim($content);
 990                 }
 991
 992                 $content = JsonLD::fetchElement($jsonld, 'brand', 'url', '@type', 'Organization');
 993                 if (!empty($content)) {
 994                         $jsonldinfo['publisher_url'] = trim($content);
 995                 }
 996
 997                 Logger::info('Fetched Organization information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
 998                 return array_merge($siteinfo, $jsonldinfo);
 999         }
1000
1001         /**
1002          * Fetch data from the provided JSON-LD Person type
1003          * @see https://schema.org/Person
1004          *
1005          * @param array $siteinfo
1006          * @param array $jsonld
1007          * @return array siteinfo
1008          */
1009         private static function parseJsonLdWebPerson(array $siteinfo, array $jsonld)
1010         {
1011                 $jsonldinfo = [];
1012
1013                 $content = JsonLD::fetchElement($jsonld, 'name');
1014                 if (!empty($content)) {
1015                         $jsonldinfo['author_name'] = trim($content);
1016                 }
1017
1018                 $content = JsonLD::fetchElement($jsonld, 'description');
1019                 if (!empty($content)) {
1020                         $jsonldinfo['author_description'] = trim($content);
1021                 }
1022
1023                 $content = JsonLD::fetchElement($jsonld, 'sameAs');
1024                 if (!empty($content) && is_string($content)) {
1025                         $jsonldinfo['author_url'] = trim($content);
1026                 }
1027
1028                 $content = JsonLD::fetchElement($jsonld, 'url');
1029                 if (!empty($content)) {
1030                         $jsonldinfo['author_url'] = trim($content);
1031                 }
1032
1033                 $content = JsonLD::fetchElement($jsonld, 'image', 'url', '@type', 'ImageObject');
1034                 if (!empty($content)) {
1035                         $jsonldinfo['author_img'] = trim($content);
1036                 }
1037
1038                 Logger::info('Fetched Person information', ['url' => $siteinfo['url'], 'fetched' => $jsonldinfo]);
1039                 return array_merge($siteinfo, $jsonldinfo);
1040         }
1041
1042         /**
1043          * Fetch data from the provided JSON-LD MediaObject type
1044          * @see https://schema.org/MediaObject
1045          *
1046          * @param array $siteinfo
1047          * @param array $jsonld
1048          * @return array siteinfo
1049          */
1050         private static function parseJsonLdMediaObject(array $siteinfo, array $jsonld, string $name)
1051         {
1052                 $media = [];
1053
1054                 $content = JsonLD::fetchElement($jsonld, 'caption');
1055                 if (!empty($content)) {
1056                         $media['caption'] = trim($content);
1057                 }
1058
1059                 $content = JsonLD::fetchElement($jsonld, 'url');
1060                 if (!empty($content)) {
1061                         $media['src'] = trim($content);
1062                 }
1063
1064                 $content = JsonLD::fetchElement($jsonld, 'description');
1065                 if (!empty($content)) {
1066                         $media['description'] = trim($content);
1067                 }
1068
1069                 $content = JsonLD::fetchElement($jsonld, 'name');
1070                 if (!empty($content) && (($media['description'] ?? '') != trim($content))) {
1071                         $media['name'] = trim($content);
1072                 }
1073
1074                 $content = JsonLD::fetchElement($jsonld, 'contentUrl');
1075                 if (!empty($content)) {
1076                         $media['content'] = trim($content);
1077                 }
1078
1079                 $content = JsonLD::fetchElement($jsonld, 'embedUrl');
1080                 if (!empty($content)) {
1081                         $media['embed'] = trim($content);
1082                 }
1083
1084                 $content = JsonLD::fetchElement($jsonld, 'height');
1085                 if (!empty($content)) {
1086                         $media['height'] = trim($content);
1087                 }
1088
1089                 $content = JsonLD::fetchElement($jsonld, 'width');
1090                 if (!empty($content)) {
1091                         $media['width'] = trim($content);
1092                 }
1093
1094                 $content = JsonLD::fetchElement($jsonld, 'thumbnailUrl');
1095                 if (!empty($content)) {
1096                         $media['preview'] = trim($content);
1097                 }
1098
1099                 $content = JsonLD::fetchElement($jsonld, 'image');
1100                 if (!empty($content)) {
1101                         $media['image'] = trim($content);
1102                 }
1103
1104                 Logger::info('Fetched Media information', ['url' => $siteinfo['url'], 'fetched' => $media]);
1105                 $siteinfo[$name][] = $media;
1106                 return $siteinfo;
1107         }
1108 }