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\DBA;
  30 use Friendica\DI;
  31
  32 /**
  33  * Get information about a given URL
  34  *
  35  * Class with methods for extracting certain content from an url
  36  */
  37 class ParseUrl
  38 {
  39         /**
  40          * Maximum number of characters for the description
  41          */
  42         const MAX_DESC_COUNT = 250;
  43
  44         /**
  45          * Minimum number of characters for the description
  46          */
  47         const MIN_DESC_COUNT = 100;
  48
  49         /**
  50          * Search for chached embeddable data of an url otherwise fetch it
  51          *
  52          * @param string $url         The url of the page which should be scraped
  53          * @param bool   $no_guessing If true the parse doens't search for
  54          *                            preview pictures
  55          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
  56          *                            to avoid endless loops
  57          *
  58          * @return array which contains needed data for embedding
  59          *    string 'url'      => The url of the parsed page
  60          *    string 'type'     => Content type
  61          *    string 'title'    => (optional) The title of the content
  62          *    string 'text'     => (optional) The description for the content
  63          *    string 'image'    => (optional) A preview image of the content (only available if $no_geuessing = false)
  64          *    array  'images'   => (optional) Array of preview pictures
  65          *    string 'keywords' => (optional) The tags which belong to the content
  66          *
  67          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
  68          * @see   ParseUrl::getSiteinfo() for more information about scraping
  69          * embeddable content
  70          */
  71         public static function getSiteinfoCached($url, $no_guessing = false, $do_oembed = true)
  72         {
  73                 if ($url == "") {
  74                         return false;
  75                 }
  76
  77                 $parsed_url = DBA::selectFirst('parsed_url', ['content'],
  78                         ['url' => Strings::normaliseLink($url), 'guessing' => !$no_guessing, 'oembed' => $do_oembed]
  79                 );
  80                 if (!empty($parsed_url['content'])) {
  81                         $data = unserialize($parsed_url['content']);
  82                         return $data;
  83                 }
  84
  85                 $data = self::getSiteinfo($url, $no_guessing, $do_oembed);
  86
  87                 DBA::insert(
  88                         'parsed_url',
  89                         [
  90                                 'url' => substr(Strings::normaliseLink($url), 0, 255), 'guessing' => !$no_guessing,
  91                                 'oembed' => $do_oembed, 'content' => serialize($data),
  92                                 'created' => DateTimeFormat::utcNow()
  93                         ],
  94                         true
  95                 );
  96
  97                 return $data;
  98         }
  99
 100         /**
 101          * Parse a page for embeddable content information
 102          *
 103          * This method parses to url for meta data which can be used to embed
 104          * the content. If available it prioritizes Open Graph meta tags.
 105          * If this is not available it uses the twitter cards meta tags.
 106          * As fallback it uses standard html elements with meta informations
 107          * like \<title\>Awesome Title\</title\> or
 108          * \<meta name="description" content="An awesome description"\>
 109          *
 110          * @param string $url         The url of the page which should be scraped
 111          * @param bool   $no_guessing If true the parse doens't search for
 112          *                            preview pictures
 113          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
 114          *                            to avoid endless loops
 115          * @param int    $count       Internal counter to avoid endless loops
 116          *
 117          * @return array which contains needed data for embedding
 118          *    string 'url'      => The url of the parsed page
 119          *    string 'type'     => Content type
 120          *    string 'title'    => (optional) The title of the content
 121          *    string 'text'     => (optional) The description for the content
 122          *    string 'image'    => (optional) A preview image of the content (only available if $no_guessing = false)
 123          *    array  'images'   => (optional) Array of preview pictures
 124          *    string 'keywords' => (optional) The tags which belong to the content
 125          *
 126          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 127          * @todo  https://developers.google.com/+/plugins/snippet/
 128          * @verbatim
 129          * <meta itemprop="name" content="Awesome title">
 130          * <meta itemprop="description" content="An awesome description">
 131          * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
 132          *
 133          * <body itemscope itemtype="http://schema.org/Product">
 134          *   <h1 itemprop="name">Shiny Trinket</h1>
 135          *   <img itemprop="image" src="{image-url}" />
 136          *   <p itemprop="description">Shiny trinkets are shiny.</p>
 137          * </body>
 138          * @endverbatim
 139          */
 140         public static function getSiteinfo($url, $no_guessing = false, $do_oembed = true, $count = 1)
 141         {
 142                 // Check if the URL does contain a scheme
 143                 $scheme = parse_url($url, PHP_URL_SCHEME);
 144
 145                 if ($scheme == '') {
 146                         $url = 'http://' . ltrim($url, '/');
 147                 }
 148
 149                 $url = trim($url, "'\"");
 150
 151                 $url = Network::stripTrackingQueryParams($url);
 152
 153                 $siteinfo = [
 154                         'url' => $url,
 155                         'type' => 'link',
 156                 ];
 157
 158                 if ($count > 10) {
 159                         Logger::log('Endless loop detected for ' . $url, Logger::DEBUG);
 160                         return $siteinfo;
 161                 }
 162
 163                 $curlResult = DI::httpRequest()->get($url);
 164                 if (!$curlResult->isSuccess()) {
 165                         return $siteinfo;
 166                 }
 167
 168                 // If the file is too large then exit
 169                 if (($curlResult->getInfo()['download_content_length'] ?? 0) > 1000000) {
 170                         return $siteinfo;
 171                 }
 172
 173                 // If it isn't a HTML file then exit
 174                 if (($curlResult->getContentType() != '') && !strstr(strtolower($curlResult->getContentType()), 'html')) {
 175                         return $siteinfo;
 176                 }
 177
 178                 $header = $curlResult->getHeader();
 179                 $body = $curlResult->getBody();
 180
 181                 if ($do_oembed) {
 182                         $oembed_data = OEmbed::fetchURL($url);
 183
 184                         if (!empty($oembed_data->type)) {
 185                                 if (!in_array($oembed_data->type, ['error', 'rich', ''])) {
 186                                         $siteinfo['type'] = $oembed_data->type;
 187                                 }
 188
 189                                 // See https://github.com/friendica/friendica/pull/5763#discussion_r217913178
 190                                 if ($siteinfo['type'] != 'photo') {
 191                                         if (isset($oembed_data->title)) {
 192                                                 $siteinfo['title'] = trim($oembed_data->title);
 193                                         }
 194                                         if (isset($oembed_data->description)) {
 195                                                 $siteinfo['text'] = trim($oembed_data->description);
 196                                         }
 197                                         if (isset($oembed_data->thumbnail_url)) {
 198                                                 $siteinfo['image'] = $oembed_data->thumbnail_url;
 199                                         }
 200                                 }
 201                         }
 202                 }
 203
 204                 // Fetch the first mentioned charset. Can be in body or header
 205                 $charset = '';
 206                 if (preg_match('/charset=(.*?)[\'"\s\n]/', $header, $matches)) {
 207                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 208                 }
 209
 210                 if ($charset && strtoupper($charset) != 'UTF-8') {
 211                         // See https://github.com/friendica/friendica/issues/5470#issuecomment-418351211
 212                         $charset = str_ireplace('latin-1', 'latin1', $charset);
 213
 214                         Logger::log('detected charset ' . $charset, Logger::DEBUG);
 215                         $body = iconv($charset, 'UTF-8//TRANSLIT', $body);
 216                 }
 217
 218                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', 'UTF-8');
 219
 220                 $doc = new DOMDocument();
 221                 @$doc->loadHTML($body);
 222
 223                 XML::deleteNode($doc, 'style');
 224                 XML::deleteNode($doc, 'script');
 225                 XML::deleteNode($doc, 'option');
 226                 XML::deleteNode($doc, 'h1');
 227                 XML::deleteNode($doc, 'h2');
 228                 XML::deleteNode($doc, 'h3');
 229                 XML::deleteNode($doc, 'h4');
 230                 XML::deleteNode($doc, 'h5');
 231                 XML::deleteNode($doc, 'h6');
 232                 XML::deleteNode($doc, 'ol');
 233                 XML::deleteNode($doc, 'ul');
 234
 235                 $xpath = new DOMXPath($doc);
 236
 237                 $list = $xpath->query('//meta[@content]');
 238                 foreach ($list as $node) {
 239                         $meta_tag = [];
 240                         if ($node->attributes->length) {
 241                                 foreach ($node->attributes as $attribute) {
 242                                         $meta_tag[$attribute->name] = $attribute->value;
 243                                 }
 244                         }
 245
 246                         if (@$meta_tag['http-equiv'] == 'refresh') {
 247                                 $path = $meta_tag['content'];
 248                                 $pathinfo = explode(';', $path);
 249                                 $content = '';
 250                                 foreach ($pathinfo as $value) {
 251                                         if (substr(strtolower($value), 0, 4) == 'url=') {
 252                                                 $content = substr($value, 4);
 253                                         }
 254                                 }
 255                                 if ($content != '') {
 256                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 257                                         return $siteinfo;
 258                                 }
 259                         }
 260                 }
 261
 262                 $list = $xpath->query('//title');
 263                 if ($list->length > 0) {
 264                         $siteinfo['title'] = trim($list->item(0)->nodeValue);
 265                 }
 266
 267                 $list = $xpath->query('//meta[@name]');
 268                 foreach ($list as $node) {
 269                         $meta_tag = [];
 270                         if ($node->attributes->length) {
 271                                 foreach ($node->attributes as $attribute) {
 272                                         $meta_tag[$attribute->name] = $attribute->value;
 273                                 }
 274                         }
 275
 276                         if (empty($meta_tag['content'])) {
 277                                 continue;
 278                         }
 279
 280                         $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 281
 282                         switch (strtolower($meta_tag['name'])) {
 283                                 case 'fulltitle':
 284                                         $siteinfo['title'] = trim($meta_tag['content']);
 285                                         break;
 286                                 case 'description':
 287                                         $siteinfo['text'] = trim($meta_tag['content']);
 288                                         break;
 289                                 case 'thumbnail':
 290                                         $siteinfo['image'] = $meta_tag['content'];
 291                                         break;
 292                                 case 'twitter:image':
 293                                         $siteinfo['image'] = $meta_tag['content'];
 294                                         break;
 295                                 case 'twitter:image:src':
 296                                         $siteinfo['image'] = $meta_tag['content'];
 297                                         break;
 298                                 case 'twitter:card':
 299                                         // Detect photo pages
 300                                         if ($meta_tag['content'] == 'summary_large_image') {
 301                                                 $siteinfo['type'] = 'photo';
 302                                         }
 303                                         break;
 304                                 case 'twitter:description':
 305                                         $siteinfo['text'] = trim($meta_tag['content']);
 306                                         break;
 307                                 case 'twitter:title':
 308                                         $siteinfo['title'] = trim($meta_tag['content']);
 309                                         break;
 310                                 case 'dc.title':
 311                                         $siteinfo['title'] = trim($meta_tag['content']);
 312                                         break;
 313                                 case 'dc.description':
 314                                         $siteinfo['text'] = trim($meta_tag['content']);
 315                                         break;
 316                                 case 'keywords':
 317                                         $keywords = explode(',', $meta_tag['content']);
 318                                         break;
 319                                 case 'news_keywords':
 320                                         $keywords = explode(',', $meta_tag['content']);
 321                                         break;
 322                         }
 323                 }
 324
 325                 if (isset($keywords)) {
 326                         $siteinfo['keywords'] = [];
 327                         foreach ($keywords as $keyword) {
 328                                 if (!in_array(trim($keyword), $siteinfo['keywords'])) {
 329                                         $siteinfo['keywords'][] = trim($keyword);
 330                                 }
 331                         }
 332                 }
 333
 334                 $list = $xpath->query('//meta[@property]');
 335                 foreach ($list as $node) {
 336                         $meta_tag = [];
 337                         if ($node->attributes->length) {
 338                                 foreach ($node->attributes as $attribute) {
 339                                         $meta_tag[$attribute->name] = $attribute->value;
 340                                 }
 341                         }
 342
 343                         if (!empty($meta_tag['content'])) {
 344                                 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 345
 346                                 switch (strtolower($meta_tag['property'])) {
 347                                         case 'og:image':
 348                                                 $siteinfo['image'] = $meta_tag['content'];
 349                                                 break;
 350                                         case 'og:title':
 351                                                 $siteinfo['title'] = trim($meta_tag['content']);
 352                                                 break;
 353                                         case 'og:description':
 354                                                 $siteinfo['text'] = trim($meta_tag['content']);
 355                                                 break;
 356                                 }
 357                         }
 358                 }
 359
 360                 // Prevent to have a photo type without an image
 361                 if ((empty($siteinfo['image']) || !empty($siteinfo['text'])) && ($siteinfo['type'] == 'photo')) {
 362                         $siteinfo['type'] = 'link';
 363                 }
 364
 365                 if (!empty($siteinfo['image'])) {
 366                         $src = self::completeUrl($siteinfo['image'], $url);
 367
 368                         unset($siteinfo['image']);
 369
 370                         $photodata = Images::getInfoFromURLCached($src);
 371
 372                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 373                                 $siteinfo['images'][] = ['src' => $src,
 374                                         'width' => $photodata[0],
 375                                         'height' => $photodata[1]];
 376                         }
 377                 }
 378
 379                 if (!empty($siteinfo['text']) && mb_strlen($siteinfo['text']) > self::MAX_DESC_COUNT) {
 380                         $siteinfo['text'] = mb_substr($siteinfo['text'], 0, self::MAX_DESC_COUNT) . '…';
 381                         $pos = mb_strrpos($siteinfo['text'], '.');
 382                         if ($pos > self::MIN_DESC_COUNT) {
 383                                 $siteinfo['text'] = mb_substr($siteinfo['text'], 0, $pos + 1);
 384                         }
 385                 }
 386
 387                 Logger::info('Siteinfo fetched', ['url' => $url, 'siteinfo' => $siteinfo]);
 388
 389                 Hook::callAll('getsiteinfo', $siteinfo);
 390
 391                 return $siteinfo;
 392         }
 393
 394         /**
 395          * Convert tags from CSV to an array
 396          *
 397          * @param string $string Tags
 398          * @return array with formatted Hashtags
 399          */
 400         public static function convertTagsToArray($string)
 401         {
 402                 $arr_tags = str_getcsv($string);
 403                 if (count($arr_tags)) {
 404                         // add the # sign to every tag
 405                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 406
 407                         return $arr_tags;
 408                 }
 409         }
 410
 411         /**
 412          * Add a hasht sign to a string
 413          *
 414          * This method is used as callback function
 415          *
 416          * @param string $tag The pure tag name
 417          * @param int    $k   Counter for internal use
 418          * @return void
 419          */
 420         private static function arrAddHashes(&$tag, $k)
 421         {
 422                 $tag = "#" . $tag;
 423         }
 424
 425         /**
 426          * Add a scheme to an url
 427          *
 428          * The src attribute of some html elements (e.g. images)
 429          * can miss the scheme so we need to add the correct
 430          * scheme
 431          *
 432          * @param string $url    The url which possibly does have
 433          *                       a missing scheme (a link to an image)
 434          * @param string $scheme The url with a correct scheme
 435          *                       (e.g. the url from the webpage which does contain the image)
 436          *
 437          * @return string The url with a scheme
 438          */
 439         private static function completeUrl($url, $scheme)
 440         {
 441                 $urlarr = parse_url($url);
 442
 443                 // If the url does allready have an scheme
 444                 // we can stop the process here
 445                 if (isset($urlarr["scheme"])) {
 446                         return($url);
 447                 }
 448
 449                 $schemearr = parse_url($scheme);
 450
 451                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 452
 453                 if (!empty($schemearr["port"])) {
 454                         $complete .= ":".$schemearr["port"];
 455                 }
 456
 457                 if (!empty($urlarr["path"])) {
 458                         if (strpos($urlarr["path"], "/") !== 0) {
 459                                 $complete .= "/";
 460                         }
 461
 462                         $complete .= $urlarr["path"];
 463                 }
 464
 465                 if (!empty($urlarr["query"])) {
 466                         $complete .= "?".$urlarr["query"];
 467                 }
 468
 469                 if (!empty($urlarr["fragment"])) {
 470                         $complete .= "#".$urlarr["fragment"];
 471                 }
 472
 473                 return($complete);
 474         }
 475 }