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
  31 /**
  32  * Get information about a given URL
  33  *
  34  * Class with methods for extracting certain content from an url
  35  */
  36 class ParseUrl
  37 {
  38         /**
  39          * Maximum number of characters for the description
  40          */
  41         const MAX_DESC_COUNT = 250;
  42
  43         /**
  44          * Minimum number of characters for the description
  45          */
  46         const MIN_DESC_COUNT = 100;
  47
  48         /**
  49          * Search for chached embeddable data of an url otherwise fetch it
  50          *
  51          * @param string $url         The url of the page which should be scraped
  52          * @param bool   $no_guessing If true the parse doens't search for
  53          *                            preview pictures
  54          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
  55          *                            to avoid endless loops
  56          *
  57          * @return array which contains needed data for embedding
  58          *    string 'url' => The url of the parsed page
  59          *    string 'type' => Content type
  60          *    string 'title' => The title of the content
  61          *    string 'text' => The description for the content
  62          *    string 'image' => A preview image of the content (only available
  63          *                if $no_geuessing = false
  64          *    array'images' = Array of preview pictures
  65          *    string 'keywords' => The tags which belong to the content
  66          *
  67          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
  68          * @see   ParseUrl::getSiteinfo() for more information about scraping
  69          * embeddable content
  70          */
  71         public static function getSiteinfoCached($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, $do_oembed);
  86
  87                 DBA::insert(
  88                         'parsed_url',
  89                         [
  90                                 'url' => Strings::normaliseLink($url), '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' => The title of the content
 121          *    string 'text' => The description for the content
 122          *    string 'image' => A preview image of the content (only available
 123          *                if $no_geuessing = false
 124          *    array'images' = Array of preview pictures
 125          *    string 'keywords' => The tags which belong to the content
 126          *
 127          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 128          * @todo  https://developers.google.com/+/plugins/snippet/
 129          * @verbatim
 130          * <meta itemprop="name" content="Awesome title">
 131          * <meta itemprop="description" content="An awesome description">
 132          * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
 133          *
 134          * <body itemscope itemtype="http://schema.org/Product">
 135          *   <h1 itemprop="name">Shiny Trinket</h1>
 136          *   <img itemprop="image" src="{image-url}" />
 137          *   <p itemprop="description">Shiny trinkets are shiny.</p>
 138          * </body>
 139          * @endverbatim
 140          */
 141         public static function getSiteinfo($url, $no_guessing = false, $do_oembed = true, $count = 1)
 142         {
 143                 $siteinfo = [];
 144
 145                 // Check if the URL does contain a scheme
 146                 $scheme = parse_url($url, PHP_URL_SCHEME);
 147
 148                 if ($scheme == '') {
 149                         $url = 'http://' . trim($url, '/');
 150                 }
 151
 152                 if ($count > 10) {
 153                         Logger::log('Endless loop detected for ' . $url, Logger::DEBUG);
 154                         return $siteinfo;
 155                 }
 156
 157                 $url = trim($url, "'");
 158                 $url = trim($url, '"');
 159
 160                 $url = Network::stripTrackingQueryParams($url);
 161
 162                 $siteinfo['url'] = $url;
 163                 $siteinfo['type'] = 'link';
 164
 165                 $curlResult = Network::curl($url);
 166                 if (!$curlResult->isSuccess()) {
 167                         return $siteinfo;
 168                 }
 169
 170                 // If the file is too large then exit
 171                 if (($curlResult->getInfo()['download_content_length'] ?? 0) > 1000000) {
 172                         return $siteinfo;
 173                 }
 174
 175                 // If it isn't a HTML file then exit
 176                 if (($curlResult->getContentType() != '') && !strstr(strtolower($curlResult->getContentType()), 'html')) {
 177                         return $siteinfo;
 178                 }
 179
 180                 $header = $curlResult->getHeader();
 181                 $body = $curlResult->getBody();
 182
 183                 if ($do_oembed) {
 184                         $oembed_data = OEmbed::fetchURL($url);
 185
 186                         if (!empty($oembed_data->type)) {
 187                                 if (!in_array($oembed_data->type, ['error', 'rich', ''])) {
 188                                         $siteinfo['type'] = $oembed_data->type;
 189                                 }
 190
 191                                 // See https://github.com/friendica/friendica/pull/5763#discussion_r217913178
 192                                 if ($siteinfo['type'] != 'photo') {
 193                                         if (isset($oembed_data->title)) {
 194                                                 $siteinfo['title'] = trim($oembed_data->title);
 195                                         }
 196                                         if (isset($oembed_data->description)) {
 197                                                 $siteinfo['text'] = trim($oembed_data->description);
 198                                         }
 199                                         if (isset($oembed_data->thumbnail_url)) {
 200                                                 $siteinfo['image'] = $oembed_data->thumbnail_url;
 201                                         }
 202                                 }
 203                         }
 204                 }
 205
 206                 // Fetch the first mentioned charset. Can be in body or header
 207                 $charset = '';
 208                 if (preg_match('/charset=(.*?)[\'"\s\n]/', $header, $matches)) {
 209                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 210                 }
 211
 212                 if ($charset && strtoupper($charset) != 'UTF-8') {
 213                         // See https://github.com/friendica/friendica/issues/5470#issuecomment-418351211
 214                         $charset = str_ireplace('latin-1', 'latin1', $charset);
 215
 216                         Logger::log('detected charset ' . $charset, Logger::DEBUG);
 217                         $body = iconv($charset, 'UTF-8//TRANSLIT', $body);
 218                 }
 219
 220                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', 'UTF-8');
 221
 222                 $doc = new DOMDocument();
 223                 @$doc->loadHTML($body);
 224
 225                 XML::deleteNode($doc, 'style');
 226                 XML::deleteNode($doc, 'script');
 227                 XML::deleteNode($doc, 'option');
 228                 XML::deleteNode($doc, 'h1');
 229                 XML::deleteNode($doc, 'h2');
 230                 XML::deleteNode($doc, 'h3');
 231                 XML::deleteNode($doc, 'h4');
 232                 XML::deleteNode($doc, 'h5');
 233                 XML::deleteNode($doc, 'h6');
 234                 XML::deleteNode($doc, 'ol');
 235                 XML::deleteNode($doc, 'ul');
 236
 237                 $xpath = new DOMXPath($doc);
 238
 239                 $list = $xpath->query('//meta[@content]');
 240                 foreach ($list as $node) {
 241                         $meta_tag = [];
 242                         if ($node->attributes->length) {
 243                                 foreach ($node->attributes as $attribute) {
 244                                         $meta_tag[$attribute->name] = $attribute->value;
 245                                 }
 246                         }
 247
 248                         if (@$meta_tag['http-equiv'] == 'refresh') {
 249                                 $path = $meta_tag['content'];
 250                                 $pathinfo = explode(';', $path);
 251                                 $content = '';
 252                                 foreach ($pathinfo as $value) {
 253                                         if (substr(strtolower($value), 0, 4) == 'url=') {
 254                                                 $content = substr($value, 4);
 255                                         }
 256                                 }
 257                                 if ($content != '') {
 258                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 259                                         return $siteinfo;
 260                                 }
 261                         }
 262                 }
 263
 264                 $list = $xpath->query('//title');
 265                 if ($list->length > 0) {
 266                         $siteinfo['title'] = trim($list->item(0)->nodeValue);
 267                 }
 268
 269                 $list = $xpath->query('//meta[@name]');
 270                 foreach ($list as $node) {
 271                         $meta_tag = [];
 272                         if ($node->attributes->length) {
 273                                 foreach ($node->attributes as $attribute) {
 274                                         $meta_tag[$attribute->name] = $attribute->value;
 275                                 }
 276                         }
 277
 278                         if (empty($meta_tag['content'])) {
 279                                 continue;
 280                         }
 281
 282                         $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 283
 284                         switch (strtolower($meta_tag['name'])) {
 285                                 case 'fulltitle':
 286                                         $siteinfo['title'] = trim($meta_tag['content']);
 287                                         break;
 288                                 case 'description':
 289                                         $siteinfo['text'] = trim($meta_tag['content']);
 290                                         break;
 291                                 case 'thumbnail':
 292                                         $siteinfo['image'] = $meta_tag['content'];
 293                                         break;
 294                                 case 'twitter:image':
 295                                         $siteinfo['image'] = $meta_tag['content'];
 296                                         break;
 297                                 case 'twitter:image:src':
 298                                         $siteinfo['image'] = $meta_tag['content'];
 299                                         break;
 300                                 case 'twitter:card':
 301                                         // Detect photo pages
 302                                         if ($meta_tag['content'] == 'summary_large_image') {
 303                                                 $siteinfo['type'] = 'photo';
 304                                         }
 305                                         break;
 306                                 case 'twitter:description':
 307                                         $siteinfo['text'] = trim($meta_tag['content']);
 308                                         break;
 309                                 case 'twitter:title':
 310                                         $siteinfo['title'] = trim($meta_tag['content']);
 311                                         break;
 312                                 case 'dc.title':
 313                                         $siteinfo['title'] = trim($meta_tag['content']);
 314                                         break;
 315                                 case 'dc.description':
 316                                         $siteinfo['text'] = trim($meta_tag['content']);
 317                                         break;
 318                                 case 'keywords':
 319                                         $keywords = explode(',', $meta_tag['content']);
 320                                         break;
 321                                 case 'news_keywords':
 322                                         $keywords = explode(',', $meta_tag['content']);
 323                                         break;
 324                         }
 325                 }
 326
 327                 if (isset($keywords)) {
 328                         $siteinfo['keywords'] = [];
 329                         foreach ($keywords as $keyword) {
 330                                 if (!in_array(trim($keyword), $siteinfo['keywords'])) {
 331                                         $siteinfo['keywords'][] = trim($keyword);
 332                                 }
 333                         }
 334                 }
 335
 336                 $list = $xpath->query('//meta[@property]');
 337                 foreach ($list as $node) {
 338                         $meta_tag = [];
 339                         if ($node->attributes->length) {
 340                                 foreach ($node->attributes as $attribute) {
 341                                         $meta_tag[$attribute->name] = $attribute->value;
 342                                 }
 343                         }
 344
 345                         if (!empty($meta_tag['content'])) {
 346                                 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 347
 348                                 switch (strtolower($meta_tag['property'])) {
 349                                         case 'og:image':
 350                                                 $siteinfo['image'] = $meta_tag['content'];
 351                                                 break;
 352                                         case 'og:title':
 353                                                 $siteinfo['title'] = trim($meta_tag['content']);
 354                                                 break;
 355                                         case 'og:description':
 356                                                 $siteinfo['text'] = trim($meta_tag['content']);
 357                                                 break;
 358                                 }
 359                         }
 360                 }
 361
 362                 // Prevent to have a photo type without an image
 363                 if ((empty($siteinfo['image']) || !empty($siteinfo['text'])) && ($siteinfo['type'] == 'photo')) {
 364                         $siteinfo['type'] = 'link';
 365                 }
 366
 367                 if (!empty($siteinfo['image'])) {
 368                         $src = self::completeUrl($siteinfo['image'], $url);
 369
 370                         unset($siteinfo['image']);
 371
 372                         $photodata = Images::getInfoFromURLCached($src);
 373
 374                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 375                                 $siteinfo['images'][] = ['src' => $src,
 376                                         'width' => $photodata[0],
 377                                         'height' => $photodata[1]];
 378                         }
 379                 }
 380
 381                 if (!empty($siteinfo['text']) && mb_strlen($siteinfo['text']) > self::MAX_DESC_COUNT) {
 382                         $siteinfo['text'] = mb_substr($siteinfo['text'], 0, self::MAX_DESC_COUNT) . '…';
 383                         $pos = mb_strrpos($siteinfo['text'], '.');
 384                         if ($pos > self::MIN_DESC_COUNT) {
 385                                 $siteinfo['text'] = mb_substr($siteinfo['text'], 0, $pos + 1);
 386                         }
 387                 }
 388
 389                 Logger::info('Siteinfo fetched', ['url' => $url, 'siteinfo' => $siteinfo]);
 390
 391                 Hook::callAll('getsiteinfo', $siteinfo);
 392
 393                 return $siteinfo;
 394         }
 395
 396         /**
 397          * Convert tags from CSV to an array
 398          *
 399          * @param string $string Tags
 400          * @return array with formatted Hashtags
 401          */
 402         public static function convertTagsToArray($string)
 403         {
 404                 $arr_tags = str_getcsv($string);
 405                 if (count($arr_tags)) {
 406                         // add the # sign to every tag
 407                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 408
 409                         return $arr_tags;
 410                 }
 411         }
 412
 413         /**
 414          * Add a hasht sign to a string
 415          *
 416          * This method is used as callback function
 417          *
 418          * @param string $tag The pure tag name
 419          * @param int    $k   Counter for internal use
 420          * @return void
 421          */
 422         private static function arrAddHashes(&$tag, $k)
 423         {
 424                 $tag = "#" . $tag;
 425         }
 426
 427         /**
 428          * Add a scheme to an url
 429          *
 430          * The src attribute of some html elements (e.g. images)
 431          * can miss the scheme so we need to add the correct
 432          * scheme
 433          *
 434          * @param string $url    The url which possibly does have
 435          *                       a missing scheme (a link to an image)
 436          * @param string $scheme The url with a correct scheme
 437          *                       (e.g. the url from the webpage which does contain the image)
 438          *
 439          * @return string The url with a scheme
 440          */
 441         private static function completeUrl($url, $scheme)
 442         {
 443                 $urlarr = parse_url($url);
 444
 445                 // If the url does allready have an scheme
 446                 // we can stop the process here
 447                 if (isset($urlarr["scheme"])) {
 448                         return($url);
 449                 }
 450
 451                 $schemearr = parse_url($scheme);
 452
 453                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 454
 455                 if (!empty($schemearr["port"])) {
 456                         $complete .= ":".$schemearr["port"];
 457                 }
 458
 459                 if (!empty($urlarr["path"])) {
 460                         if (strpos($urlarr["path"], "/") !== 0) {
 461                                 $complete .= "/";
 462                         }
 463
 464                         $complete .= $urlarr["path"];
 465                 }
 466
 467                 if (!empty($urlarr["query"])) {
 468                         $complete .= "?".$urlarr["query"];
 469                 }
 470
 471                 if (!empty($urlarr["fragment"])) {
 472                         $complete .= "#".$urlarr["fragment"];
 473                 }
 474
 475                 return($complete);
 476         }
 477 }