src/Util/ParseUrl.php

   1 <?php
   2 /**
   3  * @file src/Util/ParseUrl.php
   4  * @brief Get informations about a given URL
   5  */
   6 namespace Friendica\Util;
   7
   8 use DOMDocument;
   9 use DOMXPath;
  10 use Friendica\Content\OEmbed;
  11 use Friendica\Core\Hook;
  12 use Friendica\Core\Logger;
  13 use Friendica\Database\DBA;
  14 use Friendica\Object\Image;
  15
  16 /**
  17  * @brief Class with methods for extracting certain content from an url
  18  */
  19 class ParseUrl
  20 {
  21         /**
  22          * @brief Search for chached embeddable data of an url otherwise fetch it
  23          *
  24          * @param string $url         The url of the page which should be scraped
  25          * @param bool   $no_guessing If true the parse doens't search for
  26          *                            preview pictures
  27          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
  28          *                            to avoid endless loops
  29          *
  30          * @return array which contains needed data for embedding
  31          *    string 'url' => The url of the parsed page
  32          *    string 'type' => Content type
  33          *    string 'title' => The title of the content
  34          *    string 'text' => The description for the content
  35          *    string 'image' => A preview image of the content (only available
  36          *                if $no_geuessing = false
  37          *    array'images' = Array of preview pictures
  38          *    string 'keywords' => The tags which belong to the content
  39          *
  40          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
  41          * @see   ParseUrl::getSiteinfo() for more information about scraping
  42          * embeddable content
  43          */
  44         public static function getSiteinfoCached($url, $no_guessing = false, $do_oembed = true)
  45         {
  46                 if ($url == "") {
  47                         return false;
  48                 }
  49
  50                 $parsed_url = DBA::selectFirst('parsed_url', ['content'],
  51                         ['url' => Strings::normaliseLink($url), 'guessing' => !$no_guessing, 'oembed' => $do_oembed]
  52                 );
  53                 if (!empty($parsed_url['content'])) {
  54                         $data = unserialize($parsed_url['content']);
  55                         return $data;
  56                 }
  57
  58                 $data = self::getSiteinfo($url, $no_guessing, $do_oembed);
  59
  60                 DBA::insert(
  61                         'parsed_url',
  62                         [
  63                                 'url' => Strings::normaliseLink($url), 'guessing' => !$no_guessing,
  64                                 'oembed' => $do_oembed, 'content' => serialize($data),
  65                                 'created' => DateTimeFormat::utcNow()
  66                         ],
  67                         true
  68                 );
  69
  70                 return $data;
  71         }
  72
  73         /**
  74          * @brief Parse a page for embeddable content information
  75          *
  76          * This method parses to url for meta data which can be used to embed
  77          * the content. If available it prioritizes Open Graph meta tags.
  78          * If this is not available it uses the twitter cards meta tags.
  79          * As fallback it uses standard html elements with meta informations
  80          * like \<title\>Awesome Title\</title\> or
  81          * \<meta name="description" content="An awesome description"\>
  82          *
  83          * @param string $url         The url of the page which should be scraped
  84          * @param bool   $no_guessing If true the parse doens't search for
  85          *                            preview pictures
  86          * @param bool   $do_oembed   The false option is used by the function fetch_oembed()
  87          *                            to avoid endless loops
  88          * @param int    $count       Internal counter to avoid endless loops
  89          *
  90          * @return array which contains needed data for embedding
  91          *    string 'url' => The url of the parsed page
  92          *    string 'type' => Content type
  93          *    string 'title' => The title of the content
  94          *    string 'text' => The description for the content
  95          *    string 'image' => A preview image of the content (only available
  96          *                if $no_geuessing = false
  97          *    array'images' = Array of preview pictures
  98          *    string 'keywords' => The tags which belong to the content
  99          *
 100          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 101          * @todo  https://developers.google.com/+/plugins/snippet/
 102          * @verbatim
 103          * <meta itemprop="name" content="Awesome title">
 104          * <meta itemprop="description" content="An awesome description">
 105          * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
 106          *
 107          * <body itemscope itemtype="http://schema.org/Product">
 108          *   <h1 itemprop="name">Shiny Trinket</h1>
 109          *   <img itemprop="image" src="{image-url}" />
 110          *   <p itemprop="description">Shiny trinkets are shiny.</p>
 111          * </body>
 112          * @endverbatim
 113          */
 114         public static function getSiteinfo($url, $no_guessing = false, $do_oembed = true, $count = 1)
 115         {
 116                 $a = \get_app();
 117
 118                 $siteinfo = [];
 119
 120                 // Check if the URL does contain a scheme
 121                 $scheme = parse_url($url, PHP_URL_SCHEME);
 122
 123                 if ($scheme == '') {
 124                         $url = 'http://' . trim($url, '/');
 125                 }
 126
 127                 if ($count > 10) {
 128                         Logger::log('Endless loop detected for ' . $url, Logger::DEBUG);
 129                         return $siteinfo;
 130                 }
 131
 132                 $url = trim($url, "'");
 133                 $url = trim($url, '"');
 134
 135                 $url = Network::stripTrackingQueryParams($url);
 136
 137                 $siteinfo['url'] = $url;
 138                 $siteinfo['type'] = 'link';
 139
 140                 $curlResult = Network::curl($url);
 141                 if (!$curlResult->isSuccess()) {
 142                         return $siteinfo;
 143                 }
 144
 145                 // If the file is too large then exit
 146                 if (defaults($curlResult->getInfo(), 'download_content_length', 0) > 1000000) {
 147                         return $siteinfo;
 148                 }
 149
 150                 // If it isn't a HTML file then exit
 151                 if (($curlResult->getContentType() != '') && !strstr(strtolower($curlResult->getContentType()), 'html')) {
 152                         return $siteinfo;
 153                 }
 154
 155                 $header = $curlResult->getHeader();
 156                 $body = $curlResult->getBody();
 157
 158                 if ($do_oembed) {
 159                         $oembed_data = OEmbed::fetchURL($url);
 160
 161                         if (!empty($oembed_data->type)) {
 162                                 if (!in_array($oembed_data->type, ['error', 'rich', ''])) {
 163                                         $siteinfo['type'] = $oembed_data->type;
 164                                 }
 165
 166                                 // See https://github.com/friendica/friendica/pull/5763#discussion_r217913178
 167                                 if ($siteinfo['type'] != 'photo') {
 168                                         if (isset($oembed_data->title)) {
 169                                                 $siteinfo['title'] = trim($oembed_data->title);
 170                                         }
 171                                         if (isset($oembed_data->description)) {
 172                                                 $siteinfo['text'] = trim($oembed_data->description);
 173                                         }
 174                                         if (isset($oembed_data->thumbnail_url)) {
 175                                                 $siteinfo['image'] = $oembed_data->thumbnail_url;
 176                                         }
 177                                 }
 178                         }
 179                 }
 180
 181                 // Fetch the first mentioned charset. Can be in body or header
 182                 $charset = '';
 183                 if (preg_match('/charset=(.*?)[\'"\s\n]/', $header, $matches)) {
 184                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 185                 }
 186
 187                 if ($charset && strtoupper($charset) != 'UTF-8') {
 188                         // See https://github.com/friendica/friendica/issues/5470#issuecomment-418351211
 189                         $charset = str_ireplace('latin-1', 'latin1', $charset);
 190
 191                         Logger::log('detected charset ' . $charset, Logger::DEBUG);
 192                         $body = iconv($charset, 'UTF-8//TRANSLIT', $body);
 193                 }
 194
 195                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', 'UTF-8');
 196
 197                 $doc = new DOMDocument();
 198                 @$doc->loadHTML($body);
 199
 200                 XML::deleteNode($doc, 'style');
 201                 XML::deleteNode($doc, 'script');
 202                 XML::deleteNode($doc, 'option');
 203                 XML::deleteNode($doc, 'h1');
 204                 XML::deleteNode($doc, 'h2');
 205                 XML::deleteNode($doc, 'h3');
 206                 XML::deleteNode($doc, 'h4');
 207                 XML::deleteNode($doc, 'h5');
 208                 XML::deleteNode($doc, 'h6');
 209                 XML::deleteNode($doc, 'ol');
 210                 XML::deleteNode($doc, 'ul');
 211
 212                 $xpath = new DOMXPath($doc);
 213
 214                 $list = $xpath->query('//meta[@content]');
 215                 foreach ($list as $node) {
 216                         $meta_tag = [];
 217                         if ($node->attributes->length) {
 218                                 foreach ($node->attributes as $attribute) {
 219                                         $meta_tag[$attribute->name] = $attribute->value;
 220                                 }
 221                         }
 222
 223                         if (@$meta_tag['http-equiv'] == 'refresh') {
 224                                 $path = $meta_tag['content'];
 225                                 $pathinfo = explode(';', $path);
 226                                 $content = '';
 227                                 foreach ($pathinfo as $value) {
 228                                         if (substr(strtolower($value), 0, 4) == 'url=') {
 229                                                 $content = substr($value, 4);
 230                                         }
 231                                 }
 232                                 if ($content != '') {
 233                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 234                                         return $siteinfo;
 235                                 }
 236                         }
 237                 }
 238
 239                 $list = $xpath->query('//title');
 240                 if ($list->length > 0) {
 241                         $siteinfo['title'] = trim($list->item(0)->nodeValue);
 242                 }
 243
 244                 $list = $xpath->query('//meta[@name]');
 245                 foreach ($list as $node) {
 246                         $meta_tag = [];
 247                         if ($node->attributes->length) {
 248                                 foreach ($node->attributes as $attribute) {
 249                                         $meta_tag[$attribute->name] = $attribute->value;
 250                                 }
 251                         }
 252
 253                         if (empty($meta_tag['content'])) {
 254                                 continue;
 255                         }
 256
 257                         $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 258
 259                         switch (strtolower($meta_tag['name'])) {
 260                                 case 'fulltitle':
 261                                         $siteinfo['title'] = trim($meta_tag['content']);
 262                                         break;
 263                                 case 'description':
 264                                         $siteinfo['text'] = trim($meta_tag['content']);
 265                                         break;
 266                                 case 'thumbnail':
 267                                         $siteinfo['image'] = $meta_tag['content'];
 268                                         break;
 269                                 case 'twitter:image':
 270                                         $siteinfo['image'] = $meta_tag['content'];
 271                                         break;
 272                                 case 'twitter:image:src':
 273                                         $siteinfo['image'] = $meta_tag['content'];
 274                                         break;
 275                                 case 'twitter:card':
 276                                         // Detect photo pages
 277                                         if ($meta_tag['content'] == 'summary_large_image') {
 278                                                 $siteinfo['type'] = 'photo';
 279                                         }
 280                                         break;
 281                                 case 'twitter:description':
 282                                         $siteinfo['text'] = trim($meta_tag['content']);
 283                                         break;
 284                                 case 'twitter:title':
 285                                         $siteinfo['title'] = trim($meta_tag['content']);
 286                                         break;
 287                                 case 'dc.title':
 288                                         $siteinfo['title'] = trim($meta_tag['content']);
 289                                         break;
 290                                 case 'dc.description':
 291                                         $siteinfo['text'] = trim($meta_tag['content']);
 292                                         break;
 293                                 case 'keywords':
 294                                         $keywords = explode(',', $meta_tag['content']);
 295                                         break;
 296                                 case 'news_keywords':
 297                                         $keywords = explode(',', $meta_tag['content']);
 298                                         break;
 299                         }
 300                 }
 301
 302                 if (isset($keywords)) {
 303                         $siteinfo['keywords'] = [];
 304                         foreach ($keywords as $keyword) {
 305                                 if (!in_array(trim($keyword), $siteinfo['keywords'])) {
 306                                         $siteinfo['keywords'][] = trim($keyword);
 307                                 }
 308                         }
 309                 }
 310
 311                 $list = $xpath->query('//meta[@property]');
 312                 foreach ($list as $node) {
 313                         $meta_tag = [];
 314                         if ($node->attributes->length) {
 315                                 foreach ($node->attributes as $attribute) {
 316                                         $meta_tag[$attribute->name] = $attribute->value;
 317                                 }
 318                         }
 319
 320                         if (!empty($meta_tag['content'])) {
 321                                 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 322
 323                                 switch (strtolower($meta_tag['property'])) {
 324                                         case 'og:image':
 325                                                 $siteinfo['image'] = $meta_tag['content'];
 326                                                 break;
 327                                         case 'og:title':
 328                                                 $siteinfo['title'] = trim($meta_tag['content']);
 329                                                 break;
 330                                         case 'og:description':
 331                                                 $siteinfo['text'] = trim($meta_tag['content']);
 332                                                 break;
 333                                 }
 334                         }
 335                 }
 336
 337                 // Prevent to have a photo type without an image
 338                 if ((empty($siteinfo['image']) || !empty($siteinfo['text'])) && ($siteinfo['type'] == 'photo')) {
 339                         $siteinfo['type'] = 'link';
 340                 }
 341
 342                 if (empty($siteinfo['image']) && !$no_guessing) {
 343                         $list = $xpath->query('//img[@src]');
 344                         foreach ($list as $node) {
 345                                 $img_tag = [];
 346                                 if ($node->attributes->length) {
 347                                         foreach ($node->attributes as $attribute) {
 348                                                 $img_tag[$attribute->name] = $attribute->value;
 349                                         }
 350                                 }
 351
 352                                 $src = self::completeUrl($img_tag['src'], $url);
 353                                 $photodata = Image::getInfoFromURL($src);
 354
 355                                 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
 356                                         if ($photodata[0] > 300) {
 357                                                 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
 358                                                 $photodata[0] = 300;
 359                                         }
 360                                         if ($photodata[1] > 300) {
 361                                                 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
 362                                                 $photodata[1] = 300;
 363                                         }
 364                                         $siteinfo['images'][] = [
 365                                                 'src'    => $src,
 366                                                 'width'  => $photodata[0],
 367                                                 'height' => $photodata[1]
 368                                         ];
 369                                 }
 370                         }
 371                 } elseif (!empty($siteinfo['image'])) {
 372                         $src = self::completeUrl($siteinfo['image'], $url);
 373
 374                         unset($siteinfo['image']);
 375
 376                         $photodata = Image::getInfoFromURL($src);
 377
 378                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 379                                 $siteinfo['images'][] = ['src' => $src,
 380                                         'width' => $photodata[0],
 381                                         'height' => $photodata[1]];
 382                         }
 383                 }
 384
 385                 if ((@$siteinfo['text'] == '') && (@$siteinfo['title'] != '') && !$no_guessing) {
 386                         $text = '';
 387
 388                         $list = $xpath->query('//div[@class="article"]');
 389                         foreach ($list as $node) {
 390                                 if (strlen($node->nodeValue) > 40) {
 391                                         $text .= ' ' . trim($node->nodeValue);
 392                                 }
 393                         }
 394
 395                         if ($text == '') {
 396                                 $list = $xpath->query('//div[@class="content"]');
 397                                 foreach ($list as $node) {
 398                                         if (strlen($node->nodeValue) > 40) {
 399                                                 $text .= ' ' . trim($node->nodeValue);
 400                                         }
 401                                 }
 402                         }
 403
 404                         // If none text was found then take the paragraph content
 405                         if ($text == '') {
 406                                 $list = $xpath->query('//p');
 407                                 foreach ($list as $node) {
 408                                         if (strlen($node->nodeValue) > 40) {
 409                                                 $text .= ' ' . trim($node->nodeValue);
 410                                         }
 411                                 }
 412                         }
 413
 414                         if ($text != '') {
 415                                 $text = trim(str_replace(["\n", "\r"], [' ', ' '], $text));
 416
 417                                 while (strpos($text, '  ')) {
 418                                         $text = trim(str_replace('  ', ' ', $text));
 419                                 }
 420
 421                                 $siteinfo['text'] = trim(html_entity_decode(substr($text, 0, 350), ENT_QUOTES, 'UTF-8') . '...');
 422                         }
 423                 }
 424
 425                 Logger::log('Siteinfo for ' . $url . ' ' . print_r($siteinfo, true), Logger::DEBUG);
 426
 427                 Hook::callAll('getsiteinfo', $siteinfo);
 428
 429                 return $siteinfo;
 430         }
 431
 432         /**
 433          * @brief Convert tags from CSV to an array
 434          *
 435          * @param string $string Tags
 436          * @return array with formatted Hashtags
 437          */
 438         public static function convertTagsToArray($string)
 439         {
 440                 $arr_tags = str_getcsv($string);
 441                 if (count($arr_tags)) {
 442                         // add the # sign to every tag
 443                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 444
 445                         return $arr_tags;
 446                 }
 447         }
 448
 449         /**
 450          * @brief Add a hasht sign to a string
 451          *
 452          *  This method is used as callback function
 453          *
 454          * @param string $tag The pure tag name
 455          * @param int    $k   Counter for internal use
 456          * @return void
 457          */
 458         private static function arrAddHashes(&$tag, $k)
 459         {
 460                 $tag = "#" . $tag;
 461         }
 462
 463         /**
 464          * @brief Add a scheme to an url
 465          *
 466          * The src attribute of some html elements (e.g. images)
 467          * can miss the scheme so we need to add the correct
 468          * scheme
 469          *
 470          * @param string $url    The url which possibly does have
 471          *                       a missing scheme (a link to an image)
 472          * @param string $scheme The url with a correct scheme
 473          *                       (e.g. the url from the webpage which does contain the image)
 474          *
 475          * @return string The url with a scheme
 476          */
 477         private static function completeUrl($url, $scheme)
 478         {
 479                 $urlarr = parse_url($url);
 480
 481                 // If the url does allready have an scheme
 482                 // we can stop the process here
 483                 if (isset($urlarr["scheme"])) {
 484                         return($url);
 485                 }
 486
 487                 $schemearr = parse_url($scheme);
 488
 489                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 490
 491                 if (!empty($schemearr["port"])) {
 492                         $complete .= ":".$schemearr["port"];
 493                 }
 494
 495                 if (!empty($urlarr["path"])) {
 496                         if (strpos($urlarr["path"], "/") !== 0) {
 497                                 $complete .= "/";
 498                         }
 499
 500                         $complete .= $urlarr["path"];
 501                 }
 502
 503                 if (!empty($urlarr["query"])) {
 504                         $complete .= "?".$urlarr["query"];
 505                 }
 506
 507                 if (!empty($urlarr["fragment"])) {
 508                         $complete .= "#".$urlarr["fragment"];
 509                 }
 510
 511                 return($complete);
 512         }
 513 }