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\Addon;
  12 use Friendica\Core\Logger;
  13 use Friendica\Database\DBA;
  14 use Friendica\Object\Image;
  15 use Friendica\Util\Strings;
  16
  17 require_once 'include/dba.php';
  18
  19 /**
  20  * @brief Class with methods for extracting certain content from an url
  21  */
  22 class ParseUrl
  23 {
  24         /**
  25          * @brief Search for chached embeddable data of an url otherwise fetch it
  26          *
  27          * @param string $url         The url of the page which should be scraped
  28          * @param bool $no_guessing If true the parse doens't search for
  29          *                          preview pictures
  30          * @param bool $do_oembed   The false option is used by the function fetch_oembed()
  31          *                          to avoid endless loops
  32          *
  33          * @return array which contains needed data for embedding
  34          *    string 'url' => The url of the parsed page
  35          *    string 'type' => Content type
  36          *    string 'title' => The title of the content
  37          *    string 'text' => The description for the content
  38          *    string 'image' => A preview image of the content (only available
  39          *                if $no_geuessing = false
  40          *    array'images' = Array of preview pictures
  41          *    string 'keywords' => The tags which belong to the content
  42          *
  43          * @see ParseUrl::getSiteinfo() for more information about scraping
  44          * embeddable content
  45          */
  46         public static function getSiteinfoCached($url, $no_guessing = false, $do_oembed = true)
  47         {
  48                 if ($url == "") {
  49                         return false;
  50                 }
  51
  52                 $parsed_url = DBA::selectFirst('parsed_url', ['content'],
  53                         ['url' => Strings::normaliseLink($url), 'guessing' => !$no_guessing, 'oembed' => $do_oembed]
  54                 );
  55                 if (!empty($parsed_url['content'])) {
  56                         $data = unserialize($parsed_url['content']);
  57                         return $data;
  58                 }
  59
  60                 $data = self::getSiteinfo($url, $no_guessing, $do_oembed);
  61
  62                 DBA::insert(
  63                         'parsed_url',
  64                         [
  65                                 'url' => Strings::normaliseLink($url), 'guessing' => !$no_guessing,
  66                                 'oembed' => $do_oembed, 'content' => serialize($data),
  67                                 'created' => DateTimeFormat::utcNow()
  68                         ],
  69                         true
  70                 );
  71
  72                 return $data;
  73         }
  74         /**
  75          * @brief Parse a page for embeddable content information
  76          *
  77          * This method parses to url for meta data which can be used to embed
  78          * the content. If available it prioritizes Open Graph meta tags.
  79          * If this is not available it uses the twitter cards meta tags.
  80          * As fallback it uses standard html elements with meta informations
  81          * like \<title\>Awesome Title\</title\> or
  82          * \<meta name="description" content="An awesome description"\>
  83          *
  84          * @param string $url         The url of the page which should be scraped
  85          * @param bool $no_guessing If true the parse doens't search for
  86          *                          preview pictures
  87          * @param bool $do_oembed   The false option is used by the function fetch_oembed()
  88          *                          to avoid endless loops
  89          * @param int $count       Internal counter to avoid endless loops
  90          *
  91          * @return array which contains needed data for embedding
  92          *    string 'url' => The url of the parsed page
  93          *    string 'type' => Content type
  94          *    string 'title' => The title of the content
  95          *    string 'text' => The description for the content
  96          *    string 'image' => A preview image of the content (only available
  97          *                if $no_geuessing = false
  98          *    array'images' = Array of preview pictures
  99          *    string 'keywords' => The tags which belong to the content
 100          *
 101          * @todo https://developers.google.com/+/plugins/snippet/
 102          * @verbatim
 103          * <meta itemprop="name" content="Awesome title">
 104          * <meta itemprop="description" content="An awesome description">
 105          * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
 106          *
 107          * <body itemscope itemtype="http://schema.org/Product">
 108          *   <h1 itemprop="name">Shiny Trinket</h1>
 109          *   <img itemprop="image" src="{image-url}" />
 110          *   <p itemprop="description">Shiny trinkets are shiny.</p>
 111          * </body>
 112          * @endverbatim
 113          */
 114         public static function getSiteinfo($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 == '') {
 188                         $charset = 'utf-8';
 189                 }
 190
 191                 if (($charset != '') && (strtoupper($charset) != 'UTF-8')) {
 192                         Logger::log('detected charset ' . $charset, Logger::DEBUG);
 193                         $body = iconv($charset, 'UTF-8//TRANSLIT', $body);
 194                 }
 195
 196                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', 'UTF-8');
 197
 198                 $doc = new DOMDocument();
 199                 @$doc->loadHTML($body);
 200
 201                 XML::deleteNode($doc, 'style');
 202                 XML::deleteNode($doc, 'script');
 203                 XML::deleteNode($doc, 'option');
 204                 XML::deleteNode($doc, 'h1');
 205                 XML::deleteNode($doc, 'h2');
 206                 XML::deleteNode($doc, 'h3');
 207                 XML::deleteNode($doc, 'h4');
 208                 XML::deleteNode($doc, 'h5');
 209                 XML::deleteNode($doc, 'h6');
 210                 XML::deleteNode($doc, 'ol');
 211                 XML::deleteNode($doc, 'ul');
 212
 213                 $xpath = new DOMXPath($doc);
 214
 215                 $list = $xpath->query('//meta[@content]');
 216                 foreach ($list as $node) {
 217                         $meta_tag = [];
 218                         if ($node->attributes->length) {
 219                                 foreach ($node->attributes as $attribute) {
 220                                         $meta_tag[$attribute->name] = $attribute->value;
 221                                 }
 222                         }
 223
 224                         if (@$meta_tag['http-equiv'] == 'refresh') {
 225                                 $path = $meta_tag['content'];
 226                                 $pathinfo = explode(';', $path);
 227                                 $content = '';
 228                                 foreach ($pathinfo as $value) {
 229                                         if (substr(strtolower($value), 0, 4) == 'url=') {
 230                                                 $content = substr($value, 4);
 231                                         }
 232                                 }
 233                                 if ($content != '') {
 234                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 235                                         return $siteinfo;
 236                                 }
 237                         }
 238                 }
 239
 240                 $list = $xpath->query('//title');
 241                 if ($list->length > 0) {
 242                         $siteinfo['title'] = trim($list->item(0)->nodeValue);
 243                 }
 244
 245                 $list = $xpath->query('//meta[@name]');
 246                 foreach ($list as $node) {
 247                         $meta_tag = [];
 248                         if ($node->attributes->length) {
 249                                 foreach ($node->attributes as $attribute) {
 250                                         $meta_tag[$attribute->name] = $attribute->value;
 251                                 }
 252                         }
 253
 254                         if (empty($meta_tag['content'])) {
 255                                 continue;
 256                         }
 257
 258                         $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 259
 260                         switch (strtolower($meta_tag['name'])) {
 261                                 case 'fulltitle':
 262                                         $siteinfo['title'] = trim($meta_tag['content']);
 263                                         break;
 264                                 case 'description':
 265                                         $siteinfo['text'] = trim($meta_tag['content']);
 266                                         break;
 267                                 case 'thumbnail':
 268                                         $siteinfo['image'] = $meta_tag['content'];
 269                                         break;
 270                                 case 'twitter:image':
 271                                         $siteinfo['image'] = $meta_tag['content'];
 272                                         break;
 273                                 case 'twitter:image:src':
 274                                         $siteinfo['image'] = $meta_tag['content'];
 275                                         break;
 276                                 case 'twitter:card':
 277                                         // Detect photo pages
 278                                         if ($meta_tag['content'] == 'summary_large_image') {
 279                                                 $siteinfo['type'] = 'photo';
 280                                         }
 281                                         break;
 282                                 case 'twitter:description':
 283                                         $siteinfo['text'] = trim($meta_tag['content']);
 284                                         break;
 285                                 case 'twitter:title':
 286                                         $siteinfo['title'] = trim($meta_tag['content']);
 287                                         break;
 288                                 case 'dc.title':
 289                                         $siteinfo['title'] = trim($meta_tag['content']);
 290                                         break;
 291                                 case 'dc.description':
 292                                         $siteinfo['text'] = trim($meta_tag['content']);
 293                                         break;
 294                                 case 'keywords':
 295                                         $keywords = explode(',', $meta_tag['content']);
 296                                         break;
 297                                 case 'news_keywords':
 298                                         $keywords = explode(',', $meta_tag['content']);
 299                                         break;
 300                         }
 301                 }
 302
 303                 if (isset($keywords)) {
 304                         $siteinfo['keywords'] = [];
 305                         foreach ($keywords as $keyword) {
 306                                 if (!in_array(trim($keyword), $siteinfo['keywords'])) {
 307                                         $siteinfo['keywords'][] = trim($keyword);
 308                                 }
 309                         }
 310                 }
 311
 312                 $list = $xpath->query('//meta[@property]');
 313                 foreach ($list as $node) {
 314                         $meta_tag = [];
 315                         if ($node->attributes->length) {
 316                                 foreach ($node->attributes as $attribute) {
 317                                         $meta_tag[$attribute->name] = $attribute->value;
 318                                 }
 319                         }
 320
 321                         if (!empty($meta_tag['content'])) {
 322                                 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 323
 324                                 switch (strtolower($meta_tag['property'])) {
 325                                         case 'og:image':
 326                                                 $siteinfo['image'] = $meta_tag['content'];
 327                                                 break;
 328                                         case 'og:title':
 329                                                 $siteinfo['title'] = trim($meta_tag['content']);
 330                                                 break;
 331                                         case 'og:description':
 332                                                 $siteinfo['text'] = trim($meta_tag['content']);
 333                                                 break;
 334                                 }
 335                         }
 336                 }
 337
 338                 // Prevent to have a photo type without an image
 339                 if ((empty($siteinfo['image']) || !empty($siteinfo['text'])) && ($siteinfo['type'] == 'photo')) {
 340                         $siteinfo['type'] = 'link';
 341                 }
 342
 343                 if (empty($siteinfo['image']) && !$no_guessing) {
 344                         $list = $xpath->query('//img[@src]');
 345                         foreach ($list as $node) {
 346                                 $img_tag = [];
 347                                 if ($node->attributes->length) {
 348                                         foreach ($node->attributes as $attribute) {
 349                                                 $img_tag[$attribute->name] = $attribute->value;
 350                                         }
 351                                 }
 352
 353                                 $src = self::completeUrl($img_tag['src'], $url);
 354                                 $photodata = Image::getInfoFromURL($src);
 355
 356                                 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
 357                                         if ($photodata[0] > 300) {
 358                                                 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
 359                                                 $photodata[0] = 300;
 360                                         }
 361                                         if ($photodata[1] > 300) {
 362                                                 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
 363                                                 $photodata[1] = 300;
 364                                         }
 365                                         $siteinfo['images'][] = [
 366                                                 'src'    => $src,
 367                                                 'width'  => $photodata[0],
 368                                                 'height' => $photodata[1]
 369                                         ];
 370                                 }
 371                         }
 372                 } elseif (!empty($siteinfo['image'])) {
 373                         $src = self::completeUrl($siteinfo['image'], $url);
 374
 375                         unset($siteinfo['image']);
 376
 377                         $photodata = Image::getInfoFromURL($src);
 378
 379                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 380                                 $siteinfo['images'][] = ['src' => $src,
 381                                         'width' => $photodata[0],
 382                                         'height' => $photodata[1]];
 383                         }
 384                 }
 385
 386                 if ((@$siteinfo['text'] == '') && (@$siteinfo['title'] != '') && !$no_guessing) {
 387                         $text = '';
 388
 389                         $list = $xpath->query('//div[@class="article"]');
 390                         foreach ($list as $node) {
 391                                 if (strlen($node->nodeValue) > 40) {
 392                                         $text .= ' ' . trim($node->nodeValue);
 393                                 }
 394                         }
 395
 396                         if ($text == '') {
 397                                 $list = $xpath->query('//div[@class="content"]');
 398                                 foreach ($list as $node) {
 399                                         if (strlen($node->nodeValue) > 40) {
 400                                                 $text .= ' ' . trim($node->nodeValue);
 401                                         }
 402                                 }
 403                         }
 404
 405                         // If none text was found then take the paragraph content
 406                         if ($text == '') {
 407                                 $list = $xpath->query('//p');
 408                                 foreach ($list as $node) {
 409                                         if (strlen($node->nodeValue) > 40) {
 410                                                 $text .= ' ' . trim($node->nodeValue);
 411                                         }
 412                                 }
 413                         }
 414
 415                         if ($text != '') {
 416                                 $text = trim(str_replace(["\n", "\r"], [' ', ' '], $text));
 417
 418                                 while (strpos($text, '  ')) {
 419                                         $text = trim(str_replace('  ', ' ', $text));
 420                                 }
 421
 422                                 $siteinfo['text'] = trim(html_entity_decode(substr($text, 0, 350), ENT_QUOTES, 'UTF-8') . '...');
 423                         }
 424                 }
 425
 426                 Logger::log('Siteinfo for ' . $url . ' ' . print_r($siteinfo, true), Logger::DEBUG);
 427
 428                 Addon::callHooks('getsiteinfo', $siteinfo);
 429
 430                 return $siteinfo;
 431         }
 432
 433         /**
 434          * @brief Convert tags from CSV to an array
 435          *
 436          * @param string $string Tags
 437          * @return array with formatted Hashtags
 438          */
 439         public static function convertTagsToArray($string)
 440         {
 441                 $arr_tags = str_getcsv($string);
 442                 if (count($arr_tags)) {
 443                         // add the # sign to every tag
 444                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 445
 446                         return $arr_tags;
 447                 }
 448         }
 449
 450         /**
 451          * @brief Add a hasht sign to a string
 452          *
 453          *  This method is used as callback function
 454          *
 455          * @param string $tag The pure tag name
 456          * @param int    $k   Counter for internal use
 457          * @return void
 458          */
 459         private static function arrAddHashes(&$tag, $k)
 460         {
 461                 $tag = "#" . $tag;
 462         }
 463
 464         /**
 465          * @brief Add a scheme to an url
 466          *
 467          * The src attribute of some html elements (e.g. images)
 468          * can miss the scheme so we need to add the correct
 469          * scheme
 470          *
 471          * @param string $url    The url which possibly does have
 472          *                       a missing scheme (a link to an image)
 473          * @param string $scheme The url with a correct scheme
 474          *                       (e.g. the url from the webpage which does contain the image)
 475          *
 476          * @return string The url with a scheme
 477          */
 478         private static function completeUrl($url, $scheme)
 479         {
 480                 $urlarr = parse_url($url);
 481
 482                 // If the url does allready have an scheme
 483                 // we can stop the process here
 484                 if (isset($urlarr["scheme"])) {
 485                         return($url);
 486                 }
 487
 488                 $schemearr = parse_url($scheme);
 489
 490                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 491
 492                 if (!empty($schemearr["port"])) {
 493                         $complete .= ":".$schemearr["port"];
 494                 }
 495
 496                 if (!empty($urlarr["path"])) {
 497                         if (strpos($urlarr["path"], "/") !== 0) {
 498                                 $complete .= "/";
 499                         }
 500
 501                         $complete .= $urlarr["path"];
 502                 }
 503
 504                 if (!empty($urlarr["query"])) {
 505                         $complete .= "?".$urlarr["query"];
 506                 }
 507
 508                 if (!empty($urlarr["fragment"])) {
 509                         $complete .= "#".$urlarr["fragment"];
 510                 }
 511
 512                 return($complete);
 513         }
 514 }