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