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'] = $oembed_data->type;
 162                                 }
 163
 164                                 // See https://github.com/friendica/friendica/pull/5763#discussion_r217913178
 165                                 if ($siteinfo['type'] != 'photo') {
 166                                         if (isset($oembed_data->title)) {
 167                                                 $siteinfo['title'] = trim($oembed_data->title);
 168                                         }
 169                                         if (isset($oembed_data->description)) {
 170                                                 $siteinfo['text'] = trim($oembed_data->description);
 171                                         }
 172                                         if (isset($oembed_data->thumbnail_url)) {
 173                                                 $siteinfo['image'] = $oembed_data->thumbnail_url;
 174                                         }
 175                                 }
 176                         }
 177                 }
 178
 179                 // Fetch the first mentioned charset. Can be in body or header
 180                 $charset = '';
 181                 if (preg_match('/charset=(.*?)[\'"\s\n]/', $header, $matches)) {
 182                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 183                 }
 184
 185                 if ($charset == '') {
 186                         $charset = 'utf-8';
 187                 }
 188
 189                 if (($charset != '') && (strtoupper($charset) != 'UTF-8')) {
 190                         logger('detected charset ' . $charset, LOGGER_DEBUG);
 191                         $body = iconv($charset, 'UTF-8//TRANSLIT', $body);
 192                 }
 193
 194                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', 'UTF-8');
 195
 196                 $doc = new DOMDocument();
 197                 @$doc->loadHTML($body);
 198
 199                 XML::deleteNode($doc, 'style');
 200                 XML::deleteNode($doc, 'script');
 201                 XML::deleteNode($doc, 'option');
 202                 XML::deleteNode($doc, 'h1');
 203                 XML::deleteNode($doc, 'h2');
 204                 XML::deleteNode($doc, 'h3');
 205                 XML::deleteNode($doc, 'h4');
 206                 XML::deleteNode($doc, 'h5');
 207                 XML::deleteNode($doc, 'h6');
 208                 XML::deleteNode($doc, 'ol');
 209                 XML::deleteNode($doc, 'ul');
 210
 211                 $xpath = new DOMXPath($doc);
 212
 213                 $list = $xpath->query('//meta[@content]');
 214                 foreach ($list as $node) {
 215                         $meta_tag = [];
 216                         if ($node->attributes->length) {
 217                                 foreach ($node->attributes as $attribute) {
 218                                         $meta_tag[$attribute->name] = $attribute->value;
 219                                 }
 220                         }
 221
 222                         if (@$meta_tag['http-equiv'] == 'refresh') {
 223                                 $path = $meta_tag['content'];
 224                                 $pathinfo = explode(';', $path);
 225                                 $content = '';
 226                                 foreach ($pathinfo as $value) {
 227                                         if (substr(strtolower($value), 0, 4) == 'url=') {
 228                                                 $content = substr($value, 4);
 229                                         }
 230                                 }
 231                                 if ($content != '') {
 232                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 233                                         return $siteinfo;
 234                                 }
 235                         }
 236                 }
 237
 238                 $list = $xpath->query('//title');
 239                 if ($list->length > 0) {
 240                         $siteinfo['title'] = trim($list->item(0)->nodeValue);
 241                 }
 242
 243                 $list = $xpath->query('//meta[@name]');
 244                 foreach ($list as $node) {
 245                         $meta_tag = [];
 246                         if ($node->attributes->length) {
 247                                 foreach ($node->attributes as $attribute) {
 248                                         $meta_tag[$attribute->name] = $attribute->value;
 249                                 }
 250                         }
 251
 252                         if (empty($meta_tag['content'])) {
 253                                 continue;
 254                         }
 255
 256                         $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 257
 258                         switch (strtolower($meta_tag['name'])) {
 259                                 case 'fulltitle':
 260                                         $siteinfo['title'] = trim($meta_tag['content']);
 261                                         break;
 262                                 case 'description':
 263                                         $siteinfo['text'] = trim($meta_tag['content']);
 264                                         break;
 265                                 case 'thumbnail':
 266                                         $siteinfo['image'] = $meta_tag['content'];
 267                                         break;
 268                                 case 'twitter:image':
 269                                         $siteinfo['image'] = $meta_tag['content'];
 270                                         break;
 271                                 case 'twitter:image:src':
 272                                         $siteinfo['image'] = $meta_tag['content'];
 273                                         break;
 274                                 case 'twitter:card':
 275                                         // Detect photo pages
 276                                         if ($meta_tag['content'] == 'summary_large_image') {
 277                                                 $siteinfo['type'] = 'photo';
 278                                         }
 279                                         break;
 280                                 case 'twitter:description':
 281                                         $siteinfo['text'] = trim($meta_tag['content']);
 282                                         break;
 283                                 case 'twitter:title':
 284                                         $siteinfo['title'] = trim($meta_tag['content']);
 285                                         break;
 286                                 case 'dc.title':
 287                                         $siteinfo['title'] = trim($meta_tag['content']);
 288                                         break;
 289                                 case 'dc.description':
 290                                         $siteinfo['text'] = trim($meta_tag['content']);
 291                                         break;
 292                                 case 'keywords':
 293                                         $keywords = explode(',', $meta_tag['content']);
 294                                         break;
 295                                 case 'news_keywords':
 296                                         $keywords = explode(',', $meta_tag['content']);
 297                                         break;
 298                         }
 299                 }
 300
 301                 if (isset($keywords)) {
 302                         $siteinfo['keywords'] = [];
 303                         foreach ($keywords as $keyword) {
 304                                 if (!in_array(trim($keyword), $siteinfo['keywords'])) {
 305                                         $siteinfo['keywords'][] = trim($keyword);
 306                                 }
 307                         }
 308                 }
 309
 310                 $list = $xpath->query('//meta[@property]');
 311                 foreach ($list as $node) {
 312                         $meta_tag = [];
 313                         if ($node->attributes->length) {
 314                                 foreach ($node->attributes as $attribute) {
 315                                         $meta_tag[$attribute->name] = $attribute->value;
 316                                 }
 317                         }
 318
 319                         if (!empty($meta_tag['content'])) {
 320                                 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
 321
 322                                 switch (strtolower($meta_tag['property'])) {
 323                                         case 'og:image':
 324                                                 $siteinfo['image'] = $meta_tag['content'];
 325                                                 break;
 326                                         case 'og:title':
 327                                                 $siteinfo['title'] = trim($meta_tag['content']);
 328                                                 break;
 329                                         case 'og:description':
 330                                                 $siteinfo['text'] = trim($meta_tag['content']);
 331                                                 break;
 332                                 }
 333                         }
 334                 }
 335
 336                 // Prevent to have a photo type without an image
 337                 if ((empty($siteinfo['image']) || !empty($siteinfo['text'])) && ($siteinfo['type'] == 'photo')) {
 338                         $siteinfo['type'] = 'link';
 339                 }
 340
 341                 if (empty($siteinfo['image']) && !$no_guessing) {
 342                         $list = $xpath->query('//img[@src]');
 343                         foreach ($list as $node) {
 344                                 $img_tag = [];
 345                                 if ($node->attributes->length) {
 346                                         foreach ($node->attributes as $attribute) {
 347                                                 $img_tag[$attribute->name] = $attribute->value;
 348                                         }
 349                                 }
 350
 351                                 $src = self::completeUrl($img_tag['src'], $url);
 352                                 $photodata = Image::getInfoFromURL($src);
 353
 354                                 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
 355                                         if ($photodata[0] > 300) {
 356                                                 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
 357                                                 $photodata[0] = 300;
 358                                         }
 359                                         if ($photodata[1] > 300) {
 360                                                 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
 361                                                 $photodata[1] = 300;
 362                                         }
 363                                         $siteinfo['images'][] = [
 364                                                 'src'    => $src,
 365                                                 'width'  => $photodata[0],
 366                                                 'height' => $photodata[1]
 367                                         ];
 368                                 }
 369                         }
 370                 } elseif (!empty($siteinfo['image'])) {
 371                         $src = self::completeUrl($siteinfo['image'], $url);
 372
 373                         unset($siteinfo['image']);
 374
 375                         $photodata = Image::getInfoFromURL($src);
 376
 377                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 378                                 $siteinfo['images'][] = ['src' => $src,
 379                                         'width' => $photodata[0],
 380                                         'height' => $photodata[1]];
 381                         }
 382                 }
 383
 384                 if ((@$siteinfo['text'] == '') && (@$siteinfo['title'] != '') && !$no_guessing) {
 385                         $text = '';
 386
 387                         $list = $xpath->query('//div[@class="article"]');
 388                         foreach ($list as $node) {
 389                                 if (strlen($node->nodeValue) > 40) {
 390                                         $text .= ' ' . trim($node->nodeValue);
 391                                 }
 392                         }
 393
 394                         if ($text == '') {
 395                                 $list = $xpath->query('//div[@class="content"]');
 396                                 foreach ($list as $node) {
 397                                         if (strlen($node->nodeValue) > 40) {
 398                                                 $text .= ' ' . trim($node->nodeValue);
 399                                         }
 400                                 }
 401                         }
 402
 403                         // If none text was found then take the paragraph content
 404                         if ($text == '') {
 405                                 $list = $xpath->query('//p');
 406                                 foreach ($list as $node) {
 407                                         if (strlen($node->nodeValue) > 40) {
 408                                                 $text .= ' ' . trim($node->nodeValue);
 409                                         }
 410                                 }
 411                         }
 412
 413                         if ($text != '') {
 414                                 $text = trim(str_replace(["\n", "\r"], [' ', ' '], $text));
 415
 416                                 while (strpos($text, '  ')) {
 417                                         $text = trim(str_replace('  ', ' ', $text));
 418                                 }
 419
 420                                 $siteinfo['text'] = trim(html_entity_decode(substr($text, 0, 350), ENT_QUOTES, 'UTF-8') . '...');
 421                         }
 422                 }
 423
 424                 logger('Siteinfo for ' . $url . ' ' . print_r($siteinfo, true), LOGGER_DEBUG);
 425
 426                 Addon::callHooks('getsiteinfo', $siteinfo);
 427
 428                 return $siteinfo;
 429         }
 430
 431         /**
 432          * @brief Convert tags from CSV to an array
 433          *
 434          * @param string $string Tags
 435          * @return array with formatted Hashtags
 436          */
 437         public static function convertTagsToArray($string)
 438         {
 439                 $arr_tags = str_getcsv($string);
 440                 if (count($arr_tags)) {
 441                         // add the # sign to every tag
 442                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 443
 444                         return $arr_tags;
 445                 }
 446         }
 447
 448         /**
 449          * @brief Add a hasht sign to a string
 450          *
 451          *  This method is used as callback function
 452          *
 453          * @param string $tag The pure tag name
 454          * @param int    $k   Counter for internal use
 455          * @return void
 456          */
 457         private static function arrAddHashes(&$tag, $k)
 458         {
 459                 $tag = "#" . $tag;
 460         }
 461
 462         /**
 463          * @brief Add a scheme to an url
 464          *
 465          * The src attribute of some html elements (e.g. images)
 466          * can miss the scheme so we need to add the correct
 467          * scheme
 468          *
 469          * @param string $url    The url which possibly does have
 470          *                       a missing scheme (a link to an image)
 471          * @param string $scheme The url with a correct scheme
 472          *                       (e.g. the url from the webpage which does contain the image)
 473          *
 474          * @return string The url with a scheme
 475          */
 476         private static function completeUrl($url, $scheme)
 477         {
 478                 $urlarr = parse_url($url);
 479
 480                 // If the url does allready have an scheme
 481                 // we can stop the process here
 482                 if (isset($urlarr["scheme"])) {
 483                         return($url);
 484                 }
 485
 486                 $schemearr = parse_url($scheme);
 487
 488                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 489
 490                 if (!empty($schemearr["port"])) {
 491                         $complete .= ":".$schemearr["port"];
 492                 }
 493
 494                 if (!empty($urlarr["path"])) {
 495                         if (strpos($urlarr["path"], "/") !== 0) {
 496                                 $complete .= "/";
 497                         }
 498
 499                         $complete .= $urlarr["path"];
 500                 }
 501
 502                 if (!empty($urlarr["query"])) {
 503                         $complete .= "?".$urlarr["query"];
 504                 }
 505
 506                 if (!empty($urlarr["fragment"])) {
 507                         $complete .= "#".$urlarr["fragment"];
 508                 }
 509
 510                 return($complete);
 511         }
 512 }