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