src/ParseUrl.php

   1 <?php
   2 /**
   3  * @file include/ParseUrl.php
   4  * @brief Get informations about a given URL
   5  */
   6 namespace Friendica;
   7
   8 use Friendica\Core\Config;
   9 use Friendica\Object\Photo;
  10 use Friendica\Util\XML;
  11
  12 use dba;
  13 use DomXPath;
  14 use DOMDocument;
  15
  16 require_once "include/network.php";
  17 require_once "include/oembed.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 type $url         The url of the page which should be scraped
  28          * @param type $no_guessing If true the parse doens't search for
  29          *                          preview pictures
  30          * @param type $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                 $r = q(
  53                         "SELECT * FROM `parsed_url` WHERE `url` = '%s' AND `guessing` = %d AND `oembed` = %d",
  54                         dbesc(normalise_link($url)),
  55                         intval(!$no_guessing),
  56                         intval($do_oembed)
  57                 );
  58
  59                 if ($r) {
  60                         $data = $r[0]["content"];
  61                 }
  62
  63                 if (!is_null($data)) {
  64                         $data = unserialize($data);
  65                         return $data;
  66                 }
  67
  68                 $data = self::getSiteinfo($url, $no_guessing, $do_oembed);
  69
  70                 dba::insert(
  71                         'parsed_url',
  72                         array(
  73                                 'url' => normalise_link($url), 'guessing' => !$no_guessing,
  74                                 'oembed' => $do_oembed, 'content' => serialize($data),
  75                                 'created' => datetime_convert()),
  76                         true
  77                 );
  78
  79                 return $data;
  80         }
  81         /**
  82          * @brief Parse a page for embeddable content information
  83          *
  84          * This method parses to url for meta data which can be used to embed
  85          * the content. If available it prioritizes Open Graph meta tags.
  86          * If this is not available it uses the twitter cards meta tags.
  87          * As fallback it uses standard html elements with meta informations
  88          * like \<title\>Awesome Title\</title\> or
  89          * \<meta name="description" content="An awesome description"\>
  90          *
  91          * @param type $url         The url of the page which should be scraped
  92          * @param type $no_guessing If true the parse doens't search for
  93          *                          preview pictures
  94          * @param type $do_oembed   The false option is used by the function fetch_oembed()
  95          *                          to avoid endless loops
  96          * @param type $count       Internal counter to avoid endless loops
  97          *
  98          * @return array which contains needed data for embedding
  99          *    string 'url' => The url of the parsed page
 100          *    string 'type' => Content type
 101          *    string 'title' => The title of the content
 102          *    string 'text' => The description for the content
 103          *    string 'image' => A preview image of the content (only available
 104          *                if $no_geuessing = false
 105          *    array'images' = Array of preview pictures
 106          *    string 'keywords' => The tags which belong to the content
 107          *
 108          * @todo https://developers.google.com/+/plugins/snippet/
 109          * @verbatim
 110          * <meta itemprop="name" content="Awesome title">
 111          * <meta itemprop="description" content="An awesome description">
 112          * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
 113          *
 114          * <body itemscope itemtype="http://schema.org/Product">
 115          *   <h1 itemprop="name">Shiny Trinket</h1>
 116          *   <img itemprop="image" src="{image-url}" />
 117          *   <p itemprop="description">Shiny trinkets are shiny.</p>
 118          * </body>
 119          * @endverbatim
 120          */
 121         public static function getSiteinfo($url, $no_guessing = false, $do_oembed = true, $count = 1)
 122         {
 123                 $a = get_app();
 124
 125                 $siteinfo = array();
 126
 127                 // Check if the URL does contain a scheme
 128                 $scheme = parse_url($url, PHP_URL_SCHEME);
 129
 130                 if ($scheme == "") {
 131                         $url = "http://".trim($url, "/");
 132                 }
 133
 134                 if ($count > 10) {
 135                         logger("parseurl_getsiteinfo: Endless loop detected for ".$url, LOGGER_DEBUG);
 136                         return($siteinfo);
 137                 }
 138
 139                 $url = trim($url, "'");
 140                 $url = trim($url, '"');
 141
 142                 $url = strip_tracking_query_params($url);
 143
 144                 $siteinfo["url"] = $url;
 145                 $siteinfo["type"] = "link";
 146
 147                 $data = z_fetch_url($url);
 148                 if (!$data['success']) {
 149                         return($siteinfo);
 150                 }
 151
 152                 // If the file is too large then exit
 153                 if ($data["info"]["download_content_length"] > 1000000) {
 154                         return($siteinfo);
 155                 }
 156
 157                 // If it isn't a HTML file then exit
 158                 if (($data["info"]["content_type"] != "") && !strstr(strtolower($data["info"]["content_type"]), "html")) {
 159                         return($siteinfo);
 160                 }
 161
 162                 $header = $data["header"];
 163                 $body = $data["body"];
 164
 165                 if ($do_oembed) {
 166                         $oembed_data = oembed_fetch_url($url);
 167
 168                         if (!in_array($oembed_data->type, array("error", "rich", ""))) {
 169                                 $siteinfo["type"] = $oembed_data->type;
 170                         }
 171
 172                         if (($oembed_data->type == "link") && ($siteinfo["type"] != "photo")) {
 173                                 if (isset($oembed_data->title)) {
 174                                         $siteinfo["title"] = trim($oembed_data->title);
 175                                 }
 176                                 if (isset($oembed_data->description)) {
 177                                         $siteinfo["text"] = trim($oembed_data->description);
 178                                 }
 179                                 if (isset($oembed_data->thumbnail_url)) {
 180                                         $siteinfo["image"] = $oembed_data->thumbnail_url;
 181                                 }
 182                         }
 183                 }
 184
 185                 // Fetch the first mentioned charset. Can be in body or header
 186                 $charset = "";
 187                 if (preg_match('/charset=(.*?)['."'".'"\s\n]/', $header, $matches)) {
 188                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 189                 }
 190
 191                 if ($charset == "") {
 192                         $charset = "utf-8";
 193                 }
 194
 195                 if (($charset != "") && (strtoupper($charset) != "UTF-8")) {
 196                         logger("parseurl_getsiteinfo: detected charset ".$charset, LOGGER_DEBUG);
 197                         //$body = mb_convert_encoding($body, "UTF-8", $charset);
 198                         $body = iconv($charset, "UTF-8//TRANSLIT", $body);
 199                 }
 200
 201                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', "UTF-8");
 202
 203                 $doc = new DOMDocument();
 204                 @$doc->loadHTML($body);
 205
 206                 XML::deleteNode($doc, "style");
 207                 XML::deleteNode($doc, "script");
 208                 XML::deleteNode($doc, "option");
 209                 XML::deleteNode($doc, "h1");
 210                 XML::deleteNode($doc, "h2");
 211                 XML::deleteNode($doc, "h3");
 212                 XML::deleteNode($doc, "h4");
 213                 XML::deleteNode($doc, "h5");
 214                 XML::deleteNode($doc, "h6");
 215                 XML::deleteNode($doc, "ol");
 216                 XML::deleteNode($doc, "ul");
 217
 218                 $xpath = new DomXPath($doc);
 219
 220                 $list = $xpath->query("//meta[@content]");
 221                 foreach ($list as $node) {
 222                         $attr = array();
 223                         if ($node->attributes->length) {
 224                                 foreach ($node->attributes as $attribute) {
 225                                         $attr[$attribute->name] = $attribute->value;
 226                                 }
 227                         }
 228
 229                         if (@$attr["http-equiv"] == "refresh") {
 230                                 $path = $attr["content"];
 231                                 $pathinfo = explode(";", $path);
 232                                 $content = "";
 233                                 foreach ($pathinfo as $value) {
 234                                         if (substr(strtolower($value), 0, 4) == "url=") {
 235                                                 $content = substr($value, 4);
 236                                         }
 237                                 }
 238                                 if ($content != "") {
 239                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 240                                         return($siteinfo);
 241                                 }
 242                         }
 243                 }
 244
 245                 $list = $xpath->query("//title");
 246                 if ($list->length > 0) {
 247                         $siteinfo["title"] = trim($list->item(0)->nodeValue);
 248                 }
 249
 250                 //$list = $xpath->query("head/meta[@name]");
 251                 $list = $xpath->query("//meta[@name]");
 252                 foreach ($list as $node) {
 253                         $attr = array();
 254                         if ($node->attributes->length) {
 255                                 foreach ($node->attributes as $attribute) {
 256                                         $attr[$attribute->name] = $attribute->value;
 257                                 }
 258                         }
 259
 260                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 261
 262                         if ($attr["content"] != "") {
 263                                 switch (strtolower($attr["name"])) {
 264                                         case "fulltitle":
 265                                                 $siteinfo["title"] = trim($attr["content"]);
 266                                                 break;
 267                                         case "description":
 268                                                 $siteinfo["text"] = trim($attr["content"]);
 269                                                 break;
 270                                         case "thumbnail":
 271                                                 $siteinfo["image"] = $attr["content"];
 272                                                 break;
 273                                         case "twitter:image":
 274                                                 $siteinfo["image"] = $attr["content"];
 275                                                 break;
 276                                         case "twitter:image:src":
 277                                                 $siteinfo["image"] = $attr["content"];
 278                                                 break;
 279                                         case "twitter:card":
 280                                                 if (($siteinfo["type"] == "") || ($attr["content"] == "photo")) {
 281                                                         $siteinfo["type"] = $attr["content"];
 282                                                 }
 283                                                 break;
 284                                         case "twitter:description":
 285                                                 $siteinfo["text"] = trim($attr["content"]);
 286                                                 break;
 287                                         case "twitter:title":
 288                                                 $siteinfo["title"] = trim($attr["content"]);
 289                                                 break;
 290                                         case "dc.title":
 291                                                 $siteinfo["title"] = trim($attr["content"]);
 292                                                 break;
 293                                         case "dc.description":
 294                                                 $siteinfo["text"] = trim($attr["content"]);
 295                                                 break;
 296                                         case "keywords":
 297                                                 $keywords = explode(",", $attr["content"]);
 298                                                 break;
 299                                         case "news_keywords":
 300                                                 $keywords = explode(",", $attr["content"]);
 301                                                 break;
 302                                 }
 303                         }
 304                         if ($siteinfo["type"] == "summary") {
 305                                 $siteinfo["type"] = "link";
 306                         }
 307                 }
 308
 309                 if (isset($keywords)) {
 310                         $siteinfo["keywords"] = array();
 311                         foreach ($keywords as $keyword) {
 312                                 if (!in_array(trim($keyword), $siteinfo["keywords"])) {
 313                                         $siteinfo["keywords"][] = trim($keyword);
 314                                 }
 315                         }
 316                 }
 317
 318                 //$list = $xpath->query("head/meta[@property]");
 319                 $list = $xpath->query("//meta[@property]");
 320                 foreach ($list as $node) {
 321                         $attr = array();
 322                         if ($node->attributes->length) {
 323                                 foreach ($node->attributes as $attribute) {
 324                                         $attr[$attribute->name] = $attribute->value;
 325                                 }
 326                         }
 327
 328                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 329
 330                         if ($attr["content"] != "") {
 331                                 switch (strtolower($attr["property"])) {
 332                                         case "og:image":
 333                                                 $siteinfo["image"] = $attr["content"];
 334                                                 break;
 335                                         case "og:title":
 336                                                 $siteinfo["title"] = trim($attr["content"]);
 337                                                 break;
 338                                         case "og:description":
 339                                                 $siteinfo["text"] = trim($attr["content"]);
 340                                                 break;
 341                                 }
 342                         }
 343                 }
 344
 345                 if ((@$siteinfo["image"] == "") && !$no_guessing) {
 346                         $list = $xpath->query("//img[@src]");
 347                         foreach ($list as $node) {
 348                                 $attr = array();
 349                                 if ($node->attributes->length) {
 350                                         foreach ($node->attributes as $attribute) {
 351                                                 $attr[$attribute->name] = $attribute->value;
 352                                         }
 353                                 }
 354
 355                                 $src = self::completeUrl($attr["src"], $url);
 356                                 $photodata = Photo::getInfoFromURL($src);
 357
 358                                 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
 359                                         if ($photodata[0] > 300) {
 360                                                 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
 361                                                 $photodata[0] = 300;
 362                                         }
 363                                         if ($photodata[1] > 300) {
 364                                                 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
 365                                                 $photodata[1] = 300;
 366                                         }
 367                                         $siteinfo["images"][] = array("src" => $src,
 368                                                                         "width" => $photodata[0],
 369                                                                         "height" => $photodata[1]);
 370                                 }
 371                         }
 372                 } elseif ($siteinfo["image"] != "") {
 373                         $src = self::completeUrl($siteinfo["image"], $url);
 374
 375                         unset($siteinfo["image"]);
 376
 377                         $photodata = Photo::getInfoFromURL($src);
 378
 379                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 380                                 $siteinfo["images"][] = array("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(array("\n", "\r"), array(" ", " "), $text));
 417
 418                                 while (strpos($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("parseurl_getsiteinfo: Siteinfo for ".$url." ".print_r($siteinfo, true), LOGGER_DEBUG);
 427
 428                 call_hooks("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, array("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 (@$schemearr["port"] != "") {
 493                         $complete .= ":".$schemearr["port"];
 494                 }
 495
 496                 if (strpos($urlarr["path"], "/") !== 0) {
 497                         $complete .= "/";
 498                 }
 499
 500                 $complete .= $urlarr["path"];
 501
 502                 if (@$urlarr["query"] != "") {
 503                         $complete .= "?".$urlarr["query"];
 504                 }
 505
 506                 if (@$urlarr["fragment"] != "") {
 507                         $complete .= "#".$urlarr["fragment"];
 508                 }
 509
 510                 return($complete);
 511         }
 512 }