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\Image;
  10 use Friendica\Util\XML;
  11
  12 use dba;
  13 use DOMXPath;
  14 use DOMDocument;
  15
  16 require_once 'include/dba.php';
  17 require_once "include/network.php";
  18 require_once "include/oembed.php";
  19
  20 /**
  21  * @brief Class with methods for extracting certain content from an url
  22  */
  23 class ParseUrl
  24 {
  25         /**
  26          * @brief Search for chached embeddable data of an url otherwise fetch it
  27          *
  28          * @param string $url         The url of the page which should be scraped
  29          * @param bool $no_guessing If true the parse doens't search for
  30          *                          preview pictures
  31          * @param bool $do_oembed   The false option is used by the function fetch_oembed()
  32          *                          to avoid endless loops
  33          *
  34          * @return array which contains needed data for embedding
  35          *    string 'url' => The url of the parsed page
  36          *    string 'type' => Content type
  37          *    string 'title' => The title of the content
  38          *    string 'text' => The description for the content
  39          *    string 'image' => A preview image of the content (only available
  40          *                if $no_geuessing = false
  41          *    array'images' = Array of preview pictures
  42          *    string 'keywords' => The tags which belong to the content
  43          *
  44          * @see ParseUrl::getSiteinfo() for more information about scraping
  45          * embeddable content
  46          */
  47         public static function getSiteinfoCached($url, $no_guessing = false, $do_oembed = true)
  48         {
  49                 if ($url == "") {
  50                         return false;
  51                 }
  52
  53                 $r = q(
  54                         "SELECT * FROM `parsed_url` WHERE `url` = '%s' AND `guessing` = %d AND `oembed` = %d",
  55                         dbesc(normalise_link($url)),
  56                         intval(!$no_guessing),
  57                         intval($do_oembed)
  58                 );
  59
  60                 if ($r) {
  61                         $data = $r[0]["content"];
  62                 }
  63
  64                 if (!is_null($data)) {
  65                         $data = unserialize($data);
  66                         return $data;
  67                 }
  68
  69                 $data = self::getSiteinfo($url, $no_guessing, $do_oembed);
  70
  71                 dba::insert(
  72                         'parsed_url',
  73                         array(
  74                                 'url' => normalise_link($url), 'guessing' => !$no_guessing,
  75                                 'oembed' => $do_oembed, 'content' => serialize($data),
  76                                 'created' => datetime_convert()),
  77                         true
  78                 );
  79
  80                 return $data;
  81         }
  82         /**
  83          * @brief Parse a page for embeddable content information
  84          *
  85          * This method parses to url for meta data which can be used to embed
  86          * the content. If available it prioritizes Open Graph meta tags.
  87          * If this is not available it uses the twitter cards meta tags.
  88          * As fallback it uses standard html elements with meta informations
  89          * like \<title\>Awesome Title\</title\> or
  90          * \<meta name="description" content="An awesome description"\>
  91          *
  92          * @param string $url         The url of the page which should be scraped
  93          * @param bool $no_guessing If true the parse doens't search for
  94          *                          preview pictures
  95          * @param bool $do_oembed   The false option is used by the function fetch_oembed()
  96          *                          to avoid endless loops
  97          * @param int $count       Internal counter to avoid endless loops
  98          *
  99          * @return array which contains needed data for embedding
 100          *    string 'url' => The url of the parsed page
 101          *    string 'type' => Content type
 102          *    string 'title' => The title of the content
 103          *    string 'text' => The description for the content
 104          *    string 'image' => A preview image of the content (only available
 105          *                if $no_geuessing = false
 106          *    array'images' = Array of preview pictures
 107          *    string 'keywords' => The tags which belong to the content
 108          *
 109          * @todo https://developers.google.com/+/plugins/snippet/
 110          * @verbatim
 111          * <meta itemprop="name" content="Awesome title">
 112          * <meta itemprop="description" content="An awesome description">
 113          * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
 114          *
 115          * <body itemscope itemtype="http://schema.org/Product">
 116          *   <h1 itemprop="name">Shiny Trinket</h1>
 117          *   <img itemprop="image" src="{image-url}" />
 118          *   <p itemprop="description">Shiny trinkets are shiny.</p>
 119          * </body>
 120          * @endverbatim
 121          */
 122         public static function getSiteinfo($url, $no_guessing = false, $do_oembed = true, $count = 1)
 123         {
 124                 $a = get_app();
 125
 126                 $siteinfo = array();
 127
 128                 // Check if the URL does contain a scheme
 129                 $scheme = parse_url($url, PHP_URL_SCHEME);
 130
 131                 if ($scheme == "") {
 132                         $url = "http://".trim($url, "/");
 133                 }
 134
 135                 if ($count > 10) {
 136                         logger("parseurl_getsiteinfo: Endless loop detected for ".$url, LOGGER_DEBUG);
 137                         return($siteinfo);
 138                 }
 139
 140                 $url = trim($url, "'");
 141                 $url = trim($url, '"');
 142
 143                 $url = strip_tracking_query_params($url);
 144
 145                 $siteinfo["url"] = $url;
 146                 $siteinfo["type"] = "link";
 147
 148                 $data = z_fetch_url($url);
 149                 if (!$data['success']) {
 150                         return($siteinfo);
 151                 }
 152
 153                 // If the file is too large then exit
 154                 if ($data["info"]["download_content_length"] > 1000000) {
 155                         return($siteinfo);
 156                 }
 157
 158                 // If it isn't a HTML file then exit
 159                 if (($data["info"]["content_type"] != "") && !strstr(strtolower($data["info"]["content_type"]), "html")) {
 160                         return($siteinfo);
 161                 }
 162
 163                 $header = $data["header"];
 164                 $body = $data["body"];
 165
 166                 if ($do_oembed) {
 167                         $oembed_data = oembed_fetch_url($url);
 168
 169                         if (!in_array($oembed_data->type, array("error", "rich", ""))) {
 170                                 $siteinfo["type"] = $oembed_data->type;
 171                         }
 172
 173                         if (($oembed_data->type == "link") && ($siteinfo["type"] != "photo")) {
 174                                 if (isset($oembed_data->title)) {
 175                                         $siteinfo["title"] = trim($oembed_data->title);
 176                                 }
 177                                 if (isset($oembed_data->description)) {
 178                                         $siteinfo["text"] = trim($oembed_data->description);
 179                                 }
 180                                 if (isset($oembed_data->thumbnail_url)) {
 181                                         $siteinfo["image"] = $oembed_data->thumbnail_url;
 182                                 }
 183                         }
 184                 }
 185
 186                 // Fetch the first mentioned charset. Can be in body or header
 187                 $charset = "";
 188                 if (preg_match('/charset=(.*?)['."'".'"\s\n]/', $header, $matches)) {
 189                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 190                 }
 191
 192                 if ($charset == "") {
 193                         $charset = "utf-8";
 194                 }
 195
 196                 if (($charset != "") && (strtoupper($charset) != "UTF-8")) {
 197                         logger("parseurl_getsiteinfo: detected charset ".$charset, LOGGER_DEBUG);
 198                         //$body = mb_convert_encoding($body, "UTF-8", $charset);
 199                         $body = iconv($charset, "UTF-8//TRANSLIT", $body);
 200                 }
 201
 202                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', "UTF-8");
 203
 204                 $doc = new DOMDocument();
 205                 @$doc->loadHTML($body);
 206
 207                 XML::deleteNode($doc, "style");
 208                 XML::deleteNode($doc, "script");
 209                 XML::deleteNode($doc, "option");
 210                 XML::deleteNode($doc, "h1");
 211                 XML::deleteNode($doc, "h2");
 212                 XML::deleteNode($doc, "h3");
 213                 XML::deleteNode($doc, "h4");
 214                 XML::deleteNode($doc, "h5");
 215                 XML::deleteNode($doc, "h6");
 216                 XML::deleteNode($doc, "ol");
 217                 XML::deleteNode($doc, "ul");
 218
 219                 $xpath = new DOMXPath($doc);
 220
 221                 $list = $xpath->query("//meta[@content]");
 222                 foreach ($list as $node) {
 223                         $attr = array();
 224                         if ($node->attributes->length) {
 225                                 foreach ($node->attributes as $attribute) {
 226                                         $attr[$attribute->name] = $attribute->value;
 227                                 }
 228                         }
 229
 230                         if (@$attr["http-equiv"] == "refresh") {
 231                                 $path = $attr["content"];
 232                                 $pathinfo = explode(";", $path);
 233                                 $content = "";
 234                                 foreach ($pathinfo as $value) {
 235                                         if (substr(strtolower($value), 0, 4) == "url=") {
 236                                                 $content = substr($value, 4);
 237                                         }
 238                                 }
 239                                 if ($content != "") {
 240                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 241                                         return($siteinfo);
 242                                 }
 243                         }
 244                 }
 245
 246                 $list = $xpath->query("//title");
 247                 if ($list->length > 0) {
 248                         $siteinfo["title"] = trim($list->item(0)->nodeValue);
 249                 }
 250
 251                 //$list = $xpath->query("head/meta[@name]");
 252                 $list = $xpath->query("//meta[@name]");
 253                 foreach ($list as $node) {
 254                         $attr = array();
 255                         if ($node->attributes->length) {
 256                                 foreach ($node->attributes as $attribute) {
 257                                         $attr[$attribute->name] = $attribute->value;
 258                                 }
 259                         }
 260
 261                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 262
 263                         if ($attr["content"] != "") {
 264                                 switch (strtolower($attr["name"])) {
 265                                         case "fulltitle":
 266                                                 $siteinfo["title"] = trim($attr["content"]);
 267                                                 break;
 268                                         case "description":
 269                                                 $siteinfo["text"] = trim($attr["content"]);
 270                                                 break;
 271                                         case "thumbnail":
 272                                                 $siteinfo["image"] = $attr["content"];
 273                                                 break;
 274                                         case "twitter:image":
 275                                                 $siteinfo["image"] = $attr["content"];
 276                                                 break;
 277                                         case "twitter:image:src":
 278                                                 $siteinfo["image"] = $attr["content"];
 279                                                 break;
 280                                         case "twitter:card":
 281                                                 if (($siteinfo["type"] == "") || ($attr["content"] == "photo")) {
 282                                                         $siteinfo["type"] = $attr["content"];
 283                                                 }
 284                                                 break;
 285                                         case "twitter:description":
 286                                                 $siteinfo["text"] = trim($attr["content"]);
 287                                                 break;
 288                                         case "twitter:title":
 289                                                 $siteinfo["title"] = trim($attr["content"]);
 290                                                 break;
 291                                         case "dc.title":
 292                                                 $siteinfo["title"] = trim($attr["content"]);
 293                                                 break;
 294                                         case "dc.description":
 295                                                 $siteinfo["text"] = trim($attr["content"]);
 296                                                 break;
 297                                         case "keywords":
 298                                                 $keywords = explode(",", $attr["content"]);
 299                                                 break;
 300                                         case "news_keywords":
 301                                                 $keywords = explode(",", $attr["content"]);
 302                                                 break;
 303                                 }
 304                         }
 305                         if ($siteinfo["type"] == "summary") {
 306                                 $siteinfo["type"] = "link";
 307                         }
 308                 }
 309
 310                 if (isset($keywords)) {
 311                         $siteinfo["keywords"] = array();
 312                         foreach ($keywords as $keyword) {
 313                                 if (!in_array(trim($keyword), $siteinfo["keywords"])) {
 314                                         $siteinfo["keywords"][] = trim($keyword);
 315                                 }
 316                         }
 317                 }
 318
 319                 //$list = $xpath->query("head/meta[@property]");
 320                 $list = $xpath->query("//meta[@property]");
 321                 foreach ($list as $node) {
 322                         $attr = array();
 323                         if ($node->attributes->length) {
 324                                 foreach ($node->attributes as $attribute) {
 325                                         $attr[$attribute->name] = $attribute->value;
 326                                 }
 327                         }
 328
 329                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 330
 331                         if ($attr["content"] != "") {
 332                                 switch (strtolower($attr["property"])) {
 333                                         case "og:image":
 334                                                 $siteinfo["image"] = $attr["content"];
 335                                                 break;
 336                                         case "og:title":
 337                                                 $siteinfo["title"] = trim($attr["content"]);
 338                                                 break;
 339                                         case "og:description":
 340                                                 $siteinfo["text"] = trim($attr["content"]);
 341                                                 break;
 342                                 }
 343                         }
 344                 }
 345
 346                 if ((@$siteinfo["image"] == "") && !$no_guessing) {
 347                         $list = $xpath->query("//img[@src]");
 348                         foreach ($list as $node) {
 349                                 $attr = array();
 350                                 if ($node->attributes->length) {
 351                                         foreach ($node->attributes as $attribute) {
 352                                                 $attr[$attribute->name] = $attribute->value;
 353                                         }
 354                                 }
 355
 356                                 $src = self::completeUrl($attr["src"], $url);
 357                                 $photodata = Image::getInfoFromURL($src);
 358
 359                                 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
 360                                         if ($photodata[0] > 300) {
 361                                                 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
 362                                                 $photodata[0] = 300;
 363                                         }
 364                                         if ($photodata[1] > 300) {
 365                                                 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
 366                                                 $photodata[1] = 300;
 367                                         }
 368                                         $siteinfo["images"][] = array("src" => $src,
 369                                                                         "width" => $photodata[0],
 370                                                                         "height" => $photodata[1]);
 371                                 }
 372                         }
 373                 } elseif ($siteinfo["image"] != "") {
 374                         $src = self::completeUrl($siteinfo["image"], $url);
 375
 376                         unset($siteinfo["image"]);
 377
 378                         $photodata = Image::getInfoFromURL($src);
 379
 380                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 381                                 $siteinfo["images"][] = array("src" => $src,
 382                                                                 "width" => $photodata[0],
 383                                                                 "height" => $photodata[1]);
 384                         }
 385                 }
 386
 387                 if ((@$siteinfo["text"] == "") && (@$siteinfo["title"] != "") && !$no_guessing) {
 388                         $text = "";
 389
 390                         $list = $xpath->query("//div[@class='article']");
 391                         foreach ($list as $node) {
 392                                 if (strlen($node->nodeValue) > 40) {
 393                                         $text .= " ".trim($node->nodeValue);
 394                                 }
 395                         }
 396
 397                         if ($text == "") {
 398                                 $list = $xpath->query("//div[@class='content']");
 399                                 foreach ($list as $node) {
 400                                         if (strlen($node->nodeValue) > 40) {
 401                                                 $text .= " ".trim($node->nodeValue);
 402                                         }
 403                                 }
 404                         }
 405
 406                         // If none text was found then take the paragraph content
 407                         if ($text == "") {
 408                                 $list = $xpath->query("//p");
 409                                 foreach ($list as $node) {
 410                                         if (strlen($node->nodeValue) > 40) {
 411                                                 $text .= " ".trim($node->nodeValue);
 412                                         }
 413                                 }
 414                         }
 415
 416                         if ($text != "") {
 417                                 $text = trim(str_replace(array("\n", "\r"), array(" ", " "), $text));
 418
 419                                 while (strpos($text, "  ")) {
 420                                         $text = trim(str_replace("  ", " ", $text));
 421                                 }
 422
 423                                 $siteinfo["text"] = trim(html_entity_decode(substr($text, 0, 350), ENT_QUOTES, "UTF-8").'...');
 424                         }
 425                 }
 426
 427                 logger("parseurl_getsiteinfo: Siteinfo for ".$url." ".print_r($siteinfo, true), LOGGER_DEBUG);
 428
 429                 call_hooks("getsiteinfo", $siteinfo);
 430
 431                 return($siteinfo);
 432         }
 433
 434         /**
 435          * @brief Convert tags from CSV to an array
 436          *
 437          * @param string $string Tags
 438          * @return array with formatted Hashtags
 439          */
 440         public static function convertTagsToArray($string)
 441         {
 442                 $arr_tags = str_getcsv($string);
 443                 if (count($arr_tags)) {
 444                         // add the # sign to every tag
 445                         array_walk($arr_tags, array("self", "arrAddHashes"));
 446
 447                         return $arr_tags;
 448                 }
 449         }
 450
 451         /**
 452          * @brief Add a hasht sign to a string
 453          *
 454          *  This method is used as callback function
 455          *
 456          * @param string $tag The pure tag name
 457          * @param int    $k   Counter for internal use
 458          * @return void
 459          */
 460         private static function arrAddHashes(&$tag, $k)
 461         {
 462                 $tag = "#" . $tag;
 463         }
 464
 465         /**
 466          * @brief Add a scheme to an url
 467          *
 468          * The src attribute of some html elements (e.g. images)
 469          * can miss the scheme so we need to add the correct
 470          * scheme
 471          *
 472          * @param string $url    The url which possibly does have
 473          *                       a missing scheme (a link to an image)
 474          * @param string $scheme The url with a correct scheme
 475          *                       (e.g. the url from the webpage which does contain the image)
 476          *
 477          * @return string The url with a scheme
 478          */
 479         private static function completeUrl($url, $scheme)
 480         {
 481                 $urlarr = parse_url($url);
 482
 483                 // If the url does allready have an scheme
 484                 // we can stop the process here
 485                 if (isset($urlarr["scheme"])) {
 486                         return($url);
 487                 }
 488
 489                 $schemearr = parse_url($scheme);
 490
 491                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 492
 493                 if (@$schemearr["port"] != "") {
 494                         $complete .= ":".$schemearr["port"];
 495                 }
 496
 497                 if (strpos($urlarr["path"], "/") !== 0) {
 498                         $complete .= "/";
 499                 }
 500
 501                 $complete .= $urlarr["path"];
 502
 503                 if (@$urlarr["query"] != "") {
 504                         $complete .= "?".$urlarr["query"];
 505                 }
 506
 507                 if (@$urlarr["fragment"] != "") {
 508                         $complete .= "#".$urlarr["fragment"];
 509                 }
 510
 511                 return($complete);
 512         }
 513 }