src/ParseUrl.php

   1 <?php
   2
   3 /**
   4  * @file include/ParseUrl.php
   5  * @brief Get informations about a given URL
   6  */
   7
   8 namespace Friendica;
   9
  10 use Friendica\Core\Config;
  11
  12 use xml;
  13
  14 use DomXPath;
  15 use DOMDocument;
  16
  17 require_once("include/network.php");
  18 require_once("include/Photo.php");
  19 require_once("include/oembed.php");
  20 require_once("include/xml.php");
  21
  22 /**
  23  * @brief Class with methods for extracting certain content from an url
  24  */
  25 class ParseUrl {
  26
  27         /**
  28          * @brief Search for chached embeddable data of an url otherwise fetch it
  29          *
  30          * @param type $url The url of the page which should be scraped
  31          * @param type $no_guessing If true the parse doens't search for
  32          *    preview pictures
  33          * @param type $do_oembed The false option is used by the function fetch_oembed()
  34          *    to avoid endless loops
  35          *
  36          * @return array which contains needed data for embedding
  37          *    string 'url' => The url of the parsed page
  38          *    string 'type' => Content type
  39          *    string 'title' => The title of the content
  40          *    string 'text' => The description for the content
  41          *    string 'image' => A preview image of the content (only available
  42          *                if $no_geuessing = false
  43          *    array'images' = Array of preview pictures
  44          *    string 'keywords' => The tags which belong to the content
  45          *
  46          * @see ParseUrl::getSiteinfo() for more information about scraping
  47          * embeddable content
  48          */
  49         public static function getSiteinfoCached($url, $no_guessing = false, $do_oembed = true) {
  50
  51                 if ($url == "") {
  52                         return false;
  53                 }
  54
  55                 $r = q("SELECT * FROM `parsed_url` WHERE `url` = '%s' AND `guessing` = %d AND `oembed` = %d",
  56                         dbesc(normalise_link($url)), intval(!$no_guessing), intval($do_oembed));
  57
  58                 if ($r) {
  59                         $data = $r[0]["content"];
  60                 }
  61
  62                 if (!is_null($data)) {
  63                         $data = unserialize($data);
  64                         return $data;
  65                 }
  66
  67                 $data = self::getSiteinfo($url, $no_guessing, $do_oembed);
  68
  69                 q("INSERT INTO `parsed_url` (`url`, `guessing`, `oembed`, `content`, `created`) VALUES ('%s', %d, %d, '%s', '%s')
  70                          ON DUPLICATE KEY UPDATE `content` = '%s', `created` = '%s'",
  71                         dbesc(normalise_link($url)), intval(!$no_guessing), intval($do_oembed),
  72                         dbesc(serialize($data)), dbesc(datetime_convert()),
  73                         dbesc(serialize($data)), dbesc(datetime_convert()));
  74
  75                 return $data;
  76         }
  77         /**
  78          * @brief Parse a page for embeddable content information
  79          *
  80          * This method parses to url for meta data which can be used to embed
  81          * the content. If available it prioritizes Open Graph meta tags.
  82          * If this is not available it uses the twitter cards meta tags.
  83          * As fallback it uses standard html elements with meta informations
  84          * like \<title\>Awesome Title\</title\> or
  85          * \<meta name="description" content="An awesome description"\>
  86          *
  87          * @param type $url The url of the page which should be scraped
  88          * @param type $no_guessing If true the parse doens't search for
  89          *    preview pictures
  90          * @param type $do_oembed The false option is used by the function fetch_oembed()
  91          *    to avoid endless loops
  92          * @param type $count Internal counter to avoid endless loops
  93          *
  94          * @return array which contains needed data for embedding
  95          *    string 'url' => The url of the parsed page
  96          *    string 'type' => Content type
  97          *    string 'title' => The title of the content
  98          *    string 'text' => The description for the content
  99          *    string 'image' => A preview image of the content (only available
 100          *                if $no_geuessing = false
 101          *    array'images' = Array of preview pictures
 102          *    string 'keywords' => The tags which belong to the content
 103          *
 104          * @todo https://developers.google.com/+/plugins/snippet/
 105          * @verbatim
 106          * <meta itemprop="name" content="Awesome title">
 107          * <meta itemprop="description" content="An awesome description">
 108          * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
 109          *
 110          * <body itemscope itemtype="http://schema.org/Product">
 111          *   <h1 itemprop="name">Shiny Trinket</h1>
 112          *   <img itemprop="image" src="{image-url}" />
 113          *   <p itemprop="description">Shiny trinkets are shiny.</p>
 114          * </body>
 115          * @endverbatim
 116          */
 117         public static function getSiteinfo($url, $no_guessing = false, $do_oembed = true, $count = 1) {
 118
 119                 $a = get_app();
 120
 121                 $siteinfo = array();
 122
 123                 // Check if the URL does contain a scheme
 124                 $scheme = parse_url($url, PHP_URL_SCHEME);
 125
 126                 if ($scheme == "") {
 127                         $url = "http://".trim($url, "/");
 128                 }
 129
 130                 if ($count > 10) {
 131                         logger("parseurl_getsiteinfo: Endless loop detected for ".$url, LOGGER_DEBUG);
 132                         return($siteinfo);
 133                 }
 134
 135                 $url = trim($url, "'");
 136                 $url = trim($url, '"');
 137
 138                 $url = strip_tracking_query_params($url);
 139
 140                 $siteinfo["url"] = $url;
 141                 $siteinfo["type"] = "link";
 142
 143                 $check_cert = Config::get("system", "verifyssl");
 144
 145                 $stamp1 = microtime(true);
 146
 147                 $ch = curl_init();
 148                 curl_setopt($ch, CURLOPT_URL, $url);
 149                 curl_setopt($ch, CURLOPT_HEADER, 1);
 150                 curl_setopt($ch, CURLOPT_TIMEOUT, 10);
 151                 curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
 152                 curl_setopt($ch, CURLOPT_USERAGENT, $a->get_useragent());
 153                 curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, (($check_cert) ? true : false));
 154                 if ($check_cert) {
 155                         @curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 2);
 156                 }
 157
 158                 $header = curl_exec($ch);
 159                 $curl_info = @curl_getinfo($ch);
 160                 curl_close($ch);
 161
 162                 $a->save_timestamp($stamp1, "network");
 163
 164                 if ((($curl_info["http_code"] == "301") || ($curl_info["http_code"] == "302") || ($curl_info["http_code"] == "303") || ($curl_info["http_code"] == "307"))
 165                         && (($curl_info["redirect_url"] != "") || ($curl_info["location"] != ""))) {
 166                         if ($curl_info["redirect_url"] != "") {
 167                                 $siteinfo = self::getSiteinfo($curl_info["redirect_url"], $no_guessing, $do_oembed, ++$count);
 168                         } else {
 169                                 $siteinfo = self::getSiteinfo($curl_info["location"], $no_guessing, $do_oembed, ++$count);
 170                         }
 171                         return($siteinfo);
 172                 }
 173
 174                 // If the file is too large then exit
 175                 if ($curl_info["download_content_length"] > 1000000) {
 176                         return($siteinfo);
 177                 }
 178
 179                 // If it isn't a HTML file then exit
 180                 if (($curl_info["content_type"] != "") && !strstr(strtolower($curl_info["content_type"]), "html")) {
 181                         return($siteinfo);
 182                 }
 183
 184                 if ($do_oembed) {
 185
 186                         $oembed_data = oembed_fetch_url($url);
 187
 188                         if (!in_array($oembed_data->type, array("error", "rich"))) {
 189                                 $siteinfo["type"] = $oembed_data->type;
 190                         }
 191
 192                         if (($oembed_data->type == "link") && ($siteinfo["type"] != "photo")) {
 193                                 if (isset($oembed_data->title)) {
 194                                         $siteinfo["title"] = $oembed_data->title;
 195                                 }
 196                                 if (isset($oembed_data->description)) {
 197                                         $siteinfo["text"] = trim($oembed_data->description);
 198                                 }
 199                                 if (isset($oembed_data->thumbnail_url)) {
 200                                         $siteinfo["image"] = $oembed_data->thumbnail_url;
 201                                 }
 202                         }
 203                 }
 204
 205                 // Fetch the first mentioned charset. Can be in body or header
 206                 $charset = "";
 207                 if (preg_match('/charset=(.*?)['."'".'"\s\n]/', $header, $matches)) {
 208                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 209                 }
 210
 211                 if ($charset == "") {
 212                         $charset = "utf-8";
 213                 }
 214
 215                 $pos = strpos($header, "\r\n\r\n");
 216
 217                 if ($pos) {
 218                         $body = trim(substr($header, $pos));
 219                 } else {
 220                         $body = $header;
 221                 }
 222
 223                 if (($charset != "") && (strtoupper($charset) != "UTF-8")) {
 224                         logger("parseurl_getsiteinfo: detected charset ".$charset, LOGGER_DEBUG);
 225                         //$body = mb_convert_encoding($body, "UTF-8", $charset);
 226                         $body = iconv($charset, "UTF-8//TRANSLIT", $body);
 227                 }
 228
 229                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', "UTF-8");
 230
 231                 $doc = new DOMDocument();
 232                 @$doc->loadHTML($body);
 233
 234                 xml::deleteNode($doc, "style");
 235                 xml::deleteNode($doc, "script");
 236                 xml::deleteNode($doc, "option");
 237                 xml::deleteNode($doc, "h1");
 238                 xml::deleteNode($doc, "h2");
 239                 xml::deleteNode($doc, "h3");
 240                 xml::deleteNode($doc, "h4");
 241                 xml::deleteNode($doc, "h5");
 242                 xml::deleteNode($doc, "h6");
 243                 xml::deleteNode($doc, "ol");
 244                 xml::deleteNode($doc, "ul");
 245
 246                 $xpath = new DomXPath($doc);
 247
 248                 $list = $xpath->query("//meta[@content]");
 249                 foreach ($list as $node) {
 250                         $attr = array();
 251                         if ($node->attributes->length) {
 252                                 foreach ($node->attributes as $attribute) {
 253                                         $attr[$attribute->name] = $attribute->value;
 254                                 }
 255                         }
 256
 257                         if (@$attr["http-equiv"] == "refresh") {
 258                                 $path = $attr["content"];
 259                                 $pathinfo = explode(";", $path);
 260                                 $content = "";
 261                                 foreach ($pathinfo as $value) {
 262                                         if (substr(strtolower($value), 0, 4) == "url=") {
 263                                                 $content = substr($value, 4);
 264                                         }
 265                                 }
 266                                 if ($content != "") {
 267                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 268                                         return($siteinfo);
 269                                 }
 270                         }
 271                 }
 272
 273                 $list = $xpath->query("//title");
 274                 if ($list->length > 0) {
 275                         $siteinfo["title"] = $list->item(0)->nodeValue;
 276                 }
 277
 278                 //$list = $xpath->query("head/meta[@name]");
 279                 $list = $xpath->query("//meta[@name]");
 280                 foreach ($list as $node) {
 281                         $attr = array();
 282                         if ($node->attributes->length) {
 283                                 foreach ($node->attributes as $attribute) {
 284                                         $attr[$attribute->name] = $attribute->value;
 285                                 }
 286                         }
 287
 288                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 289
 290                         if ($attr["content"] != "") {
 291                                 switch (strtolower($attr["name"])) {
 292                                         case "fulltitle":
 293                                                 $siteinfo["title"] = $attr["content"];
 294                                                 break;
 295                                         case "description":
 296                                                 $siteinfo["text"] = $attr["content"];
 297                                                 break;
 298                                         case "thumbnail":
 299                                                 $siteinfo["image"] = $attr["content"];
 300                                                 break;
 301                                         case "twitter:image":
 302                                                 $siteinfo["image"] = $attr["content"];
 303                                                 break;
 304                                         case "twitter:image:src":
 305                                                 $siteinfo["image"] = $attr["content"];
 306                                                 break;
 307                                         case "twitter:card":
 308                                                 if (($siteinfo["type"] == "") || ($attr["content"] == "photo")) {
 309                                                         $siteinfo["type"] = $attr["content"];
 310                                                 }
 311                                                 break;
 312                                         case "twitter:description":
 313                                                 $siteinfo["text"] = $attr["content"];
 314                                                 break;
 315                                         case "twitter:title":
 316                                                 $siteinfo["title"] = $attr["content"];
 317                                                 break;
 318                                         case "dc.title":
 319                                                 $siteinfo["title"] = $attr["content"];
 320                                                 break;
 321                                         case "dc.description":
 322                                                 $siteinfo["text"] = $attr["content"];
 323                                                 break;
 324                                         case "keywords":
 325                                                 $keywords = explode(",", $attr["content"]);
 326                                                 break;
 327                                         case "news_keywords":
 328                                                 $keywords = explode(",", $attr["content"]);
 329                                                 break;
 330                                 }
 331                         }
 332                         if ($siteinfo["type"] == "summary") {
 333                                 $siteinfo["type"] = "link";
 334                         }
 335                 }
 336
 337                 if (isset($keywords)) {
 338                         $siteinfo["keywords"] = array();
 339                         foreach ($keywords as $keyword) {
 340                                 if (!in_array(trim($keyword), $siteinfo["keywords"])) {
 341                                         $siteinfo["keywords"][] = trim($keyword);
 342                                 }
 343                         }
 344                 }
 345
 346                 //$list = $xpath->query("head/meta[@property]");
 347                 $list = $xpath->query("//meta[@property]");
 348                 foreach ($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                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 357
 358                         if ($attr["content"] != "") {
 359                                 switch (strtolower($attr["property"])) {
 360                                         case "og:image":
 361                                                 $siteinfo["image"] = $attr["content"];
 362                                                 break;
 363                                         case "og:title":
 364                                                 $siteinfo["title"] = $attr["content"];
 365                                                 break;
 366                                         case "og:description":
 367                                                 $siteinfo["text"] = $attr["content"];
 368                                                 break;
 369                                 }
 370                         }
 371                 }
 372
 373                 if ((@$siteinfo["image"] == "") && !$no_guessing) {
 374                         $list = $xpath->query("//img[@src]");
 375                         foreach ($list as $node) {
 376                                 $attr = array();
 377                                 if ($node->attributes->length) {
 378                                         foreach ($node->attributes as $attribute) {
 379                                                 $attr[$attribute->name] = $attribute->value;
 380                                         }
 381                                 }
 382
 383                                 $src = self::completeUrl($attr["src"], $url);
 384                                 $photodata = get_photo_info($src);
 385
 386                                 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
 387                                         if ($photodata[0] > 300) {
 388                                                 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
 389                                                 $photodata[0] = 300;
 390                                         }
 391                                         if ($photodata[1] > 300) {
 392                                                 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
 393                                                 $photodata[1] = 300;
 394                                         }
 395                                         $siteinfo["images"][] = array("src" => $src,
 396                                                                         "width" => $photodata[0],
 397                                                                         "height" => $photodata[1]);
 398                                 }
 399
 400                                 }
 401                 } elseif ($siteinfo["image"] != "") {
 402                         $src = self::completeUrl($siteinfo["image"], $url);
 403
 404                         unset($siteinfo["image"]);
 405
 406                         $photodata = get_photo_info($src);
 407
 408                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 409                                 $siteinfo["images"][] = array("src" => $src,
 410                                                                 "width" => $photodata[0],
 411                                                                 "height" => $photodata[1]);
 412                         }
 413                 }
 414
 415                 if ((@$siteinfo["text"] == "") && (@$siteinfo["title"] != "") && !$no_guessing) {
 416                         $text = "";
 417
 418                         $list = $xpath->query("//div[@class='article']");
 419                         foreach ($list as $node) {
 420                                 if (strlen($node->nodeValue) > 40) {
 421                                         $text .= " ".trim($node->nodeValue);
 422                                 }
 423                         }
 424
 425                         if ($text == "") {
 426                                 $list = $xpath->query("//div[@class='content']");
 427                                 foreach ($list as $node) {
 428                                         if (strlen($node->nodeValue) > 40) {
 429                                                 $text .= " ".trim($node->nodeValue);
 430                                         }
 431                                 }
 432                         }
 433
 434                         // If none text was found then take the paragraph content
 435                         if ($text == "") {
 436                                 $list = $xpath->query("//p");
 437                                 foreach ($list as $node) {
 438                                         if (strlen($node->nodeValue) > 40) {
 439                                                 $text .= " ".trim($node->nodeValue);
 440                                         }
 441                                 }
 442                         }
 443
 444                         if ($text != "") {
 445                                 $text = trim(str_replace(array("\n", "\r"), array(" ", " "), $text));
 446
 447                                 while (strpos($text, "  ")) {
 448                                         $text = trim(str_replace("  ", " ", $text));
 449                                 }
 450
 451                                 $siteinfo["text"] = trim(html_entity_decode(substr($text, 0, 350), ENT_QUOTES, "UTF-8").'...');
 452                         }
 453                 }
 454
 455                 logger("parseurl_getsiteinfo: Siteinfo for ".$url." ".print_r($siteinfo, true), LOGGER_DEBUG);
 456
 457                 call_hooks("getsiteinfo", $siteinfo);
 458
 459                 return($siteinfo);
 460         }
 461
 462         /**
 463          * @brief Convert tags from CSV to an array
 464          *
 465          * @param string $string Tags
 466          * @return array with formatted Hashtags
 467          */
 468         public static function convertTagsToArray($string) {
 469                 $arr_tags = str_getcsv($string);
 470                 if (count($arr_tags)) {
 471                         // add the # sign to every tag
 472                         array_walk($arr_tags, array("self", "arrAddHashes"));
 473
 474                         return $arr_tags;
 475                 }
 476         }
 477
 478         /**
 479          * @brief Add a hasht sign to a string
 480          *
 481          *  This method is used as callback function
 482          *
 483          * @param string $tag The pure tag name
 484          * @param int $k Counter for internal use
 485          */
 486         private static function arrAddHashes(&$tag, $k) {
 487                 $tag = "#" . $tag;
 488         }
 489
 490         /**
 491          * @brief Add a scheme to an url
 492          *
 493          * The src attribute of some html elements (e.g. images)
 494          * can miss the scheme so we need to add the correct
 495          * scheme
 496          *
 497          * @param string $url The url which possibly does have
 498          *    a missing scheme (a link to an image)
 499          * @param string $scheme The url with a correct scheme
 500          *    (e.g. the url from the webpage which does contain the image)
 501          *
 502          * @return string The url with a scheme
 503          */
 504         private static function completeUrl($url, $scheme) {
 505                 $urlarr = parse_url($url);
 506
 507                 // If the url does allready have an scheme
 508                 // we can stop the process here
 509                 if (isset($urlarr["scheme"])) {
 510                         return($url);
 511                 }
 512
 513                 $schemearr = parse_url($scheme);
 514
 515                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 516
 517                 if (@$schemearr["port"] != "") {
 518                         $complete .= ":".$schemearr["port"];
 519                 }
 520
 521                 if (strpos($urlarr["path"],"/") !== 0) {
 522                         $complete .= "/";
 523                 }
 524
 525                 $complete .= $urlarr["path"];
 526
 527                 if (@$urlarr["query"] != "") {
 528                         $complete .= "?".$urlarr["query"];
 529                 }
 530
 531                 if (@$urlarr["fragment"] != "") {
 532                         $complete .= "#".$urlarr["fragment"];
 533                 }
 534
 535                 return($complete);
 536         }
 537 }