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