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                 $url = original_url($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($ch, CURLOPT_NOBODY, 1);
 146                 curl_setopt($ch, CURLOPT_TIMEOUT, 3);
 147                 curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
 148                 curl_setopt($ch, CURLOPT_USERAGENT, $a->get_useragent());
 149                 curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, (($check_cert) ? true : false));
 150                 curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, (($check_cert) ? 2 : false));
 151
 152                 $header = curl_exec($ch);
 153                 $curl_info = @curl_getinfo($ch);
 154                 $http_code = $curl_info["http_code"];
 155                 curl_close($ch);
 156
 157                 $a->save_timestamp($stamp1, "network");
 158
 159                 if ((($curl_info["http_code"] == "301") || ($curl_info["http_code"] == "302") || ($curl_info["http_code"] == "303") || ($curl_info["http_code"] == "307"))
 160                         && (($curl_info["redirect_url"] != "") || ($curl_info["location"] != ""))) {
 161                         if ($curl_info["redirect_url"] != "") {
 162                                 $siteinfo = self::getSiteinfo($curl_info["redirect_url"], $no_guessing, $do_oembed, ++$count);
 163                         } else {
 164                                 $siteinfo = self::getSiteinfo($curl_info["location"], $no_guessing, $do_oembed, ++$count);
 165                         }
 166                         return($siteinfo);
 167                 }
 168
 169                 // If the file is too large then exit
 170                 if ($curl_info["download_content_length"] > 1000000) {
 171                         return($siteinfo);
 172                 }
 173
 174                 // If it isn't a HTML file then exit
 175                 if (($curl_info["content_type"] != "") && !strstr(strtolower($curl_info["content_type"]), "html")) {
 176                         return($siteinfo);
 177                 }
 178
 179                 if ($do_oembed) {
 180
 181                         $oembed_data = oembed_fetch_url($url);
 182
 183                         if (!in_array($oembed_data->type, array("error", "rich"))) {
 184                                 $siteinfo["type"] = $oembed_data->type;
 185                         }
 186
 187                         if (($oembed_data->type == "link") && ($siteinfo["type"] != "photo")) {
 188                                 if (isset($oembed_data->title)) {
 189                                         $siteinfo["title"] = $oembed_data->title;
 190                                 }
 191                                 if (isset($oembed_data->description)) {
 192                                         $siteinfo["text"] = trim($oembed_data->description);
 193                                 }
 194                                 if (isset($oembed_data->thumbnail_url)) {
 195                                         $siteinfo["image"] = $oembed_data->thumbnail_url;
 196                                 }
 197                         }
 198                 }
 199
 200                 $stamp1 = microtime(true);
 201
 202                 // Now fetch the body as well
 203                 $ch = curl_init();
 204                 curl_setopt($ch, CURLOPT_URL, $url);
 205                 curl_setopt($ch, CURLOPT_HEADER, 1);
 206                 curl_setopt($ch, CURLOPT_NOBODY, 0);
 207                 curl_setopt($ch, CURLOPT_TIMEOUT, 10);
 208                 curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
 209                 curl_setopt($ch, CURLOPT_USERAGENT, $a->get_useragent());
 210                 curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, (($check_cert) ? true : false));
 211                 curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, (($check_cert) ? 2 : false));
 212
 213                 $header = curl_exec($ch);
 214                 $curl_info = @curl_getinfo($ch);
 215                 $http_code = $curl_info["http_code"];
 216                 curl_close($ch);
 217
 218                 $a->save_timestamp($stamp1, "network");
 219
 220                 // Fetch the first mentioned charset. Can be in body or header
 221                 $charset = "";
 222                 if (preg_match('/charset=(.*?)['."'".'"\s\n]/', $header, $matches)) {
 223                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 224                 }
 225
 226                 if ($charset == "") {
 227                         $charset = "utf-8";
 228                 }
 229
 230                 $pos = strpos($header, "\r\n\r\n");
 231
 232                 if ($pos) {
 233                         $body = trim(substr($header, $pos));
 234                 } else {
 235                         $body = $header;
 236                 }
 237
 238                 if (($charset != "") && (strtoupper($charset) != "UTF-8")) {
 239                         logger("parseurl_getsiteinfo: detected charset ".$charset, LOGGER_DEBUG);
 240                         //$body = mb_convert_encoding($body, "UTF-8", $charset);
 241                         $body = iconv($charset, "UTF-8//TRANSLIT", $body);
 242                 }
 243
 244                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', "UTF-8");
 245
 246                 $doc = new \DOMDocument();
 247                 @$doc->loadHTML($body);
 248
 249                 \xml::deleteNode($doc, "style");
 250                 \xml::deleteNode($doc, "script");
 251                 \xml::deleteNode($doc, "option");
 252                 \xml::deleteNode($doc, "h1");
 253                 \xml::deleteNode($doc, "h2");
 254                 \xml::deleteNode($doc, "h3");
 255                 \xml::deleteNode($doc, "h4");
 256                 \xml::deleteNode($doc, "h5");
 257                 \xml::deleteNode($doc, "h6");
 258                 \xml::deleteNode($doc, "ol");
 259                 \xml::deleteNode($doc, "ul");
 260
 261                 $xpath = new \DomXPath($doc);
 262
 263                 $list = $xpath->query("//meta[@content]");
 264                 foreach ($list as $node) {
 265                         $attr = array();
 266                         if ($node->attributes->length) {
 267                                 foreach ($node->attributes as $attribute) {
 268                                         $attr[$attribute->name] = $attribute->value;
 269                                 }
 270                         }
 271
 272                         if (@$attr["http-equiv"] == "refresh") {
 273                                 $path = $attr["content"];
 274                                 $pathinfo = explode(";", $path);
 275                                 $content = "";
 276                                 foreach ($pathinfo as $value) {
 277                                         if (substr(strtolower($value), 0, 4) == "url=") {
 278                                                 $content = substr($value, 4);
 279                                         }
 280                                 }
 281                                 if ($content != "") {
 282                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 283                                         return($siteinfo);
 284                                 }
 285                         }
 286                 }
 287
 288                 $list = $xpath->query("//title");
 289                 if ($list->length > 0) {
 290                         $siteinfo["title"] = $list->item(0)->nodeValue;
 291                 }
 292
 293                 //$list = $xpath->query("head/meta[@name]");
 294                 $list = $xpath->query("//meta[@name]");
 295                 foreach ($list as $node) {
 296                         $attr = array();
 297                         if ($node->attributes->length) {
 298                                 foreach ($node->attributes as $attribute) {
 299                                         $attr[$attribute->name] = $attribute->value;
 300                                 }
 301                         }
 302
 303                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 304
 305                         if ($attr["content"] != "") {
 306                                 switch (strtolower($attr["name"])) {
 307                                         case "fulltitle":
 308                                                 $siteinfo["title"] = $attr["content"];
 309                                                 break;
 310                                         case "description":
 311                                                 $siteinfo["text"] = $attr["content"];
 312                                                 break;
 313                                         case "thumbnail":
 314                                                 $siteinfo["image"] = $attr["content"];
 315                                                 break;
 316                                         case "twitter:image":
 317                                                 $siteinfo["image"] = $attr["content"];
 318                                                 break;
 319                                         case "twitter:image:src":
 320                                                 $siteinfo["image"] = $attr["content"];
 321                                                 break;
 322                                         case "twitter:card":
 323                                                 if (($siteinfo["type"] == "") || ($attr["content"] == "photo")) {
 324                                                         $siteinfo["type"] = $attr["content"];
 325                                                 }
 326                                                 break;
 327                                         case "twitter:description":
 328                                                 $siteinfo["text"] = $attr["content"];
 329                                                 break;
 330                                         case "twitter:title":
 331                                                 $siteinfo["title"] = $attr["content"];
 332                                                 break;
 333                                         case "dc.title":
 334                                                 $siteinfo["title"] = $attr["content"];
 335                                                 break;
 336                                         case "dc.description":
 337                                                 $siteinfo["text"] = $attr["content"];
 338                                                 break;
 339                                         case "keywords":
 340                                                 $keywords = explode(",", $attr["content"]);
 341                                                 break;
 342                                         case "news_keywords":
 343                                                 $keywords = explode(",", $attr["content"]);
 344                                                 break;
 345                                 }
 346                         }
 347                         if ($siteinfo["type"] == "summary") {
 348                                 $siteinfo["type"] = "link";
 349                         }
 350                 }
 351
 352                 if (isset($keywords)) {
 353                         $siteinfo["keywords"] = array();
 354                         foreach ($keywords as $keyword) {
 355                                 if (!in_array(trim($keyword), $siteinfo["keywords"])) {
 356                                         $siteinfo["keywords"][] = trim($keyword);
 357                                 }
 358                         }
 359                 }
 360
 361                 //$list = $xpath->query("head/meta[@property]");
 362                 $list = $xpath->query("//meta[@property]");
 363                 foreach ($list as $node) {
 364                         $attr = array();
 365                         if ($node->attributes->length) {
 366                                 foreach ($node->attributes as $attribute) {
 367                                         $attr[$attribute->name] = $attribute->value;
 368                                 }
 369                         }
 370
 371                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 372
 373                         if ($attr["content"] != "") {
 374                                 switch (strtolower($attr["property"])) {
 375                                         case "og:image":
 376                                                 $siteinfo["image"] = $attr["content"];
 377                                                 break;
 378                                         case "og:title":
 379                                                 $siteinfo["title"] = $attr["content"];
 380                                                 break;
 381                                         case "og:description":
 382                                                 $siteinfo["text"] = $attr["content"];
 383                                                 break;
 384                                 }
 385                         }
 386                 }
 387
 388                 if ((@$siteinfo["image"] == "") && !$no_guessing) {
 389                         $list = $xpath->query("//img[@src]");
 390                         foreach ($list as $node) {
 391                                 $attr = array();
 392                                 if ($node->attributes->length) {
 393                                         foreach ($node->attributes as $attribute) {
 394                                                 $attr[$attribute->name] = $attribute->value;
 395                                         }
 396                                 }
 397
 398                                 $src = self::completeUrl($attr["src"], $url);
 399                                 $photodata = get_photo_info($src);
 400
 401                                 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
 402                                         if ($photodata[0] > 300) {
 403                                                 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
 404                                                 $photodata[0] = 300;
 405                                         }
 406                                         if ($photodata[1] > 300) {
 407                                                 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
 408                                                 $photodata[1] = 300;
 409                                         }
 410                                         $siteinfo["images"][] = array("src" => $src,
 411                                                                         "width" => $photodata[0],
 412                                                                         "height" => $photodata[1]);
 413                                 }
 414
 415                                 }
 416                 } elseif ($siteinfo["image"] != "") {
 417                         $src = self::completeUrl($siteinfo["image"], $url);
 418
 419                         unset($siteinfo["image"]);
 420
 421                         $photodata = get_photo_info($src);
 422
 423                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 424                                 $siteinfo["images"][] = array("src" => $src,
 425                                                                 "width" => $photodata[0],
 426                                                                 "height" => $photodata[1]);
 427                         }
 428                 }
 429
 430                 if ((@$siteinfo["text"] == "") && (@$siteinfo["title"] != "") && !$no_guessing) {
 431                         $text = "";
 432
 433                         $list = $xpath->query("//div[@class='article']");
 434                         foreach ($list as $node) {
 435                                 if (strlen($node->nodeValue) > 40) {
 436                                         $text .= " ".trim($node->nodeValue);
 437                                 }
 438                         }
 439
 440                         if ($text == "") {
 441                                 $list = $xpath->query("//div[@class='content']");
 442                                 foreach ($list as $node) {
 443                                         if (strlen($node->nodeValue) > 40) {
 444                                                 $text .= " ".trim($node->nodeValue);
 445                                         }
 446                                 }
 447                         }
 448
 449                         // If none text was found then take the paragraph content
 450                         if ($text == "") {
 451                                 $list = $xpath->query("//p");
 452                                 foreach ($list as $node) {
 453                                         if (strlen($node->nodeValue) > 40) {
 454                                                 $text .= " ".trim($node->nodeValue);
 455                                         }
 456                                 }
 457                         }
 458
 459                         if ($text != "") {
 460                                 $text = trim(str_replace(array("\n", "\r"), array(" ", " "), $text));
 461
 462                                 while (strpos($text, "  ")) {
 463                                         $text = trim(str_replace("  ", " ", $text));
 464                                 }
 465
 466                                 $siteinfo["text"] = trim(html_entity_decode(substr($text, 0, 350), ENT_QUOTES, "UTF-8").'...');
 467                         }
 468                 }
 469
 470                 logger("parseurl_getsiteinfo: Siteinfo for ".$url." ".print_r($siteinfo, true), LOGGER_DEBUG);
 471
 472                 call_hooks("getsiteinfo", $siteinfo);
 473
 474                 return($siteinfo);
 475         }
 476
 477         /**
 478          * @brief Convert tags from CSV to an array
 479          *
 480          * @param string $string Tags
 481          * @return array with formatted Hashtags
 482          */
 483         public static function convertTagsToArray($string) {
 484                 $arr_tags = str_getcsv($string);
 485                 if (count($arr_tags)) {
 486                         // add the # sign to every tag
 487                         array_walk($arr_tags, array("self", "arrAddHashes"));
 488
 489                         return $arr_tags;
 490                 }
 491         }
 492
 493         /**
 494          * @brief Add a hasht sign to a string
 495          *
 496          *  This method is used as callback function
 497          *
 498          * @param string $tag The pure tag name
 499          * @param int $k Counter for internal use
 500          */
 501         private static function arrAddHashes(&$tag, $k) {
 502                 $tag = "#" . $tag;
 503         }
 504
 505         /**
 506          * @brief Add a scheme to an url
 507          *
 508          * The src attribute of some html elements (e.g. images)
 509          * can miss the scheme so we need to add the correct
 510          * scheme
 511          *
 512          * @param string $url The url which possibly does have
 513          *    a missing scheme (a link to an image)
 514          * @param string $scheme The url with a correct scheme
 515          *    (e.g. the url from the webpage which does contain the image)
 516          *
 517          * @return string The url with a scheme
 518          */
 519         private static function completeUrl($url, $scheme) {
 520                 $urlarr = parse_url($url);
 521
 522                 // If the url does allready have an scheme
 523                 // we can stop the process here
 524                 if (isset($urlarr["scheme"])) {
 525                         return($url);
 526                 }
 527
 528                 $schemearr = parse_url($scheme);
 529
 530                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 531
 532                 if (@$schemearr["port"] != "") {
 533                         $complete .= ":".$schemearr["port"];
 534                 }
 535
 536                 if (strpos($urlarr["path"],"/") !== 0) {
 537                         $complete .= "/";
 538                 }
 539
 540                 $complete .= $urlarr["path"];
 541
 542                 if (@$urlarr["query"] != "") {
 543                         $complete .= "?".$urlarr["query"];
 544                 }
 545
 546                 if (@$urlarr["fragment"] != "") {
 547                         $complete .= "#".$urlarr["fragment"];
 548                 }
 549
 550                 return($complete);
 551         }
 552 }