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