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 use dba;
  14
  15 use DomXPath;
  16 use DOMDocument;
  17
  18 require_once("include/network.php");
  19 require_once("include/Photo.php");
  20 require_once("include/oembed.php");
  21 require_once("include/xml.php");
  22
  23 /**
  24  * @brief Class with methods for extracting certain content from an url
  25  */
  26 class ParseUrl {
  27
  28         /**
  29          * @brief Search for chached embeddable data of an url otherwise fetch it
  30          *
  31          * @param type $url The url of the page which should be scraped
  32          * @param type $no_guessing If true the parse doens't search for
  33          *    preview pictures
  34          * @param type $do_oembed The false option is used by the function fetch_oembed()
  35          *    to avoid endless loops
  36          *
  37          * @return array which contains needed data for embedding
  38          *    string 'url' => The url of the parsed page
  39          *    string 'type' => Content type
  40          *    string 'title' => The title of the content
  41          *    string 'text' => The description for the content
  42          *    string 'image' => A preview image of the content (only available
  43          *                if $no_geuessing = false
  44          *    array'images' = Array of preview pictures
  45          *    string 'keywords' => The tags which belong to the content
  46          *
  47          * @see ParseUrl::getSiteinfo() for more information about scraping
  48          * embeddable content
  49          */
  50         public static function getSiteinfoCached($url, $no_guessing = false, $do_oembed = true) {
  51
  52                 if ($url == "") {
  53                         return false;
  54                 }
  55
  56                 $r = q("SELECT * FROM `parsed_url` WHERE `url` = '%s' AND `guessing` = %d AND `oembed` = %d",
  57                         dbesc(normalise_link($url)), intval(!$no_guessing), intval($do_oembed));
  58
  59                 if ($r) {
  60                         $data = $r[0]["content"];
  61                 }
  62
  63                 if (!is_null($data)) {
  64                         $data = unserialize($data);
  65                         return $data;
  66                 }
  67
  68                 $data = self::getSiteinfo($url, $no_guessing, $do_oembed);
  69
  70                 dba::insert('parsed_url', array('url' => normalise_link($url), 'guessing' => !$no_guessing,
  71                                 'oembed' => $do_oembed, 'content' => serialize($data),
  72                                 'created' => datetime_convert()), true);
  73
  74                 return $data;
  75         }
  76         /**
  77          * @brief Parse a page for embeddable content information
  78          *
  79          * This method parses to url for meta data which can be used to embed
  80          * the content. If available it prioritizes Open Graph meta tags.
  81          * If this is not available it uses the twitter cards meta tags.
  82          * As fallback it uses standard html elements with meta informations
  83          * like \<title\>Awesome Title\</title\> or
  84          * \<meta name="description" content="An awesome description"\>
  85          *
  86          * @param type $url The url of the page which should be scraped
  87          * @param type $no_guessing If true the parse doens't search for
  88          *    preview pictures
  89          * @param type $do_oembed The false option is used by the function fetch_oembed()
  90          *    to avoid endless loops
  91          * @param type $count Internal counter to avoid endless loops
  92          *
  93          * @return array which contains needed data for embedding
  94          *    string 'url' => The url of the parsed page
  95          *    string 'type' => Content type
  96          *    string 'title' => The title of the content
  97          *    string 'text' => The description for the content
  98          *    string 'image' => A preview image of the content (only available
  99          *                if $no_geuessing = false
 100          *    array'images' = Array of preview pictures
 101          *    string 'keywords' => The tags which belong to the content
 102          *
 103          * @todo https://developers.google.com/+/plugins/snippet/
 104          * @verbatim
 105          * <meta itemprop="name" content="Awesome title">
 106          * <meta itemprop="description" content="An awesome description">
 107          * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
 108          *
 109          * <body itemscope itemtype="http://schema.org/Product">
 110          *   <h1 itemprop="name">Shiny Trinket</h1>
 111          *   <img itemprop="image" src="{image-url}" />
 112          *   <p itemprop="description">Shiny trinkets are shiny.</p>
 113          * </body>
 114          * @endverbatim
 115          */
 116         public static function getSiteinfo($url, $no_guessing = false, $do_oembed = true, $count = 1) {
 117
 118                 $a = get_app();
 119
 120                 $siteinfo = array();
 121
 122                 // Check if the URL does contain a scheme
 123                 $scheme = parse_url($url, PHP_URL_SCHEME);
 124
 125                 if ($scheme == "") {
 126                         $url = "http://".trim($url, "/");
 127                 }
 128
 129                 if ($count > 10) {
 130                         logger("parseurl_getsiteinfo: Endless loop detected for ".$url, LOGGER_DEBUG);
 131                         return($siteinfo);
 132                 }
 133
 134                 $url = trim($url, "'");
 135                 $url = trim($url, '"');
 136
 137                 $url = strip_tracking_query_params($url);
 138
 139                 $siteinfo["url"] = $url;
 140                 $siteinfo["type"] = "link";
 141
 142                 $check_cert = Config::get("system", "verifyssl");
 143
 144                 $stamp1 = microtime(true);
 145
 146                 $ch = curl_init();
 147                 curl_setopt($ch, CURLOPT_URL, $url);
 148                 curl_setopt($ch, CURLOPT_HEADER, 1);
 149                 curl_setopt($ch, CURLOPT_TIMEOUT, 10);
 150                 curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
 151                 curl_setopt($ch, CURLOPT_USERAGENT, $a->get_useragent());
 152                 curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, (($check_cert) ? true : false));
 153                 if ($check_cert) {
 154                         @curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 2);
 155                 }
 156
 157                 $range = intval(Config::get('system', 'curl_range_bytes', 0));
 158
 159                 if ($range > 0) {
 160                         curl_setopt($ch, CURLOPT_RANGE, '0-' . $range);
 161                 }
 162
 163                 $header = curl_exec($ch);
 164                 $curl_info = @curl_getinfo($ch);
 165                 curl_close($ch);
 166
 167                 $a->save_timestamp($stamp1, "network");
 168
 169                 if ((($curl_info["http_code"] == "301") || ($curl_info["http_code"] == "302") || ($curl_info["http_code"] == "303") || ($curl_info["http_code"] == "307"))
 170                         && (($curl_info["redirect_url"] != "") || ($curl_info["location"] != ""))) {
 171                         if ($curl_info["redirect_url"] != "") {
 172                                 $siteinfo = self::getSiteinfo($curl_info["redirect_url"], $no_guessing, $do_oembed, ++$count);
 173                         } else {
 174                                 $siteinfo = self::getSiteinfo($curl_info["location"], $no_guessing, $do_oembed, ++$count);
 175                         }
 176                         return($siteinfo);
 177                 }
 178
 179                 // If the file is too large then exit
 180                 if ($curl_info["download_content_length"] > 1000000) {
 181                         return($siteinfo);
 182                 }
 183
 184                 // If it isn't a HTML file then exit
 185                 if (($curl_info["content_type"] != "") && !strstr(strtolower($curl_info["content_type"]), "html")) {
 186                         return($siteinfo);
 187                 }
 188
 189                 if ($do_oembed) {
 190
 191                         $oembed_data = oembed_fetch_url($url);
 192
 193                         if (!in_array($oembed_data->type, array("error", "rich"))) {
 194                                 $siteinfo["type"] = $oembed_data->type;
 195                         }
 196
 197                         if (($oembed_data->type == "link") && ($siteinfo["type"] != "photo")) {
 198                                 if (isset($oembed_data->title)) {
 199                                         $siteinfo["title"] = $oembed_data->title;
 200                                 }
 201                                 if (isset($oembed_data->description)) {
 202                                         $siteinfo["text"] = trim($oembed_data->description);
 203                                 }
 204                                 if (isset($oembed_data->thumbnail_url)) {
 205                                         $siteinfo["image"] = $oembed_data->thumbnail_url;
 206                                 }
 207                         }
 208                 }
 209
 210                 // Fetch the first mentioned charset. Can be in body or header
 211                 $charset = "";
 212                 if (preg_match('/charset=(.*?)['."'".'"\s\n]/', $header, $matches)) {
 213                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 214                 }
 215
 216                 if ($charset == "") {
 217                         $charset = "utf-8";
 218                 }
 219
 220                 $pos = strpos($header, "\r\n\r\n");
 221
 222                 if ($pos) {
 223                         $body = trim(substr($header, $pos));
 224                 } else {
 225                         $body = $header;
 226                 }
 227
 228                 if (($charset != "") && (strtoupper($charset) != "UTF-8")) {
 229                         logger("parseurl_getsiteinfo: detected charset ".$charset, LOGGER_DEBUG);
 230                         //$body = mb_convert_encoding($body, "UTF-8", $charset);
 231                         $body = iconv($charset, "UTF-8//TRANSLIT", $body);
 232                 }
 233
 234                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', "UTF-8");
 235
 236                 $doc = new DOMDocument();
 237                 @$doc->loadHTML($body);
 238
 239                 xml::deleteNode($doc, "style");
 240                 xml::deleteNode($doc, "script");
 241                 xml::deleteNode($doc, "option");
 242                 xml::deleteNode($doc, "h1");
 243                 xml::deleteNode($doc, "h2");
 244                 xml::deleteNode($doc, "h3");
 245                 xml::deleteNode($doc, "h4");
 246                 xml::deleteNode($doc, "h5");
 247                 xml::deleteNode($doc, "h6");
 248                 xml::deleteNode($doc, "ol");
 249                 xml::deleteNode($doc, "ul");
 250
 251                 $xpath = new DomXPath($doc);
 252
 253                 $list = $xpath->query("//meta[@content]");
 254                 foreach ($list as $node) {
 255                         $attr = array();
 256                         if ($node->attributes->length) {
 257                                 foreach ($node->attributes as $attribute) {
 258                                         $attr[$attribute->name] = $attribute->value;
 259                                 }
 260                         }
 261
 262                         if (@$attr["http-equiv"] == "refresh") {
 263                                 $path = $attr["content"];
 264                                 $pathinfo = explode(";", $path);
 265                                 $content = "";
 266                                 foreach ($pathinfo as $value) {
 267                                         if (substr(strtolower($value), 0, 4) == "url=") {
 268                                                 $content = substr($value, 4);
 269                                         }
 270                                 }
 271                                 if ($content != "") {
 272                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 273                                         return($siteinfo);
 274                                 }
 275                         }
 276                 }
 277
 278                 $list = $xpath->query("//title");
 279                 if ($list->length > 0) {
 280                         $siteinfo["title"] = $list->item(0)->nodeValue;
 281                 }
 282
 283                 //$list = $xpath->query("head/meta[@name]");
 284                 $list = $xpath->query("//meta[@name]");
 285                 foreach ($list as $node) {
 286                         $attr = array();
 287                         if ($node->attributes->length) {
 288                                 foreach ($node->attributes as $attribute) {
 289                                         $attr[$attribute->name] = $attribute->value;
 290                                 }
 291                         }
 292
 293                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 294
 295                         if ($attr["content"] != "") {
 296                                 switch (strtolower($attr["name"])) {
 297                                         case "fulltitle":
 298                                                 $siteinfo["title"] = $attr["content"];
 299                                                 break;
 300                                         case "description":
 301                                                 $siteinfo["text"] = $attr["content"];
 302                                                 break;
 303                                         case "thumbnail":
 304                                                 $siteinfo["image"] = $attr["content"];
 305                                                 break;
 306                                         case "twitter:image":
 307                                                 $siteinfo["image"] = $attr["content"];
 308                                                 break;
 309                                         case "twitter:image:src":
 310                                                 $siteinfo["image"] = $attr["content"];
 311                                                 break;
 312                                         case "twitter:card":
 313                                                 if (($siteinfo["type"] == "") || ($attr["content"] == "photo")) {
 314                                                         $siteinfo["type"] = $attr["content"];
 315                                                 }
 316                                                 break;
 317                                         case "twitter:description":
 318                                                 $siteinfo["text"] = $attr["content"];
 319                                                 break;
 320                                         case "twitter:title":
 321                                                 $siteinfo["title"] = $attr["content"];
 322                                                 break;
 323                                         case "dc.title":
 324                                                 $siteinfo["title"] = $attr["content"];
 325                                                 break;
 326                                         case "dc.description":
 327                                                 $siteinfo["text"] = $attr["content"];
 328                                                 break;
 329                                         case "keywords":
 330                                                 $keywords = explode(",", $attr["content"]);
 331                                                 break;
 332                                         case "news_keywords":
 333                                                 $keywords = explode(",", $attr["content"]);
 334                                                 break;
 335                                 }
 336                         }
 337                         if ($siteinfo["type"] == "summary") {
 338                                 $siteinfo["type"] = "link";
 339                         }
 340                 }
 341
 342                 if (isset($keywords)) {
 343                         $siteinfo["keywords"] = array();
 344                         foreach ($keywords as $keyword) {
 345                                 if (!in_array(trim($keyword), $siteinfo["keywords"])) {
 346                                         $siteinfo["keywords"][] = trim($keyword);
 347                                 }
 348                         }
 349                 }
 350
 351                 //$list = $xpath->query("head/meta[@property]");
 352                 $list = $xpath->query("//meta[@property]");
 353                 foreach ($list as $node) {
 354                         $attr = array();
 355                         if ($node->attributes->length) {
 356                                 foreach ($node->attributes as $attribute) {
 357                                         $attr[$attribute->name] = $attribute->value;
 358                                 }
 359                         }
 360
 361                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 362
 363                         if ($attr["content"] != "") {
 364                                 switch (strtolower($attr["property"])) {
 365                                         case "og:image":
 366                                                 $siteinfo["image"] = $attr["content"];
 367                                                 break;
 368                                         case "og:title":
 369                                                 $siteinfo["title"] = $attr["content"];
 370                                                 break;
 371                                         case "og:description":
 372                                                 $siteinfo["text"] = $attr["content"];
 373                                                 break;
 374                                 }
 375                         }
 376                 }
 377
 378                 if ((@$siteinfo["image"] == "") && !$no_guessing) {
 379                         $list = $xpath->query("//img[@src]");
 380                         foreach ($list as $node) {
 381                                 $attr = array();
 382                                 if ($node->attributes->length) {
 383                                         foreach ($node->attributes as $attribute) {
 384                                                 $attr[$attribute->name] = $attribute->value;
 385                                         }
 386                                 }
 387
 388                                 $src = self::completeUrl($attr["src"], $url);
 389                                 $photodata = get_photo_info($src);
 390
 391                                 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
 392                                         if ($photodata[0] > 300) {
 393                                                 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
 394                                                 $photodata[0] = 300;
 395                                         }
 396                                         if ($photodata[1] > 300) {
 397                                                 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
 398                                                 $photodata[1] = 300;
 399                                         }
 400                                         $siteinfo["images"][] = array("src" => $src,
 401                                                                         "width" => $photodata[0],
 402                                                                         "height" => $photodata[1]);
 403                                 }
 404
 405                                 }
 406                 } elseif ($siteinfo["image"] != "") {
 407                         $src = self::completeUrl($siteinfo["image"], $url);
 408
 409                         unset($siteinfo["image"]);
 410
 411                         $photodata = get_photo_info($src);
 412
 413                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 414                                 $siteinfo["images"][] = array("src" => $src,
 415                                                                 "width" => $photodata[0],
 416                                                                 "height" => $photodata[1]);
 417                         }
 418                 }
 419
 420                 if ((@$siteinfo["text"] == "") && (@$siteinfo["title"] != "") && !$no_guessing) {
 421                         $text = "";
 422
 423                         $list = $xpath->query("//div[@class='article']");
 424                         foreach ($list as $node) {
 425                                 if (strlen($node->nodeValue) > 40) {
 426                                         $text .= " ".trim($node->nodeValue);
 427                                 }
 428                         }
 429
 430                         if ($text == "") {
 431                                 $list = $xpath->query("//div[@class='content']");
 432                                 foreach ($list as $node) {
 433                                         if (strlen($node->nodeValue) > 40) {
 434                                                 $text .= " ".trim($node->nodeValue);
 435                                         }
 436                                 }
 437                         }
 438
 439                         // If none text was found then take the paragraph content
 440                         if ($text == "") {
 441                                 $list = $xpath->query("//p");
 442                                 foreach ($list as $node) {
 443                                         if (strlen($node->nodeValue) > 40) {
 444                                                 $text .= " ".trim($node->nodeValue);
 445                                         }
 446                                 }
 447                         }
 448
 449                         if ($text != "") {
 450                                 $text = trim(str_replace(array("\n", "\r"), array(" ", " "), $text));
 451
 452                                 while (strpos($text, "  ")) {
 453                                         $text = trim(str_replace("  ", " ", $text));
 454                                 }
 455
 456                                 $siteinfo["text"] = trim(html_entity_decode(substr($text, 0, 350), ENT_QUOTES, "UTF-8").'...');
 457                         }
 458                 }
 459
 460                 logger("parseurl_getsiteinfo: Siteinfo for ".$url." ".print_r($siteinfo, true), LOGGER_DEBUG);
 461
 462                 call_hooks("getsiteinfo", $siteinfo);
 463
 464                 return($siteinfo);
 465         }
 466
 467         /**
 468          * @brief Convert tags from CSV to an array
 469          *
 470          * @param string $string Tags
 471          * @return array with formatted Hashtags
 472          */
 473         public static function convertTagsToArray($string) {
 474                 $arr_tags = str_getcsv($string);
 475                 if (count($arr_tags)) {
 476                         // add the # sign to every tag
 477                         array_walk($arr_tags, array("self", "arrAddHashes"));
 478
 479                         return $arr_tags;
 480                 }
 481         }
 482
 483         /**
 484          * @brief Add a hasht sign to a string
 485          *
 486          *  This method is used as callback function
 487          *
 488          * @param string $tag The pure tag name
 489          * @param int $k Counter for internal use
 490          */
 491         private static function arrAddHashes(&$tag, $k) {
 492                 $tag = "#" . $tag;
 493         }
 494
 495         /**
 496          * @brief Add a scheme to an url
 497          *
 498          * The src attribute of some html elements (e.g. images)
 499          * can miss the scheme so we need to add the correct
 500          * scheme
 501          *
 502          * @param string $url The url which possibly does have
 503          *    a missing scheme (a link to an image)
 504          * @param string $scheme The url with a correct scheme
 505          *    (e.g. the url from the webpage which does contain the image)
 506          *
 507          * @return string The url with a scheme
 508          */
 509         private static function completeUrl($url, $scheme) {
 510                 $urlarr = parse_url($url);
 511
 512                 // If the url does allready have an scheme
 513                 // we can stop the process here
 514                 if (isset($urlarr["scheme"])) {
 515                         return($url);
 516                 }
 517
 518                 $schemearr = parse_url($scheme);
 519
 520                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 521
 522                 if (@$schemearr["port"] != "") {
 523                         $complete .= ":".$schemearr["port"];
 524                 }
 525
 526                 if (strpos($urlarr["path"],"/") !== 0) {
 527                         $complete .= "/";
 528                 }
 529
 530                 $complete .= $urlarr["path"];
 531
 532                 if (@$urlarr["query"] != "") {
 533                         $complete .= "?".$urlarr["query"];
 534                 }
 535
 536                 if (@$urlarr["fragment"] != "") {
 537                         $complete .= "#".$urlarr["fragment"];
 538                 }
 539
 540                 return($complete);
 541         }
 542 }