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