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                 $data = z_fetch_url($url);
 143                 if (!$data['success']) {
 144                         return($siteinfo);
 145                 }
 146
 147                 // If the file is too large then exit
 148                 if ($data["info"]["download_content_length"] > 1000000) {
 149                         return($siteinfo);
 150                 }
 151
 152                 // If it isn't a HTML file then exit
 153                 if (($data["info"]["content_type"] != "") && !strstr(strtolower($data["info"]["content_type"]), "html")) {
 154                         return($siteinfo);
 155                 }
 156
 157                 $header = $data["header"];
 158                 $body = $data["body"];
 159
 160                 if ($do_oembed) {
 161
 162                         $oembed_data = oembed_fetch_url($url);
 163
 164                         if (!in_array($oembed_data->type, array("error", "rich"))) {
 165                                 $siteinfo["type"] = $oembed_data->type;
 166                         }
 167
 168                         if (($oembed_data->type == "link") && ($siteinfo["type"] != "photo")) {
 169                                 if (isset($oembed_data->title)) {
 170                                         $siteinfo["title"] = $oembed_data->title;
 171                                 }
 172                                 if (isset($oembed_data->description)) {
 173                                         $siteinfo["text"] = trim($oembed_data->description);
 174                                 }
 175                                 if (isset($oembed_data->thumbnail_url)) {
 176                                         $siteinfo["image"] = $oembed_data->thumbnail_url;
 177                                 }
 178                         }
 179                 }
 180
 181                 // Fetch the first mentioned charset. Can be in body or header
 182                 $charset = "";
 183                 if (preg_match('/charset=(.*?)['."'".'"\s\n]/', $header, $matches)) {
 184                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 185                 }
 186
 187                 if ($charset == "") {
 188                         $charset = "utf-8";
 189                 }
 190
 191                 if (($charset != "") && (strtoupper($charset) != "UTF-8")) {
 192                         logger("parseurl_getsiteinfo: detected charset ".$charset, LOGGER_DEBUG);
 193                         //$body = mb_convert_encoding($body, "UTF-8", $charset);
 194                         $body = iconv($charset, "UTF-8//TRANSLIT", $body);
 195                 }
 196
 197                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', "UTF-8");
 198
 199                 $doc = new DOMDocument();
 200                 @$doc->loadHTML($body);
 201
 202                 xml::deleteNode($doc, "style");
 203                 xml::deleteNode($doc, "script");
 204                 xml::deleteNode($doc, "option");
 205                 xml::deleteNode($doc, "h1");
 206                 xml::deleteNode($doc, "h2");
 207                 xml::deleteNode($doc, "h3");
 208                 xml::deleteNode($doc, "h4");
 209                 xml::deleteNode($doc, "h5");
 210                 xml::deleteNode($doc, "h6");
 211                 xml::deleteNode($doc, "ol");
 212                 xml::deleteNode($doc, "ul");
 213
 214                 $xpath = new DomXPath($doc);
 215
 216                 $list = $xpath->query("//meta[@content]");
 217                 foreach ($list as $node) {
 218                         $attr = array();
 219                         if ($node->attributes->length) {
 220                                 foreach ($node->attributes as $attribute) {
 221                                         $attr[$attribute->name] = $attribute->value;
 222                                 }
 223                         }
 224
 225                         if (@$attr["http-equiv"] == "refresh") {
 226                                 $path = $attr["content"];
 227                                 $pathinfo = explode(";", $path);
 228                                 $content = "";
 229                                 foreach ($pathinfo as $value) {
 230                                         if (substr(strtolower($value), 0, 4) == "url=") {
 231                                                 $content = substr($value, 4);
 232                                         }
 233                                 }
 234                                 if ($content != "") {
 235                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 236                                         return($siteinfo);
 237                                 }
 238                         }
 239                 }
 240
 241                 $list = $xpath->query("//title");
 242                 if ($list->length > 0) {
 243                         $siteinfo["title"] = $list->item(0)->nodeValue;
 244                 }
 245
 246                 //$list = $xpath->query("head/meta[@name]");
 247                 $list = $xpath->query("//meta[@name]");
 248                 foreach ($list as $node) {
 249                         $attr = array();
 250                         if ($node->attributes->length) {
 251                                 foreach ($node->attributes as $attribute) {
 252                                         $attr[$attribute->name] = $attribute->value;
 253                                 }
 254                         }
 255
 256                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 257
 258                         if ($attr["content"] != "") {
 259                                 switch (strtolower($attr["name"])) {
 260                                         case "fulltitle":
 261                                                 $siteinfo["title"] = $attr["content"];
 262                                                 break;
 263                                         case "description":
 264                                                 $siteinfo["text"] = $attr["content"];
 265                                                 break;
 266                                         case "thumbnail":
 267                                                 $siteinfo["image"] = $attr["content"];
 268                                                 break;
 269                                         case "twitter:image":
 270                                                 $siteinfo["image"] = $attr["content"];
 271                                                 break;
 272                                         case "twitter:image:src":
 273                                                 $siteinfo["image"] = $attr["content"];
 274                                                 break;
 275                                         case "twitter:card":
 276                                                 if (($siteinfo["type"] == "") || ($attr["content"] == "photo")) {
 277                                                         $siteinfo["type"] = $attr["content"];
 278                                                 }
 279                                                 break;
 280                                         case "twitter:description":
 281                                                 $siteinfo["text"] = $attr["content"];
 282                                                 break;
 283                                         case "twitter:title":
 284                                                 $siteinfo["title"] = $attr["content"];
 285                                                 break;
 286                                         case "dc.title":
 287                                                 $siteinfo["title"] = $attr["content"];
 288                                                 break;
 289                                         case "dc.description":
 290                                                 $siteinfo["text"] = $attr["content"];
 291                                                 break;
 292                                         case "keywords":
 293                                                 $keywords = explode(",", $attr["content"]);
 294                                                 break;
 295                                         case "news_keywords":
 296                                                 $keywords = explode(",", $attr["content"]);
 297                                                 break;
 298                                 }
 299                         }
 300                         if ($siteinfo["type"] == "summary") {
 301                                 $siteinfo["type"] = "link";
 302                         }
 303                 }
 304
 305                 if (isset($keywords)) {
 306                         $siteinfo["keywords"] = array();
 307                         foreach ($keywords as $keyword) {
 308                                 if (!in_array(trim($keyword), $siteinfo["keywords"])) {
 309                                         $siteinfo["keywords"][] = trim($keyword);
 310                                 }
 311                         }
 312                 }
 313
 314                 //$list = $xpath->query("head/meta[@property]");
 315                 $list = $xpath->query("//meta[@property]");
 316                 foreach ($list as $node) {
 317                         $attr = array();
 318                         if ($node->attributes->length) {
 319                                 foreach ($node->attributes as $attribute) {
 320                                         $attr[$attribute->name] = $attribute->value;
 321                                 }
 322                         }
 323
 324                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 325
 326                         if ($attr["content"] != "") {
 327                                 switch (strtolower($attr["property"])) {
 328                                         case "og:image":
 329                                                 $siteinfo["image"] = $attr["content"];
 330                                                 break;
 331                                         case "og:title":
 332                                                 $siteinfo["title"] = $attr["content"];
 333                                                 break;
 334                                         case "og:description":
 335                                                 $siteinfo["text"] = $attr["content"];
 336                                                 break;
 337                                 }
 338                         }
 339                 }
 340
 341                 if ((@$siteinfo["image"] == "") && !$no_guessing) {
 342                         $list = $xpath->query("//img[@src]");
 343                         foreach ($list as $node) {
 344                                 $attr = array();
 345                                 if ($node->attributes->length) {
 346                                         foreach ($node->attributes as $attribute) {
 347                                                 $attr[$attribute->name] = $attribute->value;
 348                                         }
 349                                 }
 350
 351                                 $src = self::completeUrl($attr["src"], $url);
 352                                 $photodata = get_photo_info($src);
 353
 354                                 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
 355                                         if ($photodata[0] > 300) {
 356                                                 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
 357                                                 $photodata[0] = 300;
 358                                         }
 359                                         if ($photodata[1] > 300) {
 360                                                 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
 361                                                 $photodata[1] = 300;
 362                                         }
 363                                         $siteinfo["images"][] = array("src" => $src,
 364                                                                         "width" => $photodata[0],
 365                                                                         "height" => $photodata[1]);
 366                                 }
 367
 368                                 }
 369                 } elseif ($siteinfo["image"] != "") {
 370                         $src = self::completeUrl($siteinfo["image"], $url);
 371
 372                         unset($siteinfo["image"]);
 373
 374                         $photodata = get_photo_info($src);
 375
 376                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 377                                 $siteinfo["images"][] = array("src" => $src,
 378                                                                 "width" => $photodata[0],
 379                                                                 "height" => $photodata[1]);
 380                         }
 381                 }
 382
 383                 if ((@$siteinfo["text"] == "") && (@$siteinfo["title"] != "") && !$no_guessing) {
 384                         $text = "";
 385
 386                         $list = $xpath->query("//div[@class='article']");
 387                         foreach ($list as $node) {
 388                                 if (strlen($node->nodeValue) > 40) {
 389                                         $text .= " ".trim($node->nodeValue);
 390                                 }
 391                         }
 392
 393                         if ($text == "") {
 394                                 $list = $xpath->query("//div[@class='content']");
 395                                 foreach ($list as $node) {
 396                                         if (strlen($node->nodeValue) > 40) {
 397                                                 $text .= " ".trim($node->nodeValue);
 398                                         }
 399                                 }
 400                         }
 401
 402                         // If none text was found then take the paragraph content
 403                         if ($text == "") {
 404                                 $list = $xpath->query("//p");
 405                                 foreach ($list as $node) {
 406                                         if (strlen($node->nodeValue) > 40) {
 407                                                 $text .= " ".trim($node->nodeValue);
 408                                         }
 409                                 }
 410                         }
 411
 412                         if ($text != "") {
 413                                 $text = trim(str_replace(array("\n", "\r"), array(" ", " "), $text));
 414
 415                                 while (strpos($text, "  ")) {
 416                                         $text = trim(str_replace("  ", " ", $text));
 417                                 }
 418
 419                                 $siteinfo["text"] = trim(html_entity_decode(substr($text, 0, 350), ENT_QUOTES, "UTF-8").'...');
 420                         }
 421                 }
 422
 423                 logger("parseurl_getsiteinfo: Siteinfo for ".$url." ".print_r($siteinfo, true), LOGGER_DEBUG);
 424
 425                 call_hooks("getsiteinfo", $siteinfo);
 426
 427                 return($siteinfo);
 428         }
 429
 430         /**
 431          * @brief Convert tags from CSV to an array
 432          *
 433          * @param string $string Tags
 434          * @return array with formatted Hashtags
 435          */
 436         public static function convertTagsToArray($string) {
 437                 $arr_tags = str_getcsv($string);
 438                 if (count($arr_tags)) {
 439                         // add the # sign to every tag
 440                         array_walk($arr_tags, array("self", "arrAddHashes"));
 441
 442                         return $arr_tags;
 443                 }
 444         }
 445
 446         /**
 447          * @brief Add a hasht sign to a string
 448          *
 449          *  This method is used as callback function
 450          *
 451          * @param string $tag The pure tag name
 452          * @param int $k Counter for internal use
 453          */
 454         private static function arrAddHashes(&$tag, $k) {
 455                 $tag = "#" . $tag;
 456         }
 457
 458         /**
 459          * @brief Add a scheme to an url
 460          *
 461          * The src attribute of some html elements (e.g. images)
 462          * can miss the scheme so we need to add the correct
 463          * scheme
 464          *
 465          * @param string $url The url which possibly does have
 466          *    a missing scheme (a link to an image)
 467          * @param string $scheme The url with a correct scheme
 468          *    (e.g. the url from the webpage which does contain the image)
 469          *
 470          * @return string The url with a scheme
 471          */
 472         private static function completeUrl($url, $scheme) {
 473                 $urlarr = parse_url($url);
 474
 475                 // If the url does allready have an scheme
 476                 // we can stop the process here
 477                 if (isset($urlarr["scheme"])) {
 478                         return($url);
 479                 }
 480
 481                 $schemearr = parse_url($scheme);
 482
 483                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 484
 485                 if (@$schemearr["port"] != "") {
 486                         $complete .= ":".$schemearr["port"];
 487                 }
 488
 489                 if (strpos($urlarr["path"],"/") !== 0) {
 490                         $complete .= "/";
 491                 }
 492
 493                 $complete .= $urlarr["path"];
 494
 495                 if (@$urlarr["query"] != "") {
 496                         $complete .= "?".$urlarr["query"];
 497                 }
 498
 499                 if (@$urlarr["fragment"] != "") {
 500                         $complete .= "#".$urlarr["fragment"];
 501                 }
 502
 503                 return($complete);
 504         }
 505 }