src/Util/ParseUrl.php

   1 <?php
   2 /**
   3  * @file src/Util/ParseUrl.php
   4  * @brief Get informations about a given URL
   5  */
   6 namespace Friendica\Util;
   7
   8 use Friendica\Content\OEmbed;
   9 use Friendica\Core\Addon;
  10 use Friendica\Object\Image;
  11 use Friendica\Util\Network;
  12 use Friendica\Util\XML;
  13
  14 use dba;
  15 use DOMXPath;
  16 use DOMDocument;
  17
  18 require_once 'include/dba.php';
  19
  20 /**
  21  * @brief Class with methods for extracting certain content from an url
  22  */
  23 class ParseUrl
  24 {
  25         /**
  26          * @brief Search for chached embeddable data of an url otherwise fetch it
  27          *
  28          * @param string $url         The url of the page which should be scraped
  29          * @param bool $no_guessing If true the parse doens't search for
  30          *                          preview pictures
  31          * @param bool $do_oembed   The false option is used by the function fetch_oembed()
  32          *                          to avoid endless loops
  33          *
  34          * @return array which contains needed data for embedding
  35          *    string 'url' => The url of the parsed page
  36          *    string 'type' => Content type
  37          *    string 'title' => The title of the content
  38          *    string 'text' => The description for the content
  39          *    string 'image' => A preview image of the content (only available
  40          *                if $no_geuessing = false
  41          *    array'images' = Array of preview pictures
  42          *    string 'keywords' => The tags which belong to the content
  43          *
  44          * @see ParseUrl::getSiteinfo() for more information about scraping
  45          * embeddable content
  46          */
  47         public static function getSiteinfoCached($url, $no_guessing = false, $do_oembed = true)
  48         {
  49                 if ($url == "") {
  50                         return false;
  51                 }
  52
  53                 $r = q(
  54                         "SELECT * FROM `parsed_url` WHERE `url` = '%s' AND `guessing` = %d AND `oembed` = %d",
  55                         dbesc(normalise_link($url)),
  56                         intval(!$no_guessing),
  57                         intval($do_oembed)
  58                 );
  59
  60                 if ($r) {
  61                         $data = $r[0]["content"];
  62                 }
  63
  64                 if (!is_null($data)) {
  65                         $data = unserialize($data);
  66                         return $data;
  67                 }
  68
  69                 $data = self::getSiteinfo($url, $no_guessing, $do_oembed);
  70
  71                 dba::insert(
  72                         'parsed_url',
  73                         [
  74                                 'url' => normalise_link($url), 'guessing' => !$no_guessing,
  75                                 'oembed' => $do_oembed, 'content' => serialize($data),
  76                                 'created' => Temporal::utcNow()
  77                         ],
  78                         true
  79                 );
  80
  81                 return $data;
  82         }
  83         /**
  84          * @brief Parse a page for embeddable content information
  85          *
  86          * This method parses to url for meta data which can be used to embed
  87          * the content. If available it prioritizes Open Graph meta tags.
  88          * If this is not available it uses the twitter cards meta tags.
  89          * As fallback it uses standard html elements with meta informations
  90          * like \<title\>Awesome Title\</title\> or
  91          * \<meta name="description" content="An awesome description"\>
  92          *
  93          * @param string $url         The url of the page which should be scraped
  94          * @param bool $no_guessing If true the parse doens't search for
  95          *                          preview pictures
  96          * @param bool $do_oembed   The false option is used by the function fetch_oembed()
  97          *                          to avoid endless loops
  98          * @param int $count       Internal counter to avoid endless loops
  99          *
 100          * @return array which contains needed data for embedding
 101          *    string 'url' => The url of the parsed page
 102          *    string 'type' => Content type
 103          *    string 'title' => The title of the content
 104          *    string 'text' => The description for the content
 105          *    string 'image' => A preview image of the content (only available
 106          *                if $no_geuessing = false
 107          *    array'images' = Array of preview pictures
 108          *    string 'keywords' => The tags which belong to the content
 109          *
 110          * @todo https://developers.google.com/+/plugins/snippet/
 111          * @verbatim
 112          * <meta itemprop="name" content="Awesome title">
 113          * <meta itemprop="description" content="An awesome description">
 114          * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
 115          *
 116          * <body itemscope itemtype="http://schema.org/Product">
 117          *   <h1 itemprop="name">Shiny Trinket</h1>
 118          *   <img itemprop="image" src="{image-url}" />
 119          *   <p itemprop="description">Shiny trinkets are shiny.</p>
 120          * </body>
 121          * @endverbatim
 122          */
 123         public static function getSiteinfo($url, $no_guessing = false, $do_oembed = true, $count = 1)
 124         {
 125                 $a = get_app();
 126
 127                 $siteinfo = [];
 128
 129                 // Check if the URL does contain a scheme
 130                 $scheme = parse_url($url, PHP_URL_SCHEME);
 131
 132                 if ($scheme == "") {
 133                         $url = "http://".trim($url, "/");
 134                 }
 135
 136                 if ($count > 10) {
 137                         logger("parseurl_getsiteinfo: Endless loop detected for ".$url, LOGGER_DEBUG);
 138                         return($siteinfo);
 139                 }
 140
 141                 $url = trim($url, "'");
 142                 $url = trim($url, '"');
 143
 144                 $url = Network::stripTrackingQueryParams($url);
 145
 146                 $siteinfo["url"] = $url;
 147                 $siteinfo["type"] = "link";
 148
 149                 $data = Network::curl($url);
 150                 if (!$data['success']) {
 151                         return($siteinfo);
 152                 }
 153
 154                 // If the file is too large then exit
 155                 if ($data["info"]["download_content_length"] > 1000000) {
 156                         return($siteinfo);
 157                 }
 158
 159                 // If it isn't a HTML file then exit
 160                 if (($data["info"]["content_type"] != "") && !strstr(strtolower($data["info"]["content_type"]), "html")) {
 161                         return($siteinfo);
 162                 }
 163
 164                 $header = $data["header"];
 165                 $body = $data["body"];
 166
 167                 if ($do_oembed) {
 168                         $oembed_data = OEmbed::fetchURL($url);
 169
 170                         if (!in_array($oembed_data->type, ["error", "rich", ""])) {
 171                                 $siteinfo["type"] = $oembed_data->type;
 172                         }
 173
 174                         if (($oembed_data->type == "link") && ($siteinfo["type"] != "photo")) {
 175                                 if (isset($oembed_data->title)) {
 176                                         $siteinfo["title"] = trim($oembed_data->title);
 177                                 }
 178                                 if (isset($oembed_data->description)) {
 179                                         $siteinfo["text"] = trim($oembed_data->description);
 180                                 }
 181                                 if (isset($oembed_data->thumbnail_url)) {
 182                                         $siteinfo["image"] = $oembed_data->thumbnail_url;
 183                                 }
 184                         }
 185                 }
 186
 187                 // Fetch the first mentioned charset. Can be in body or header
 188                 $charset = "";
 189                 if (preg_match('/charset=(.*?)['."'".'"\s\n]/', $header, $matches)) {
 190                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 191                 }
 192
 193                 if ($charset == "") {
 194                         $charset = "utf-8";
 195                 }
 196
 197                 if (($charset != "") && (strtoupper($charset) != "UTF-8")) {
 198                         logger("parseurl_getsiteinfo: detected charset ".$charset, LOGGER_DEBUG);
 199                         //$body = mb_convert_encoding($body, "UTF-8", $charset);
 200                         $body = iconv($charset, "UTF-8//TRANSLIT", $body);
 201                 }
 202
 203                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', "UTF-8");
 204
 205                 $doc = new DOMDocument();
 206                 @$doc->loadHTML($body);
 207
 208                 XML::deleteNode($doc, "style");
 209                 XML::deleteNode($doc, "script");
 210                 XML::deleteNode($doc, "option");
 211                 XML::deleteNode($doc, "h1");
 212                 XML::deleteNode($doc, "h2");
 213                 XML::deleteNode($doc, "h3");
 214                 XML::deleteNode($doc, "h4");
 215                 XML::deleteNode($doc, "h5");
 216                 XML::deleteNode($doc, "h6");
 217                 XML::deleteNode($doc, "ol");
 218                 XML::deleteNode($doc, "ul");
 219
 220                 $xpath = new DOMXPath($doc);
 221
 222                 $list = $xpath->query("//meta[@content]");
 223                 foreach ($list as $node) {
 224                         $attr = [];
 225                         if ($node->attributes->length) {
 226                                 foreach ($node->attributes as $attribute) {
 227                                         $attr[$attribute->name] = $attribute->value;
 228                                 }
 229                         }
 230
 231                         if (@$attr["http-equiv"] == "refresh") {
 232                                 $path = $attr["content"];
 233                                 $pathinfo = explode(";", $path);
 234                                 $content = "";
 235                                 foreach ($pathinfo as $value) {
 236                                         if (substr(strtolower($value), 0, 4) == "url=") {
 237                                                 $content = substr($value, 4);
 238                                         }
 239                                 }
 240                                 if ($content != "") {
 241                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 242                                         return($siteinfo);
 243                                 }
 244                         }
 245                 }
 246
 247                 $list = $xpath->query("//title");
 248                 if ($list->length > 0) {
 249                         $siteinfo["title"] = trim($list->item(0)->nodeValue);
 250                 }
 251
 252                 //$list = $xpath->query("head/meta[@name]");
 253                 $list = $xpath->query("//meta[@name]");
 254                 foreach ($list as $node) {
 255                         $attr = [];
 256                         if ($node->attributes->length) {
 257                                 foreach ($node->attributes as $attribute) {
 258                                         $attr[$attribute->name] = $attribute->value;
 259                                 }
 260                         }
 261
 262                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 263
 264                         if ($attr["content"] != "") {
 265                                 switch (strtolower($attr["name"])) {
 266                                         case "fulltitle":
 267                                                 $siteinfo["title"] = trim($attr["content"]);
 268                                                 break;
 269                                         case "description":
 270                                                 $siteinfo["text"] = trim($attr["content"]);
 271                                                 break;
 272                                         case "thumbnail":
 273                                                 $siteinfo["image"] = $attr["content"];
 274                                                 break;
 275                                         case "twitter:image":
 276                                                 $siteinfo["image"] = $attr["content"];
 277                                                 break;
 278                                         case "twitter:image:src":
 279                                                 $siteinfo["image"] = $attr["content"];
 280                                                 break;
 281                                         case "twitter:card":
 282                                                 if (($siteinfo["type"] == "") || ($attr["content"] == "photo")) {
 283                                                         $siteinfo["type"] = $attr["content"];
 284                                                 }
 285                                                 break;
 286                                         case "twitter:description":
 287                                                 $siteinfo["text"] = trim($attr["content"]);
 288                                                 break;
 289                                         case "twitter:title":
 290                                                 $siteinfo["title"] = trim($attr["content"]);
 291                                                 break;
 292                                         case "dc.title":
 293                                                 $siteinfo["title"] = trim($attr["content"]);
 294                                                 break;
 295                                         case "dc.description":
 296                                                 $siteinfo["text"] = trim($attr["content"]);
 297                                                 break;
 298                                         case "keywords":
 299                                                 $keywords = explode(",", $attr["content"]);
 300                                                 break;
 301                                         case "news_keywords":
 302                                                 $keywords = explode(",", $attr["content"]);
 303                                                 break;
 304                                 }
 305                         }
 306                         if ($siteinfo["type"] == "summary") {
 307                                 $siteinfo["type"] = "link";
 308                         }
 309                 }
 310
 311                 if (isset($keywords)) {
 312                         $siteinfo["keywords"] = [];
 313                         foreach ($keywords as $keyword) {
 314                                 if (!in_array(trim($keyword), $siteinfo["keywords"])) {
 315                                         $siteinfo["keywords"][] = trim($keyword);
 316                                 }
 317                         }
 318                 }
 319
 320                 //$list = $xpath->query("head/meta[@property]");
 321                 $list = $xpath->query("//meta[@property]");
 322                 foreach ($list as $node) {
 323                         $attr = [];
 324                         if ($node->attributes->length) {
 325                                 foreach ($node->attributes as $attribute) {
 326                                         $attr[$attribute->name] = $attribute->value;
 327                                 }
 328                         }
 329
 330                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 331
 332                         if ($attr["content"] != "") {
 333                                 switch (strtolower($attr["property"])) {
 334                                         case "og:image":
 335                                                 $siteinfo["image"] = $attr["content"];
 336                                                 break;
 337                                         case "og:title":
 338                                                 $siteinfo["title"] = trim($attr["content"]);
 339                                                 break;
 340                                         case "og:description":
 341                                                 $siteinfo["text"] = trim($attr["content"]);
 342                                                 break;
 343                                 }
 344                         }
 345                 }
 346
 347                 if ((@$siteinfo["image"] == "") && !$no_guessing) {
 348                         $list = $xpath->query("//img[@src]");
 349                         foreach ($list as $node) {
 350                                 $attr = [];
 351                                 if ($node->attributes->length) {
 352                                         foreach ($node->attributes as $attribute) {
 353                                                 $attr[$attribute->name] = $attribute->value;
 354                                         }
 355                                 }
 356
 357                                 $src = self::completeUrl($attr["src"], $url);
 358                                 $photodata = Image::getInfoFromURL($src);
 359
 360                                 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
 361                                         if ($photodata[0] > 300) {
 362                                                 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
 363                                                 $photodata[0] = 300;
 364                                         }
 365                                         if ($photodata[1] > 300) {
 366                                                 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
 367                                                 $photodata[1] = 300;
 368                                         }
 369                                         $siteinfo["images"][] = ["src" => $src,
 370                                                                         "width" => $photodata[0],
 371                                                                         "height" => $photodata[1]];
 372                                 }
 373                         }
 374                 } elseif ($siteinfo["image"] != "") {
 375                         $src = self::completeUrl($siteinfo["image"], $url);
 376
 377                         unset($siteinfo["image"]);
 378
 379                         $photodata = Image::getInfoFromURL($src);
 380
 381                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 382                                 $siteinfo["images"][] = ["src" => $src,
 383                                                                 "width" => $photodata[0],
 384                                                                 "height" => $photodata[1]];
 385                         }
 386                 }
 387
 388                 if ((@$siteinfo["text"] == "") && (@$siteinfo["title"] != "") && !$no_guessing) {
 389                         $text = "";
 390
 391                         $list = $xpath->query("//div[@class='article']");
 392                         foreach ($list as $node) {
 393                                 if (strlen($node->nodeValue) > 40) {
 394                                         $text .= " ".trim($node->nodeValue);
 395                                 }
 396                         }
 397
 398                         if ($text == "") {
 399                                 $list = $xpath->query("//div[@class='content']");
 400                                 foreach ($list as $node) {
 401                                         if (strlen($node->nodeValue) > 40) {
 402                                                 $text .= " ".trim($node->nodeValue);
 403                                         }
 404                                 }
 405                         }
 406
 407                         // If none text was found then take the paragraph content
 408                         if ($text == "") {
 409                                 $list = $xpath->query("//p");
 410                                 foreach ($list as $node) {
 411                                         if (strlen($node->nodeValue) > 40) {
 412                                                 $text .= " ".trim($node->nodeValue);
 413                                         }
 414                                 }
 415                         }
 416
 417                         if ($text != "") {
 418                                 $text = trim(str_replace(["\n", "\r"], [" ", " "], $text));
 419
 420                                 while (strpos($text, "  ")) {
 421                                         $text = trim(str_replace("  ", " ", $text));
 422                                 }
 423
 424                                 $siteinfo["text"] = trim(html_entity_decode(substr($text, 0, 350), ENT_QUOTES, "UTF-8").'...');
 425                         }
 426                 }
 427
 428                 logger("parseurl_getsiteinfo: Siteinfo for ".$url." ".print_r($siteinfo, true), LOGGER_DEBUG);
 429
 430                 Addon::callHooks("getsiteinfo", $siteinfo);
 431
 432                 return($siteinfo);
 433         }
 434
 435         /**
 436          * @brief Convert tags from CSV to an array
 437          *
 438          * @param string $string Tags
 439          * @return array with formatted Hashtags
 440          */
 441         public static function convertTagsToArray($string)
 442         {
 443                 $arr_tags = str_getcsv($string);
 444                 if (count($arr_tags)) {
 445                         // add the # sign to every tag
 446                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 447
 448                         return $arr_tags;
 449                 }
 450         }
 451
 452         /**
 453          * @brief Add a hasht sign to a string
 454          *
 455          *  This method is used as callback function
 456          *
 457          * @param string $tag The pure tag name
 458          * @param int    $k   Counter for internal use
 459          * @return void
 460          */
 461         private static function arrAddHashes(&$tag, $k)
 462         {
 463                 $tag = "#" . $tag;
 464         }
 465
 466         /**
 467          * @brief Add a scheme to an url
 468          *
 469          * The src attribute of some html elements (e.g. images)
 470          * can miss the scheme so we need to add the correct
 471          * scheme
 472          *
 473          * @param string $url    The url which possibly does have
 474          *                       a missing scheme (a link to an image)
 475          * @param string $scheme The url with a correct scheme
 476          *                       (e.g. the url from the webpage which does contain the image)
 477          *
 478          * @return string The url with a scheme
 479          */
 480         private static function completeUrl($url, $scheme)
 481         {
 482                 $urlarr = parse_url($url);
 483
 484                 // If the url does allready have an scheme
 485                 // we can stop the process here
 486                 if (isset($urlarr["scheme"])) {
 487                         return($url);
 488                 }
 489
 490                 $schemearr = parse_url($scheme);
 491
 492                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 493
 494                 if (@$schemearr["port"] != "") {
 495                         $complete .= ":".$schemearr["port"];
 496                 }
 497
 498                 if (strpos($urlarr["path"], "/") !== 0) {
 499                         $complete .= "/";
 500                 }
 501
 502                 $complete .= $urlarr["path"];
 503
 504                 if (@$urlarr["query"] != "") {
 505                         $complete .= "?".$urlarr["query"];
 506                 }
 507
 508                 if (@$urlarr["fragment"] != "") {
 509                         $complete .= "#".$urlarr["fragment"];
 510                 }
 511
 512                 return($complete);
 513         }
 514 }