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