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                 $parsed_url = dba::selectFirst('parsed_url', ['content'],
  54                         ['url' => normalise_link($url), 'guessing' => !$no_guessing, 'oembed' => $do_oembed]
  55                 );
  56                 if (!empty($parsed_url['content'])) {
  57                         $data = unserialize($parsed_url['content']);
  58                         return $data;
  59                 }
  60
  61                 $data = self::getSiteinfo($url, $no_guessing, $do_oembed);
  62
  63                 dba::insert(
  64                         'parsed_url',
  65                         [
  66                                 'url' => normalise_link($url), 'guessing' => !$no_guessing,
  67                                 'oembed' => $do_oembed, 'content' => serialize($data),
  68                                 'created' => DateTimeFormat::utcNow()
  69                         ],
  70                         true
  71                 );
  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 string $url         The url of the page which should be scraped
  86          * @param bool $no_guessing If true the parse doens't search for
  87          *                          preview pictures
  88          * @param bool $do_oembed   The false option is used by the function fetch_oembed()
  89          *                          to avoid endless loops
  90          * @param int $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 = [];
 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                 $url = Network::stripTrackingQueryParams($url);
 137
 138                 $siteinfo["url"] = $url;
 139                 $siteinfo["type"] = "link";
 140
 141                 $data = Network::curl($url);
 142                 if (!$data['success']) {
 143                         return($siteinfo);
 144                 }
 145
 146                 // If the file is too large then exit
 147                 if ($data["info"]["download_content_length"] > 1000000) {
 148                         return($siteinfo);
 149                 }
 150
 151                 // If it isn't a HTML file then exit
 152                 if (($data["info"]["content_type"] != "") && !strstr(strtolower($data["info"]["content_type"]), "html")) {
 153                         return($siteinfo);
 154                 }
 155
 156                 $header = $data["header"];
 157                 $body = $data["body"];
 158
 159                 if ($do_oembed) {
 160                         $oembed_data = OEmbed::fetchURL($url);
 161
 162                         if (!in_array($oembed_data->type, ["error", "rich", ""])) {
 163                                 $siteinfo["type"] = $oembed_data->type;
 164                         }
 165
 166                         if (($oembed_data->type == "link") && ($siteinfo["type"] != "photo")) {
 167                                 if (isset($oembed_data->title)) {
 168                                         $siteinfo["title"] = trim($oembed_data->title);
 169                                 }
 170                                 if (isset($oembed_data->description)) {
 171                                         $siteinfo["text"] = trim($oembed_data->description);
 172                                 }
 173                                 if (isset($oembed_data->thumbnail_url)) {
 174                                         $siteinfo["image"] = $oembed_data->thumbnail_url;
 175                                 }
 176                         }
 177                 }
 178
 179                 // Fetch the first mentioned charset. Can be in body or header
 180                 $charset = "";
 181                 if (preg_match('/charset=(.*?)['."'".'"\s\n]/', $header, $matches)) {
 182                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
 183                 }
 184
 185                 if ($charset == "") {
 186                         $charset = "utf-8";
 187                 }
 188
 189                 if (($charset != "") && (strtoupper($charset) != "UTF-8")) {
 190                         logger("parseurl_getsiteinfo: detected charset ".$charset, LOGGER_DEBUG);
 191                         //$body = mb_convert_encoding($body, "UTF-8", $charset);
 192                         $body = iconv($charset, "UTF-8//TRANSLIT", $body);
 193                 }
 194
 195                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', "UTF-8");
 196
 197                 $doc = new DOMDocument();
 198                 @$doc->loadHTML($body);
 199
 200                 XML::deleteNode($doc, "style");
 201                 XML::deleteNode($doc, "script");
 202                 XML::deleteNode($doc, "option");
 203                 XML::deleteNode($doc, "h1");
 204                 XML::deleteNode($doc, "h2");
 205                 XML::deleteNode($doc, "h3");
 206                 XML::deleteNode($doc, "h4");
 207                 XML::deleteNode($doc, "h5");
 208                 XML::deleteNode($doc, "h6");
 209                 XML::deleteNode($doc, "ol");
 210                 XML::deleteNode($doc, "ul");
 211
 212                 $xpath = new DOMXPath($doc);
 213
 214                 $list = $xpath->query("//meta[@content]");
 215                 foreach ($list as $node) {
 216                         $attr = [];
 217                         if ($node->attributes->length) {
 218                                 foreach ($node->attributes as $attribute) {
 219                                         $attr[$attribute->name] = $attribute->value;
 220                                 }
 221                         }
 222
 223                         if (@$attr["http-equiv"] == "refresh") {
 224                                 $path = $attr["content"];
 225                                 $pathinfo = explode(";", $path);
 226                                 $content = "";
 227                                 foreach ($pathinfo as $value) {
 228                                         if (substr(strtolower($value), 0, 4) == "url=") {
 229                                                 $content = substr($value, 4);
 230                                         }
 231                                 }
 232                                 if ($content != "") {
 233                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
 234                                         return($siteinfo);
 235                                 }
 236                         }
 237                 }
 238
 239                 $list = $xpath->query("//title");
 240                 if ($list->length > 0) {
 241                         $siteinfo["title"] = trim($list->item(0)->nodeValue);
 242                 }
 243
 244                 //$list = $xpath->query("head/meta[@name]");
 245                 $list = $xpath->query("//meta[@name]");
 246                 foreach ($list as $node) {
 247                         $attr = [];
 248                         if ($node->attributes->length) {
 249                                 foreach ($node->attributes as $attribute) {
 250                                         $attr[$attribute->name] = $attribute->value;
 251                                 }
 252                         }
 253
 254                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 255
 256                         if ($attr["content"] != "") {
 257                                 switch (strtolower($attr["name"])) {
 258                                         case "fulltitle":
 259                                                 $siteinfo["title"] = trim($attr["content"]);
 260                                                 break;
 261                                         case "description":
 262                                                 $siteinfo["text"] = trim($attr["content"]);
 263                                                 break;
 264                                         case "thumbnail":
 265                                                 $siteinfo["image"] = $attr["content"];
 266                                                 break;
 267                                         case "twitter:image":
 268                                                 $siteinfo["image"] = $attr["content"];
 269                                                 break;
 270                                         case "twitter:image:src":
 271                                                 $siteinfo["image"] = $attr["content"];
 272                                                 break;
 273                                         case "twitter:card":
 274                                                 if (($siteinfo["type"] == "") || ($attr["content"] == "photo")) {
 275                                                         $siteinfo["type"] = $attr["content"];
 276                                                 }
 277                                                 break;
 278                                         case "twitter:description":
 279                                                 $siteinfo["text"] = trim($attr["content"]);
 280                                                 break;
 281                                         case "twitter:title":
 282                                                 $siteinfo["title"] = trim($attr["content"]);
 283                                                 break;
 284                                         case "dc.title":
 285                                                 $siteinfo["title"] = trim($attr["content"]);
 286                                                 break;
 287                                         case "dc.description":
 288                                                 $siteinfo["text"] = trim($attr["content"]);
 289                                                 break;
 290                                         case "keywords":
 291                                                 $keywords = explode(",", $attr["content"]);
 292                                                 break;
 293                                         case "news_keywords":
 294                                                 $keywords = explode(",", $attr["content"]);
 295                                                 break;
 296                                 }
 297                         }
 298                         if ($siteinfo["type"] == "summary") {
 299                                 $siteinfo["type"] = "link";
 300                         }
 301                 }
 302
 303                 if (isset($keywords)) {
 304                         $siteinfo["keywords"] = [];
 305                         foreach ($keywords as $keyword) {
 306                                 if (!in_array(trim($keyword), $siteinfo["keywords"])) {
 307                                         $siteinfo["keywords"][] = trim($keyword);
 308                                 }
 309                         }
 310                 }
 311
 312                 //$list = $xpath->query("head/meta[@property]");
 313                 $list = $xpath->query("//meta[@property]");
 314                 foreach ($list as $node) {
 315                         $attr = [];
 316                         if ($node->attributes->length) {
 317                                 foreach ($node->attributes as $attribute) {
 318                                         $attr[$attribute->name] = $attribute->value;
 319                                 }
 320                         }
 321
 322                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
 323
 324                         if ($attr["content"] != "") {
 325                                 switch (strtolower($attr["property"])) {
 326                                         case "og:image":
 327                                                 $siteinfo["image"] = $attr["content"];
 328                                                 break;
 329                                         case "og:title":
 330                                                 $siteinfo["title"] = trim($attr["content"]);
 331                                                 break;
 332                                         case "og:description":
 333                                                 $siteinfo["text"] = trim($attr["content"]);
 334                                                 break;
 335                                 }
 336                         }
 337                 }
 338
 339                 if ((@$siteinfo["image"] == "") && !$no_guessing) {
 340                         $list = $xpath->query("//img[@src]");
 341                         foreach ($list as $node) {
 342                                 $attr = [];
 343                                 if ($node->attributes->length) {
 344                                         foreach ($node->attributes as $attribute) {
 345                                                 $attr[$attribute->name] = $attribute->value;
 346                                         }
 347                                 }
 348
 349                                 $src = self::completeUrl($attr["src"], $url);
 350                                 $photodata = Image::getInfoFromURL($src);
 351
 352                                 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
 353                                         if ($photodata[0] > 300) {
 354                                                 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
 355                                                 $photodata[0] = 300;
 356                                         }
 357                                         if ($photodata[1] > 300) {
 358                                                 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
 359                                                 $photodata[1] = 300;
 360                                         }
 361                                         $siteinfo["images"][] = ["src" => $src,
 362                                                                         "width" => $photodata[0],
 363                                                                         "height" => $photodata[1]];
 364                                 }
 365                         }
 366                 } elseif ($siteinfo["image"] != "") {
 367                         $src = self::completeUrl($siteinfo["image"], $url);
 368
 369                         unset($siteinfo["image"]);
 370
 371                         $photodata = Image::getInfoFromURL($src);
 372
 373                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
 374                                 $siteinfo["images"][] = ["src" => $src,
 375                                                                 "width" => $photodata[0],
 376                                                                 "height" => $photodata[1]];
 377                         }
 378                 }
 379
 380                 if ((@$siteinfo["text"] == "") && (@$siteinfo["title"] != "") && !$no_guessing) {
 381                         $text = "";
 382
 383                         $list = $xpath->query("//div[@class='article']");
 384                         foreach ($list as $node) {
 385                                 if (strlen($node->nodeValue) > 40) {
 386                                         $text .= " ".trim($node->nodeValue);
 387                                 }
 388                         }
 389
 390                         if ($text == "") {
 391                                 $list = $xpath->query("//div[@class='content']");
 392                                 foreach ($list as $node) {
 393                                         if (strlen($node->nodeValue) > 40) {
 394                                                 $text .= " ".trim($node->nodeValue);
 395                                         }
 396                                 }
 397                         }
 398
 399                         // If none text was found then take the paragraph content
 400                         if ($text == "") {
 401                                 $list = $xpath->query("//p");
 402                                 foreach ($list as $node) {
 403                                         if (strlen($node->nodeValue) > 40) {
 404                                                 $text .= " ".trim($node->nodeValue);
 405                                         }
 406                                 }
 407                         }
 408
 409                         if ($text != "") {
 410                                 $text = trim(str_replace(["\n", "\r"], [" ", " "], $text));
 411
 412                                 while (strpos($text, "  ")) {
 413                                         $text = trim(str_replace("  ", " ", $text));
 414                                 }
 415
 416                                 $siteinfo["text"] = trim(html_entity_decode(substr($text, 0, 350), ENT_QUOTES, "UTF-8").'...');
 417                         }
 418                 }
 419
 420                 logger("parseurl_getsiteinfo: Siteinfo for ".$url." ".print_r($siteinfo, true), LOGGER_DEBUG);
 421
 422                 Addon::callHooks("getsiteinfo", $siteinfo);
 423
 424                 return($siteinfo);
 425         }
 426
 427         /**
 428          * @brief Convert tags from CSV to an array
 429          *
 430          * @param string $string Tags
 431          * @return array with formatted Hashtags
 432          */
 433         public static function convertTagsToArray($string)
 434         {
 435                 $arr_tags = str_getcsv($string);
 436                 if (count($arr_tags)) {
 437                         // add the # sign to every tag
 438                         array_walk($arr_tags, ["self", "arrAddHashes"]);
 439
 440                         return $arr_tags;
 441                 }
 442         }
 443
 444         /**
 445          * @brief Add a hasht sign to a string
 446          *
 447          *  This method is used as callback function
 448          *
 449          * @param string $tag The pure tag name
 450          * @param int    $k   Counter for internal use
 451          * @return void
 452          */
 453         private static function arrAddHashes(&$tag, $k)
 454         {
 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         {
 474                 $urlarr = parse_url($url);
 475
 476                 // If the url does allready have an scheme
 477                 // we can stop the process here
 478                 if (isset($urlarr["scheme"])) {
 479                         return($url);
 480                 }
 481
 482                 $schemearr = parse_url($scheme);
 483
 484                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
 485
 486                 if (@$schemearr["port"] != "") {
 487                         $complete .= ":".$schemearr["port"];
 488                 }
 489
 490                 if (strpos($urlarr["path"], "/") !== 0) {
 491                         $complete .= "/";
 492                 }
 493
 494                 $complete .= $urlarr["path"];
 495
 496                 if (@$urlarr["query"] != "") {
 497                         $complete .= "?".$urlarr["query"];
 498                 }
 499
 500                 if (@$urlarr["fragment"] != "") {
 501                         $complete .= "#".$urlarr["fragment"];
 502                 }
 503
 504                 return($complete);
 505         }
 506 }