]> git.mxchange.org Git - friendica.git/blob - include/ParseUrl.php
Merge remote-tracking branch 'refs/remotes/friendica/develop' into develop
[friendica.git] / include / 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 require_once("include/network.php");
13 require_once("include/Photo.php");
14 require_once("include/oembed.php");
15 require_once("include/xml.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 type $url The url of the page which should be scraped
26          * @param type $no_guessing If true the parse doens't search for
27          *    preview pictures
28          * @param type $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                 $r = q("SELECT * FROM `parsed_url` WHERE `url` = '%s' AND `guessing` = %d AND `oembed` = %d",
51                         dbesc(normalise_link($url)), intval(!$no_guessing), intval($do_oembed));
52
53                 if ($r) {
54                         $data = $r[0]["content"];
55                 }
56
57                 if (!is_null($data)) {
58                         $data = unserialize($data);
59                         return $data;
60                 }
61
62                 $data = self::getSiteinfo($url, $no_guessing, $do_oembed);
63
64                 q("INSERT INTO `parsed_url` (`url`, `guessing`, `oembed`, `content`, `created`) VALUES ('%s', %d, %d, '%s', '%s')
65                          ON DUPLICATE KEY UPDATE `content` = '%s', `created` = '%s'",
66                         dbesc(normalise_link($url)), intval(!$no_guessing), intval($do_oembed),
67                         dbesc(serialize($data)), dbesc(datetime_convert()),
68                         dbesc(serialize($data)), dbesc(datetime_convert()));
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 type $url The url of the page which should be scraped
83          * @param type $no_guessing If true the parse doens't search for
84          *    preview pictures
85          * @param type $do_oembed The false option is used by the function fetch_oembed()
86          *    to avoid endless loops
87          * @param type $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 = array();
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                 $url = strip_tracking_query_params($url);
134
135                 $siteinfo["url"] = $url;
136                 $siteinfo["type"] = "link";
137
138                 $check_cert = Config::get("system", "verifyssl");
139
140                 $stamp1 = microtime(true);
141
142                 $ch = curl_init();
143                 curl_setopt($ch, CURLOPT_URL, $url);
144                 curl_setopt($ch, CURLOPT_HEADER, 1);
145                 curl_setopt($ch, CURLOPT_TIMEOUT, 10);
146                 curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
147                 curl_setopt($ch, CURLOPT_USERAGENT, $a->get_useragent());
148                 curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, (($check_cert) ? true : false));
149                 if ($check_cert) {
150                         @curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 2);
151                 }
152
153                 $header = curl_exec($ch);
154                 $curl_info = @curl_getinfo($ch);
155                 curl_close($ch);
156
157                 $a->save_timestamp($stamp1, "network");
158
159                 if ((($curl_info["http_code"] == "301") || ($curl_info["http_code"] == "302") || ($curl_info["http_code"] == "303") || ($curl_info["http_code"] == "307"))
160                         && (($curl_info["redirect_url"] != "") || ($curl_info["location"] != ""))) {
161                         if ($curl_info["redirect_url"] != "") {
162                                 $siteinfo = self::getSiteinfo($curl_info["redirect_url"], $no_guessing, $do_oembed, ++$count);
163                         } else {
164                                 $siteinfo = self::getSiteinfo($curl_info["location"], $no_guessing, $do_oembed, ++$count);
165                         }
166                         return($siteinfo);
167                 }
168
169                 // If the file is too large then exit
170                 if ($curl_info["download_content_length"] > 1000000) {
171                         return($siteinfo);
172                 }
173
174                 // If it isn't a HTML file then exit
175                 if (($curl_info["content_type"] != "") && !strstr(strtolower($curl_info["content_type"]), "html")) {
176                         return($siteinfo);
177                 }
178
179                 if ($do_oembed) {
180
181                         $oembed_data = oembed_fetch_url($url);
182
183                         if (!in_array($oembed_data->type, array("error", "rich"))) {
184                                 $siteinfo["type"] = $oembed_data->type;
185                         }
186
187                         if (($oembed_data->type == "link") && ($siteinfo["type"] != "photo")) {
188                                 if (isset($oembed_data->title)) {
189                                         $siteinfo["title"] = $oembed_data->title;
190                                 }
191                                 if (isset($oembed_data->description)) {
192                                         $siteinfo["text"] = trim($oembed_data->description);
193                                 }
194                                 if (isset($oembed_data->thumbnail_url)) {
195                                         $siteinfo["image"] = $oembed_data->thumbnail_url;
196                                 }
197                         }
198                 }
199
200                 // Fetch the first mentioned charset. Can be in body or header
201                 $charset = "";
202                 if (preg_match('/charset=(.*?)['."'".'"\s\n]/', $header, $matches)) {
203                         $charset = trim(trim(trim(array_pop($matches)), ';,'));
204                 }
205
206                 if ($charset == "") {
207                         $charset = "utf-8";
208                 }
209
210                 $pos = strpos($header, "\r\n\r\n");
211
212                 if ($pos) {
213                         $body = trim(substr($header, $pos));
214                 } else {
215                         $body = $header;
216                 }
217
218                 if (($charset != "") && (strtoupper($charset) != "UTF-8")) {
219                         logger("parseurl_getsiteinfo: detected charset ".$charset, LOGGER_DEBUG);
220                         //$body = mb_convert_encoding($body, "UTF-8", $charset);
221                         $body = iconv($charset, "UTF-8//TRANSLIT", $body);
222                 }
223
224                 $body = mb_convert_encoding($body, 'HTML-ENTITIES', "UTF-8");
225
226                 $doc = new \DOMDocument();
227                 @$doc->loadHTML($body);
228
229                 \xml::deleteNode($doc, "style");
230                 \xml::deleteNode($doc, "script");
231                 \xml::deleteNode($doc, "option");
232                 \xml::deleteNode($doc, "h1");
233                 \xml::deleteNode($doc, "h2");
234                 \xml::deleteNode($doc, "h3");
235                 \xml::deleteNode($doc, "h4");
236                 \xml::deleteNode($doc, "h5");
237                 \xml::deleteNode($doc, "h6");
238                 \xml::deleteNode($doc, "ol");
239                 \xml::deleteNode($doc, "ul");
240
241                 $xpath = new \DomXPath($doc);
242
243                 $list = $xpath->query("//meta[@content]");
244                 foreach ($list as $node) {
245                         $attr = array();
246                         if ($node->attributes->length) {
247                                 foreach ($node->attributes as $attribute) {
248                                         $attr[$attribute->name] = $attribute->value;
249                                 }
250                         }
251
252                         if (@$attr["http-equiv"] == "refresh") {
253                                 $path = $attr["content"];
254                                 $pathinfo = explode(";", $path);
255                                 $content = "";
256                                 foreach ($pathinfo as $value) {
257                                         if (substr(strtolower($value), 0, 4) == "url=") {
258                                                 $content = substr($value, 4);
259                                         }
260                                 }
261                                 if ($content != "") {
262                                         $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
263                                         return($siteinfo);
264                                 }
265                         }
266                 }
267
268                 $list = $xpath->query("//title");
269                 if ($list->length > 0) {
270                         $siteinfo["title"] = $list->item(0)->nodeValue;
271                 }
272
273                 //$list = $xpath->query("head/meta[@name]");
274                 $list = $xpath->query("//meta[@name]");
275                 foreach ($list as $node) {
276                         $attr = array();
277                         if ($node->attributes->length) {
278                                 foreach ($node->attributes as $attribute) {
279                                         $attr[$attribute->name] = $attribute->value;
280                                 }
281                         }
282
283                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
284
285                         if ($attr["content"] != "") {
286                                 switch (strtolower($attr["name"])) {
287                                         case "fulltitle":
288                                                 $siteinfo["title"] = $attr["content"];
289                                                 break;
290                                         case "description":
291                                                 $siteinfo["text"] = $attr["content"];
292                                                 break;
293                                         case "thumbnail":
294                                                 $siteinfo["image"] = $attr["content"];
295                                                 break;
296                                         case "twitter:image":
297                                                 $siteinfo["image"] = $attr["content"];
298                                                 break;
299                                         case "twitter:image:src":
300                                                 $siteinfo["image"] = $attr["content"];
301                                                 break;
302                                         case "twitter:card":
303                                                 if (($siteinfo["type"] == "") || ($attr["content"] == "photo")) {
304                                                         $siteinfo["type"] = $attr["content"];
305                                                 }
306                                                 break;
307                                         case "twitter:description":
308                                                 $siteinfo["text"] = $attr["content"];
309                                                 break;
310                                         case "twitter:title":
311                                                 $siteinfo["title"] = $attr["content"];
312                                                 break;
313                                         case "dc.title":
314                                                 $siteinfo["title"] = $attr["content"];
315                                                 break;
316                                         case "dc.description":
317                                                 $siteinfo["text"] = $attr["content"];
318                                                 break;
319                                         case "keywords":
320                                                 $keywords = explode(",", $attr["content"]);
321                                                 break;
322                                         case "news_keywords":
323                                                 $keywords = explode(",", $attr["content"]);
324                                                 break;
325                                 }
326                         }
327                         if ($siteinfo["type"] == "summary") {
328                                 $siteinfo["type"] = "link";
329                         }
330                 }
331
332                 if (isset($keywords)) {
333                         $siteinfo["keywords"] = array();
334                         foreach ($keywords as $keyword) {
335                                 if (!in_array(trim($keyword), $siteinfo["keywords"])) {
336                                         $siteinfo["keywords"][] = trim($keyword);
337                                 }
338                         }
339                 }
340
341                 //$list = $xpath->query("head/meta[@property]");
342                 $list = $xpath->query("//meta[@property]");
343                 foreach ($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                         $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
352
353                         if ($attr["content"] != "") {
354                                 switch (strtolower($attr["property"])) {
355                                         case "og:image":
356                                                 $siteinfo["image"] = $attr["content"];
357                                                 break;
358                                         case "og:title":
359                                                 $siteinfo["title"] = $attr["content"];
360                                                 break;
361                                         case "og:description":
362                                                 $siteinfo["text"] = $attr["content"];
363                                                 break;
364                                 }
365                         }
366                 }
367
368                 if ((@$siteinfo["image"] == "") && !$no_guessing) {
369                         $list = $xpath->query("//img[@src]");
370                         foreach ($list as $node) {
371                                 $attr = array();
372                                 if ($node->attributes->length) {
373                                         foreach ($node->attributes as $attribute) {
374                                                 $attr[$attribute->name] = $attribute->value;
375                                         }
376                                 }
377
378                                 $src = self::completeUrl($attr["src"], $url);
379                                 $photodata = get_photo_info($src);
380
381                                 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
382                                         if ($photodata[0] > 300) {
383                                                 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
384                                                 $photodata[0] = 300;
385                                         }
386                                         if ($photodata[1] > 300) {
387                                                 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
388                                                 $photodata[1] = 300;
389                                         }
390                                         $siteinfo["images"][] = array("src" => $src,
391                                                                         "width" => $photodata[0],
392                                                                         "height" => $photodata[1]);
393                                 }
394
395                                 }
396                 } elseif ($siteinfo["image"] != "") {
397                         $src = self::completeUrl($siteinfo["image"], $url);
398
399                         unset($siteinfo["image"]);
400
401                         $photodata = get_photo_info($src);
402
403                         if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
404                                 $siteinfo["images"][] = array("src" => $src,
405                                                                 "width" => $photodata[0],
406                                                                 "height" => $photodata[1]);
407                         }
408                 }
409
410                 if ((@$siteinfo["text"] == "") && (@$siteinfo["title"] != "") && !$no_guessing) {
411                         $text = "";
412
413                         $list = $xpath->query("//div[@class='article']");
414                         foreach ($list as $node) {
415                                 if (strlen($node->nodeValue) > 40) {
416                                         $text .= " ".trim($node->nodeValue);
417                                 }
418                         }
419
420                         if ($text == "") {
421                                 $list = $xpath->query("//div[@class='content']");
422                                 foreach ($list as $node) {
423                                         if (strlen($node->nodeValue) > 40) {
424                                                 $text .= " ".trim($node->nodeValue);
425                                         }
426                                 }
427                         }
428
429                         // If none text was found then take the paragraph content
430                         if ($text == "") {
431                                 $list = $xpath->query("//p");
432                                 foreach ($list as $node) {
433                                         if (strlen($node->nodeValue) > 40) {
434                                                 $text .= " ".trim($node->nodeValue);
435                                         }
436                                 }
437                         }
438
439                         if ($text != "") {
440                                 $text = trim(str_replace(array("\n", "\r"), array(" ", " "), $text));
441
442                                 while (strpos($text, "  ")) {
443                                         $text = trim(str_replace("  ", " ", $text));
444                                 }
445
446                                 $siteinfo["text"] = trim(html_entity_decode(substr($text, 0, 350), ENT_QUOTES, "UTF-8").'...');
447                         }
448                 }
449
450                 logger("parseurl_getsiteinfo: Siteinfo for ".$url." ".print_r($siteinfo, true), LOGGER_DEBUG);
451
452                 call_hooks("getsiteinfo", $siteinfo);
453
454                 return($siteinfo);
455         }
456
457         /**
458          * @brief Convert tags from CSV to an array
459          *
460          * @param string $string Tags
461          * @return array with formatted Hashtags
462          */
463         public static function convertTagsToArray($string) {
464                 $arr_tags = str_getcsv($string);
465                 if (count($arr_tags)) {
466                         // add the # sign to every tag
467                         array_walk($arr_tags, array("self", "arrAddHashes"));
468
469                         return $arr_tags;
470                 }
471         }
472
473         /**
474          * @brief Add a hasht sign to a string
475          *
476          *  This method is used as callback function
477          *
478          * @param string $tag The pure tag name
479          * @param int $k Counter for internal use
480          */
481         private static function arrAddHashes(&$tag, $k) {
482                 $tag = "#" . $tag;
483         }
484
485         /**
486          * @brief Add a scheme to an url
487          *
488          * The src attribute of some html elements (e.g. images)
489          * can miss the scheme so we need to add the correct
490          * scheme
491          *
492          * @param string $url The url which possibly does have
493          *    a missing scheme (a link to an image)
494          * @param string $scheme The url with a correct scheme
495          *    (e.g. the url from the webpage which does contain the image)
496          *
497          * @return string The url with a scheme
498          */
499         private static function completeUrl($url, $scheme) {
500                 $urlarr = parse_url($url);
501
502                 // If the url does allready have an scheme
503                 // we can stop the process here
504                 if (isset($urlarr["scheme"])) {
505                         return($url);
506                 }
507
508                 $schemearr = parse_url($scheme);
509
510                 $complete = $schemearr["scheme"]."://".$schemearr["host"];
511
512                 if (@$schemearr["port"] != "") {
513                         $complete .= ":".$schemearr["port"];
514                 }
515
516                 if (strpos($urlarr["path"],"/") !== 0) {
517                         $complete .= "/";
518                 }
519
520                 $complete .= $urlarr["path"];
521
522                 if (@$urlarr["query"] != "") {
523                         $complete .= "?".$urlarr["query"];
524                 }
525
526                 if (@$urlarr["fragment"] != "") {
527                         $complete .= "#".$urlarr["fragment"];
528                 }
529
530                 return($complete);
531         }
532 }