3 * @file src/Util/ParseUrl.php
4 * @brief Get informations about a given URL
6 namespace Friendica\Util;
10 use Friendica\Content\OEmbed;
11 use Friendica\Core\Addon;
12 use Friendica\Database\DBA;
13 use Friendica\Object\Image;
15 require_once 'include/dba.php';
18 * @brief Class with methods for extracting certain content from an url
23 * @brief Search for chached embeddable data of an url otherwise fetch it
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
28 * @param bool $do_oembed The false option is used by the function fetch_oembed()
29 * to avoid endless loops
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 $no_geuessing = false
38 * array'images' = Array of preview pictures
39 * string 'keywords' => The tags which belong to the content
41 * @see ParseUrl::getSiteinfo() for more information about scraping
44 public static function getSiteinfoCached($url, $no_guessing = false, $do_oembed = true)
50 $parsed_url = DBA::selectFirst('parsed_url', ['content'],
51 ['url' => normalise_link($url), 'guessing' => !$no_guessing, 'oembed' => $do_oembed]
53 if (!empty($parsed_url['content'])) {
54 $data = unserialize($parsed_url['content']);
58 $data = self::getSiteinfo($url, $no_guessing, $do_oembed);
63 'url' => normalise_link($url), 'guessing' => !$no_guessing,
64 'oembed' => $do_oembed, 'content' => serialize($data),
65 'created' => DateTimeFormat::utcNow()
73 * @brief Parse a page for embeddable content information
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"\>
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
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
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 $no_geuessing = false
96 * array'images' = Array of preview pictures
97 * string 'keywords' => The tags which belong to the content
99 * @todo https://developers.google.com/+/plugins/snippet/
101 * <meta itemprop="name" content="Awesome title">
102 * <meta itemprop="description" content="An awesome description">
103 * <meta itemprop="image" content="http://maple.libertreeproject.org/images/tree-icon.png">
105 * <body itemscope itemtype="http://schema.org/Product">
106 * <h1 itemprop="name">Shiny Trinket</h1>
107 * <img itemprop="image" src="{image-url}" />
108 * <p itemprop="description">Shiny trinkets are shiny.</p>
112 public static function getSiteinfo($url, $no_guessing = false, $do_oembed = true, $count = 1)
118 // Check if the URL does contain a scheme
119 $scheme = parse_url($url, PHP_URL_SCHEME);
122 $url = 'http://' . trim($url, '/');
126 logger('Endless loop detected for ' . $url, LOGGER_DEBUG);
130 $url = trim($url, "'");
131 $url = trim($url, '"');
133 $url = Network::stripTrackingQueryParams($url);
135 $siteinfo['url'] = $url;
136 $siteinfo['type'] = 'link';
138 $data = Network::curl($url);
139 if (!$data['success']) {
143 // If the file is too large then exit
144 if ($data['info']['download_content_length'] > 1000000) {
148 // If it isn't a HTML file then exit
149 if (($data['info']['content_type'] != '') && !strstr(strtolower($data['info']['content_type']), 'html')) {
153 $header = $data['header'];
154 $body = $data['body'];
157 $oembed_data = OEmbed::fetchURL($url);
159 if (!empty($oembed_data->type)) {
160 if (!in_array($oembed_data->type, ['error', 'rich', ''])) {
161 $siteinfo['type'] = $oembed_data->type;
164 if (($oembed_data->type == 'link') && ($siteinfo['type'] != 'photo')) {
165 if (isset($oembed_data->title)) {
166 $siteinfo['title'] = trim($oembed_data->title);
168 if (isset($oembed_data->description)) {
169 $siteinfo['text'] = trim($oembed_data->description);
171 if (isset($oembed_data->thumbnail_url)) {
172 $siteinfo['image'] = $oembed_data->thumbnail_url;
178 // Fetch the first mentioned charset. Can be in body or header
180 if (preg_match('/charset=(.*?)[\'"\s\n]/', $header, $matches)) {
181 $charset = trim(trim(trim(array_pop($matches)), ';,'));
184 if ($charset == '') {
188 if (($charset != '') && (strtoupper($charset) != 'UTF-8')) {
189 logger('detected charset ' . $charset, LOGGER_DEBUG);
190 $body = iconv($charset, 'UTF-8//TRANSLIT', $body);
193 $body = mb_convert_encoding($body, 'HTML-ENTITIES', 'UTF-8');
195 $doc = new DOMDocument();
196 @$doc->loadHTML($body);
198 XML::deleteNode($doc, 'style');
199 XML::deleteNode($doc, 'script');
200 XML::deleteNode($doc, 'option');
201 XML::deleteNode($doc, 'h1');
202 XML::deleteNode($doc, 'h2');
203 XML::deleteNode($doc, 'h3');
204 XML::deleteNode($doc, 'h4');
205 XML::deleteNode($doc, 'h5');
206 XML::deleteNode($doc, 'h6');
207 XML::deleteNode($doc, 'ol');
208 XML::deleteNode($doc, 'ul');
210 $xpath = new DOMXPath($doc);
212 $list = $xpath->query('//meta[@content]');
213 foreach ($list as $node) {
215 if ($node->attributes->length) {
216 foreach ($node->attributes as $attribute) {
217 $meta_tag[$attribute->name] = $attribute->value;
221 if (@$meta_tag['http-equiv'] == 'refresh') {
222 $path = $meta_tag['content'];
223 $pathinfo = explode(';', $path);
225 foreach ($pathinfo as $value) {
226 if (substr(strtolower($value), 0, 4) == 'url=') {
227 $content = substr($value, 4);
230 if ($content != '') {
231 $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
237 $list = $xpath->query('//title');
238 if ($list->length > 0) {
239 $siteinfo['title'] = trim($list->item(0)->nodeValue);
242 $list = $xpath->query('//meta[@name]');
243 foreach ($list as $node) {
245 if ($node->attributes->length) {
246 foreach ($node->attributes as $attribute) {
247 $meta_tag[$attribute->name] = $attribute->value;
251 if (empty($meta_tag['content'])) {
255 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
257 switch (strtolower($meta_tag['name'])) {
259 $siteinfo['title'] = trim($meta_tag['content']);
262 $siteinfo['text'] = trim($meta_tag['content']);
265 $siteinfo['image'] = $meta_tag['content'];
267 case 'twitter:image':
268 $siteinfo['image'] = $meta_tag['content'];
270 case 'twitter:image:src':
271 $siteinfo['image'] = $meta_tag['content'];
274 // Convert Twitter types in our own
275 if ($meta_tag['content'] == 'summary_large_image') {
276 $siteinfo['type'] = 'photo';
278 $siteinfo['type'] = $meta_tag['content'];
281 case 'twitter:description':
282 $siteinfo['text'] = trim($meta_tag['content']);
284 case 'twitter:title':
285 $siteinfo['title'] = trim($meta_tag['content']);
288 $siteinfo['title'] = trim($meta_tag['content']);
290 case 'dc.description':
291 $siteinfo['text'] = trim($meta_tag['content']);
294 $keywords = explode(',', $meta_tag['content']);
296 case 'news_keywords':
297 $keywords = explode(',', $meta_tag['content']);
302 if (isset($keywords)) {
303 $siteinfo['keywords'] = [];
304 foreach ($keywords as $keyword) {
305 if (!in_array(trim($keyword), $siteinfo['keywords'])) {
306 $siteinfo['keywords'][] = trim($keyword);
311 $list = $xpath->query('//meta[@property]');
312 foreach ($list as $node) {
314 if ($node->attributes->length) {
315 foreach ($node->attributes as $attribute) {
316 $meta_tag[$attribute->name] = $attribute->value;
320 if (!empty($meta_tag['content'])) {
321 $meta_tag['content'] = trim(html_entity_decode($meta_tag['content'], ENT_QUOTES, 'UTF-8'));
323 switch (strtolower($meta_tag['property'])) {
325 $siteinfo['image'] = $meta_tag['content'];
328 $siteinfo['title'] = trim($meta_tag['content']);
330 case 'og:description':
331 $siteinfo['text'] = trim($meta_tag['content']);
337 if ((@$siteinfo['image'] == '') && !$no_guessing) {
338 $list = $xpath->query('//img[@src]');
339 foreach ($list as $node) {
341 if ($node->attributes->length) {
342 foreach ($node->attributes as $attribute) {
343 $img_tag[$attribute->name] = $attribute->value;
347 $src = self::completeUrl($img_tag['src'], $url);
348 $photodata = Image::getInfoFromURL($src);
350 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
351 if ($photodata[0] > 300) {
352 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
355 if ($photodata[1] > 300) {
356 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
359 $siteinfo['images'][] = [
361 'width' => $photodata[0],
362 'height' => $photodata[1]
366 } elseif (!empty($siteinfo['image'])) {
367 $src = self::completeUrl($siteinfo['image'], $url);
369 unset($siteinfo['image']);
371 $photodata = Image::getInfoFromURL($src);
373 if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
374 $siteinfo['images'][] = ['src' => $src,
375 'width' => $photodata[0],
376 'height' => $photodata[1]];
380 if ((@$siteinfo['text'] == '') && (@$siteinfo['title'] != '') && !$no_guessing) {
383 $list = $xpath->query('//div[@class="article"]');
384 foreach ($list as $node) {
385 if (strlen($node->nodeValue) > 40) {
386 $text .= ' ' . trim($node->nodeValue);
391 $list = $xpath->query('//div[@class="content"]');
392 foreach ($list as $node) {
393 if (strlen($node->nodeValue) > 40) {
394 $text .= ' ' . trim($node->nodeValue);
399 // If none text was found then take the paragraph content
401 $list = $xpath->query('//p');
402 foreach ($list as $node) {
403 if (strlen($node->nodeValue) > 40) {
404 $text .= ' ' . trim($node->nodeValue);
410 $text = trim(str_replace(["\n", "\r"], [' ', ' '], $text));
412 while (strpos($text, ' ')) {
413 $text = trim(str_replace(' ', ' ', $text));
416 $siteinfo['text'] = trim(html_entity_decode(substr($text, 0, 350), ENT_QUOTES, 'UTF-8') . '...');
420 logger('Siteinfo for ' . $url . ' ' . print_r($siteinfo, true), LOGGER_DEBUG);
422 Addon::callHooks('getsiteinfo', $siteinfo);
428 * @brief Convert tags from CSV to an array
430 * @param string $string Tags
431 * @return array with formatted Hashtags
433 public static function convertTagsToArray($string)
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"]);
445 * @brief Add a hasht sign to a string
447 * This method is used as callback function
449 * @param string $tag The pure tag name
450 * @param int $k Counter for internal use
453 private static function arrAddHashes(&$tag, $k)
459 * @brief Add a scheme to an url
461 * The src attribute of some html elements (e.g. images)
462 * can miss the scheme so we need to add the correct
465 * @param string $url The url which possibly does have
466 * a missing scheme (a link to an image)
467 * @param string $scheme The url with a correct scheme
468 * (e.g. the url from the webpage which does contain the image)
470 * @return string The url with a scheme
472 private static function completeUrl($url, $scheme)
474 $urlarr = parse_url($url);
476 // If the url does allready have an scheme
477 // we can stop the process here
478 if (isset($urlarr["scheme"])) {
482 $schemearr = parse_url($scheme);
484 $complete = $schemearr["scheme"]."://".$schemearr["host"];
486 if (@$schemearr["port"] != "") {
487 $complete .= ":".$schemearr["port"];
490 if (strpos($urlarr["path"], "/") !== 0) {
494 $complete .= $urlarr["path"];
496 if (@$urlarr["query"] != "") {
497 $complete .= "?".$urlarr["query"];
500 if (@$urlarr["fragment"] != "") {
501 $complete .= "#".$urlarr["fragment"];