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("parseurl_getsiteinfo: 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($charset) != "UTF-8")) {
189 logger("parseurl_getsiteinfo: detected charset ".$charset, LOGGER_DEBUG);
190 //$body = mb_convert_encoding($body, "UTF-8", $charset);
191 $body = iconv($charset, "UTF-8//TRANSLIT", $body);
194 $body = mb_convert_encoding($body, 'HTML-ENTITIES', "UTF-8");
196 $doc = new DOMDocument();
197 @$doc->loadHTML($body);
199 XML::deleteNode($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");
211 $xpath = new DOMXPath($doc);
213 $list = $xpath->query("//meta[@content]");
214 foreach ($list as $node) {
216 if ($node->attributes->length) {
217 foreach ($node->attributes as $attribute) {
218 $attr[$attribute->name] = $attribute->value;
222 if (@$attr["http-equiv"] == "refresh") {
223 $path = $attr["content"];
224 $pathinfo = explode(";", $path);
226 foreach ($pathinfo as $value) {
227 if (substr(strtolower($value), 0, 4) == "url=") {
228 $content = substr($value, 4);
231 if ($content != "") {
232 $siteinfo = self::getSiteinfo($content, $no_guessing, $do_oembed, ++$count);
238 $list = $xpath->query("//title");
239 if ($list->length > 0) {
240 $siteinfo["title"] = trim($list->item(0)->nodeValue);
243 //$list = $xpath->query("head/meta[@name]");
244 $list = $xpath->query("//meta[@name]");
245 foreach ($list as $node) {
247 if ($node->attributes->length) {
248 foreach ($node->attributes as $attribute) {
249 $attr[$attribute->name] = $attribute->value;
253 if (!empty($attr["content"])) {
254 $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
256 switch (strtolower($attr["name"])) {
258 $siteinfo["title"] = trim($attr["content"]);
261 $siteinfo["text"] = trim($attr["content"]);
264 $siteinfo["image"] = $attr["content"];
266 case "twitter:image":
267 $siteinfo["image"] = $attr["content"];
269 case "twitter:image:src":
270 $siteinfo["image"] = $attr["content"];
273 if (($siteinfo["type"] == "") || ($attr["content"] == "photo")) {
274 $siteinfo["type"] = $attr["content"];
277 case "twitter:description":
278 $siteinfo["text"] = trim($attr["content"]);
280 case "twitter:title":
281 $siteinfo["title"] = trim($attr["content"]);
284 $siteinfo["title"] = trim($attr["content"]);
286 case "dc.description":
287 $siteinfo["text"] = trim($attr["content"]);
290 $keywords = explode(",", $attr["content"]);
292 case "news_keywords":
293 $keywords = explode(",", $attr["content"]);
297 if ($siteinfo["type"] == "summary") {
298 $siteinfo["type"] = "link";
302 if (isset($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("head/meta[@property]");
312 $list = $xpath->query("//meta[@property]");
313 foreach ($list as $node) {
315 if ($node->attributes->length) {
316 foreach ($node->attributes as $attribute) {
317 $attr[$attribute->name] = $attribute->value;
321 if (!empty($attr["content"])) {
322 $attr["content"] = trim(html_entity_decode($attr["content"], ENT_QUOTES, "UTF-8"));
324 switch (strtolower($attr["property"])) {
326 $siteinfo["image"] = $attr["content"];
329 $siteinfo["title"] = trim($attr["content"]);
331 case "og:description":
332 $siteinfo["text"] = trim($attr["content"]);
338 if ((@$siteinfo["image"] == "") && !$no_guessing) {
339 $list = $xpath->query("//img[@src]");
340 foreach ($list as $node) {
342 if ($node->attributes->length) {
343 foreach ($node->attributes as $attribute) {
344 $attr[$attribute->name] = $attribute->value;
348 $src = self::completeUrl($attr["src"], $url);
349 $photodata = Image::getInfoFromURL($src);
351 if (($photodata) && ($photodata[0] > 150) && ($photodata[1] > 150)) {
352 if ($photodata[0] > 300) {
353 $photodata[1] = round($photodata[1] * (300 / $photodata[0]));
356 if ($photodata[1] > 300) {
357 $photodata[0] = round($photodata[0] * (300 / $photodata[1]));
360 $siteinfo["images"][] = ["src" => $src,
361 "width" => $photodata[0],
362 "height" => $photodata[1]];
365 } elseif (!empty($siteinfo["image"])) {
366 $src = self::completeUrl($siteinfo["image"], $url);
368 unset($siteinfo["image"]);
370 $photodata = Image::getInfoFromURL($src);
372 if (($photodata) && ($photodata[0] > 10) && ($photodata[1] > 10)) {
373 $siteinfo["images"][] = ["src" => $src,
374 "width" => $photodata[0],
375 "height" => $photodata[1]];
379 if ((@$siteinfo["text"] == "") && (@$siteinfo["title"] != "") && !$no_guessing) {
382 $list = $xpath->query("//div[@class='article']");
383 foreach ($list as $node) {
384 if (strlen($node->nodeValue) > 40) {
385 $text .= " ".trim($node->nodeValue);
390 $list = $xpath->query("//div[@class='content']");
391 foreach ($list as $node) {
392 if (strlen($node->nodeValue) > 40) {
393 $text .= " ".trim($node->nodeValue);
398 // If none text was found then take the paragraph content
400 $list = $xpath->query("//p");
401 foreach ($list as $node) {
402 if (strlen($node->nodeValue) > 40) {
403 $text .= " ".trim($node->nodeValue);
409 $text = trim(str_replace(["\n", "\r"], [" ", " "], $text));
411 while (strpos($text, " ")) {
412 $text = trim(str_replace(" ", " ", $text));
415 $siteinfo["text"] = trim(html_entity_decode(substr($text, 0, 350), ENT_QUOTES, "UTF-8").'...');
419 logger("parseurl_getsiteinfo: Siteinfo for ".$url." ".print_r($siteinfo, true), LOGGER_DEBUG);
421 Addon::callHooks("getsiteinfo", $siteinfo);
427 * @brief Convert tags from CSV to an array
429 * @param string $string Tags
430 * @return array with formatted Hashtags
432 public static function convertTagsToArray($string)
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"]);
444 * @brief Add a hasht sign to a string
446 * This method is used as callback function
448 * @param string $tag The pure tag name
449 * @param int $k Counter for internal use
452 private static function arrAddHashes(&$tag, $k)
458 * @brief Add a scheme to an url
460 * The src attribute of some html elements (e.g. images)
461 * can miss the scheme so we need to add the correct
464 * @param string $url The url which possibly does have
465 * a missing scheme (a link to an image)
466 * @param string $scheme The url with a correct scheme
467 * (e.g. the url from the webpage which does contain the image)
469 * @return string The url with a scheme
471 private static function completeUrl($url, $scheme)
473 $urlarr = parse_url($url);
475 // If the url does allready have an scheme
476 // we can stop the process here
477 if (isset($urlarr["scheme"])) {
481 $schemearr = parse_url($scheme);
483 $complete = $schemearr["scheme"]."://".$schemearr["host"];
485 if (@$schemearr["port"] != "") {
486 $complete .= ":".$schemearr["port"];
489 if (strpos($urlarr["path"], "/") !== 0) {
493 $complete .= $urlarr["path"];
495 if (@$urlarr["query"] != "") {
496 $complete .= "?".$urlarr["query"];
499 if (@$urlarr["fragment"] != "") {
500 $complete .= "#".$urlarr["fragment"];