application/bookmark/LinkUtils.php

   1 <?php
   2
   3 use Shaarli\Bookmark\Bookmark;
   4
   5 /**
   6  * Get cURL callback function for CURLOPT_WRITEFUNCTION
   7  *
   8  * @param string $charset     to extract from the downloaded page (reference)
   9  * @param string $title       to extract from the downloaded page (reference)
  10  * @param string $description to extract from the downloaded page (reference)
  11  * @param string $keywords    to extract from the downloaded page (reference)
  12  * @param bool   $retrieveDescription Automatically tries to retrieve description and keywords from HTML content
  13  * @param string $curlGetInfo Optionally overrides curl_getinfo function
  14  *
  15  * @return Closure
  16  */
  17 function get_curl_download_callback(
  18     &$charset,
  19     &$title,
  20     &$description,
  21     &$keywords,
  22     $retrieveDescription,
  23     $curlGetInfo = 'curl_getinfo'
  24 ) {
  25     $isRedirected = false;
  26     $currentChunk = 0;
  27     $foundChunk = null;
  28
  29     /**
  30      * cURL callback function for CURLOPT_WRITEFUNCTION (called during the download).
  31      *
  32      * While downloading the remote page, we check that the HTTP code is 200 and content type is 'html/text'
  33      * Then we extract the title and the charset and stop the download when it's done.
  34      *
  35      * @param resource $ch   cURL resource
  36      * @param string   $data chunk of data being downloaded
  37      *
  38      * @return int|bool length of $data or false if we need to stop the download
  39      */
  40     return function (&$ch, $data) use (
  41         $retrieveDescription,
  42         $curlGetInfo,
  43         &$charset,
  44         &$title,
  45         &$description,
  46         &$keywords,
  47         &$isRedirected,
  48         &$currentChunk,
  49         &$foundChunk
  50     ) {
  51         $currentChunk++;
  52         $responseCode = $curlGetInfo($ch, CURLINFO_RESPONSE_CODE);
  53         if (!empty($responseCode) && in_array($responseCode, [301, 302])) {
  54             $isRedirected = true;
  55             return strlen($data);
  56         }
  57         if (!empty($responseCode) && $responseCode !== 200) {
  58             return false;
  59         }
  60         // After a redirection, the content type will keep the previous request value
  61         // until it finds the next content-type header.
  62         if (! $isRedirected || strpos(strtolower($data), 'content-type') !== false) {
  63             $contentType = $curlGetInfo($ch, CURLINFO_CONTENT_TYPE);
  64         }
  65         if (!empty($contentType) && strpos($contentType, 'text/html') === false) {
  66             return false;
  67         }
  68         if (!empty($contentType) && empty($charset)) {
  69             $charset = header_extract_charset($contentType);
  70         }
  71         if (empty($charset)) {
  72             $charset = html_extract_charset($data);
  73         }
  74         if (empty($title)) {
  75             $title = html_extract_title($data);
  76             $foundChunk = ! empty($title) ? $currentChunk : $foundChunk;
  77         }
  78         if ($retrieveDescription && empty($description)) {
  79             $description = html_extract_tag('description', $data);
  80             $foundChunk = ! empty($description) ? $currentChunk : $foundChunk;
  81         }
  82         if ($retrieveDescription && empty($keywords)) {
  83             $keywords = html_extract_tag('keywords', $data);
  84             if (! empty($keywords)) {
  85                 $foundChunk = $currentChunk;
  86                 // Keywords use the format tag1, tag2 multiple words, tag
  87                 // So we format them to match Shaarli's separator and glue multiple words with '-'
  88                 $keywords = implode(' ', array_map(function($keyword) {
  89                     return implode('-', preg_split('/\s+/', trim($keyword)));
  90                 }, explode(',', $keywords)));
  91             }
  92         }
  93
  94         // We got everything we want, stop the download.
  95         // If we already found either the title, description or keywords,
  96         // it's highly unlikely that we'll found the other metas further than
  97         // in the same chunk of data or the next one. So we also stop the download after that.
  98         if ((!empty($responseCode) && !empty($contentType) && !empty($charset)) && $foundChunk !== null
  99             && (! $retrieveDescription
 100                 || $foundChunk < $currentChunk
 101                 || (!empty($title) && !empty($description) && !empty($keywords))
 102             )
 103         ) {
 104             return false;
 105         }
 106
 107         return strlen($data);
 108     };
 109 }
 110
 111 /**
 112  * Extract title from an HTML document.
 113  *
 114  * @param string $html HTML content where to look for a title.
 115  *
 116  * @return bool|string Extracted title if found, false otherwise.
 117  */
 118 function html_extract_title($html)
 119 {
 120     if (preg_match('!<title.*?>(.*?)</title>!is', $html, $matches)) {
 121         return trim(str_replace("\n", '', $matches[1]));
 122     }
 123     return false;
 124 }
 125
 126 /**
 127  * Extract charset from HTTP header if it's defined.
 128  *
 129  * @param string $header HTTP header Content-Type line.
 130  *
 131  * @return bool|string Charset string if found (lowercase), false otherwise.
 132  */
 133 function header_extract_charset($header)
 134 {
 135     preg_match('/charset="?([^; ]+)/i', $header, $match);
 136     if (! empty($match[1])) {
 137         return strtolower(trim($match[1]));
 138     }
 139
 140     return false;
 141 }
 142
 143 /**
 144  * Extract charset HTML content (tag <meta charset>).
 145  *
 146  * @param string $html HTML content where to look for charset.
 147  *
 148  * @return bool|string Charset string if found, false otherwise.
 149  */
 150 function html_extract_charset($html)
 151 {
 152     // Get encoding specified in HTML header.
 153     preg_match('#<meta .*charset=["\']?([^";\'>/]+)["\']? */?>#Usi', $html, $enc);
 154     if (!empty($enc[1])) {
 155         return strtolower($enc[1]);
 156     }
 157
 158     return false;
 159 }
 160
 161 /**
 162  * Extract meta tag from HTML content in either:
 163  *   - OpenGraph: <meta property="og:[tag]" ...>
 164  *   - Meta tag: <meta name="[tag]" ...>
 165  *
 166  * @param string $tag  Name of the tag to retrieve.
 167  * @param string $html HTML content where to look for charset.
 168  *
 169  * @return bool|string Charset string if found, false otherwise.
 170  */
 171 function html_extract_tag($tag, $html)
 172 {
 173     $propertiesKey = ['property', 'name', 'itemprop'];
 174     $properties = implode('|', $propertiesKey);
 175     // Try to retrieve OpenGraph image.
 176     $ogRegex = '#<meta[^>]+(?:'. $properties .')=["\']?(?:og:)?'. $tag .'["\'\s][^>]*content=["\']?(.*?)["\'/>]#';
 177     // If the attributes are not in the order property => content (e.g. Github)
 178     // New regex to keep this readable... more or less.
 179     $ogRegexReverse = '#<meta[^>]+content=["\']([^"\']+)[^>]+(?:'. $properties .')=["\']?(?:og)?:'. $tag .'["\'\s/>]#';
 180
 181     if (preg_match($ogRegex, $html, $matches) > 0
 182         || preg_match($ogRegexReverse, $html, $matches) > 0
 183     ) {
 184         return $matches[1];
 185     }
 186
 187     return false;
 188 }
 189
 190 /**
 191  * In a string, converts URLs to clickable bookmarks.
 192  *
 193  * @param string $text       input string.
 194  *
 195  * @return string returns $text with all bookmarks converted to HTML bookmarks.
 196  *
 197  * @see Function inspired from http://www.php.net/manual/en/function.preg-replace.php#85722
 198  */
 199 function text2clickable($text)
 200 {
 201     $regex = '!(((?:https?|ftp|file)://|apt:|magnet:)\S+[a-z0-9\(\)]/?)!si';
 202     return preg_replace($regex, '<a href="$1">$1</a>', $text);
 203 }
 204
 205 /**
 206  * Auto-link hashtags.
 207  *
 208  * @param string $description Given description.
 209  * @param string $indexUrl    Root URL.
 210  *
 211  * @return string Description with auto-linked hashtags.
 212  */
 213 function hashtag_autolink($description, $indexUrl = '')
 214 {
 215     /*
 216      * To support unicode: http://stackoverflow.com/a/35498078/1484919
 217      * \p{Pc} - to match underscore
 218      * \p{N} - numeric character in any script
 219      * \p{L} - letter from any language
 220      * \p{Mn} - any non marking space (accents, umlauts, etc)
 221      */
 222     $regex = '/(^|\s)#([\p{Pc}\p{N}\p{L}\p{Mn}]+)/mui';
 223     $replacement = '$1<a href="'. $indexUrl .'?addtag=$2" title="Hashtag $2">#$2</a>';
 224     return preg_replace($regex, $replacement, $description);
 225 }
 226
 227 /**
 228  * This function inserts &nbsp; where relevant so that multiple spaces are properly displayed in HTML
 229  * even in the absence of <pre>  (This is used in description to keep text formatting).
 230  *
 231  * @param string $text input text.
 232  *
 233  * @return string formatted text.
 234  */
 235 function space2nbsp($text)
 236 {
 237     return preg_replace('/(^| ) /m', '$1&nbsp;', $text);
 238 }
 239
 240 /**
 241  * Format Shaarli's description
 242  *
 243  * @param string $description shaare's description.
 244  * @param string $indexUrl    URL to Shaarli's index.
 245
 246  * @return string formatted description.
 247  */
 248 function format_description($description, $indexUrl = '')
 249 {
 250     return nl2br(space2nbsp(hashtag_autolink(text2clickable($description), $indexUrl)));
 251 }
 252
 253 /**
 254  * Generate a small hash for a link.
 255  *
 256  * @param DateTime $date Link creation date.
 257  * @param int      $id   Link ID.
 258  *
 259  * @return string the small hash generated from link data.
 260  */
 261 function link_small_hash($date, $id)
 262 {
 263     return smallHash($date->format(Bookmark::LINK_DATE_FORMAT) . $id);
 264 }
 265
 266 /**
 267  * Returns whether or not the link is an internal note.
 268  * Its URL starts by `?` because it's actually a permalink.
 269  *
 270  * @param string $linkUrl
 271  *
 272  * @return bool true if internal note, false otherwise.
 273  */
 274 function is_note($linkUrl)
 275 {
 276     return isset($linkUrl[0]) && $linkUrl[0] === '?';
 277 }