repository/url/lib.php

   1 <?php
   2
   3 // This file is part of Moodle - http://moodle.org/
   4 //
   5 // Moodle is free software: you can redistribute it and/or modify
   6 // it under the terms of the GNU General Public License as published by
   7 // the Free Software Foundation, either version 3 of the License, or
   8 // (at your option) any later version.
   9 //
  10 // Moodle is distributed in the hope that it will be useful,
  11 // but WITHOUT ANY WARRANTY; without even the implied warranty of
  12 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13 // GNU General Public License for more details.
  14 //
  15 // You should have received a copy of the GNU General Public License
  16 // along with Moodle.  If not, see <http://www.gnu.org/licenses/>.
  17
  18 /**
  19  * This plugin is used to access files by providing an url
  20  *
  21  * @since Moodle 2.0
  22  * @package    repository_url
  23  * @copyright  2010 Dongsheng Cai {@link http://dongsheng.org}
  24  * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
  25  */
  26 require_once($CFG->dirroot . '/repository/lib.php');
  27 require_once(__DIR__.'/locallib.php');
  28
  29 /**
  30  * repository_url class
  31  * A subclass of repository, which is used to download a file from a specific url
  32  *
  33  * @since Moodle 2.0
  34  * @package    repository_url
  35  * @copyright  2009 Dongsheng Cai {@link http://dongsheng.org}
  36  * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
  37  */
  38 class repository_url extends repository {
  39     var $processedfiles = array();
  40
  41     /**
  42      * @param int $repositoryid
  43      * @param object $context
  44      * @param array $options
  45      */
  46     public function __construct($repositoryid, $context = SYSCONTEXTID, $options = array()){
  47         global $CFG;
  48         parent::__construct($repositoryid, $context, $options);
  49         $this->file_url = optional_param('file', '', PARAM_RAW);
  50         $this->file_url = $this->escape_url($this->file_url);
  51     }
  52
  53     public function check_login() {
  54         if (!empty($this->file_url)) {
  55             return true;
  56         } else {
  57             return false;
  58         }
  59     }
  60     /**
  61      * @return mixed
  62      */
  63     public function print_login() {
  64         $strdownload = get_string('download', 'repository');
  65         $strname     = get_string('rename', 'repository_url');
  66         $strurl      = get_string('url', 'repository_url');
  67         if ($this->options['ajax']) {
  68             $url = new stdClass();
  69             $url->label = $strurl.': ';
  70             $url->id   = 'fileurl';
  71             $url->type = 'text';
  72             $url->name = 'file';
  73
  74             $ret['login'] = array($url);
  75             $ret['login_btn_label'] = get_string('download', 'repository_url');
  76             $ret['allowcaching'] = true; // indicates that login form can be cached in filepicker.js
  77             return $ret;
  78         } else {
  79             echo <<<EOD
  80 <table>
  81 <tr>
  82 <td>{$strurl}: </td><td><input name="file" type="text" /></td>
  83 </tr>
  84 </table>
  85 <input type="submit" value="{$strdownload}" />
  86 EOD;
  87
  88         }
  89     }
  90
  91     /**
  92      * @param mixed $path
  93      * @param string $search
  94      * @return array
  95      */
  96     public function get_listing($path='', $page='') {
  97         $ret = array();
  98         $ret['list'] = array();
  99         $ret['nosearch'] = true;
 100         $ret['norefresh'] = true;
 101         $ret['nologin'] = true;
 102
 103         $this->file_url = clean_param($this->file_url, PARAM_URL);
 104         if (empty($this->file_url)) {
 105             throw new repository_exception('validfiletype', 'repository_url');
 106         }
 107
 108         $this->parse_file(null, $this->file_url, $ret, true);
 109         return $ret;
 110     }
 111
 112     /**
 113      * Parses one file (either html or css)
 114      *
 115      * @param string $baseurl (optional) URL of the file where link to this file was found
 116      * @param string $relativeurl relative or absolute link to the file
 117      * @param array $list
 118      * @param bool $mainfile true only for main HTML false and false for all embedded/linked files
 119      */
 120     protected function parse_file($baseurl, $relativeurl, &$list, $mainfile = false) {
 121         if (preg_match('/([\'"])(.*)\1/', $relativeurl, $matches)) {
 122             $relativeurl = $matches[2];
 123         }
 124         if (empty($baseurl)) {
 125             $url = $relativeurl;
 126         } else {
 127             $url = htmlspecialchars_decode(url_to_absolute($baseurl, $relativeurl));
 128         }
 129         if (in_array($url, $this->processedfiles)) {
 130             // avoid endless recursion
 131             return;
 132         }
 133         $this->processedfiles[] = $url;
 134         $curl = new curl;
 135         $curl->setopt(array('CURLOPT_FOLLOWLOCATION' => true, 'CURLOPT_MAXREDIRS' => 3));
 136         $msg = $curl->head($url);
 137         $info = $curl->get_info();
 138         if ($info['http_code'] != 200) {
 139             if ($mainfile) {
 140                 $list['error'] = $msg;
 141             }
 142         } else {
 143             $csstoanalyze = '';
 144             if ($mainfile && (strstr($info['content_type'], 'text/html') || empty($info['content_type']))) {
 145                 // parse as html
 146                 $htmlcontent = $curl->get($info['url']);
 147                 $ddoc = new DOMDocument();
 148                 @$ddoc->loadHTML($htmlcontent);
 149                 // extract <img>
 150                 $tags = $ddoc->getElementsByTagName('img');
 151                 foreach ($tags as $tag) {
 152                     $url = $tag->getAttribute('src');
 153                     $this->add_image_to_list($info['url'], $url, $list);
 154                 }
 155                 // analyse embedded css (<style>)
 156                 $tags = $ddoc->getElementsByTagName('style');
 157                 foreach ($tags as $tag) {
 158                     if ($tag->getAttribute('type') == 'text/css') {
 159                         $csstoanalyze .= $tag->textContent."\n";
 160                     }
 161                 }
 162                 // analyse links to css (<link type='text/css' href='...'>)
 163                 $tags = $ddoc->getElementsByTagName('link');
 164                 foreach ($tags as $tag) {
 165                     if ($tag->getAttribute('type') == 'text/css' && strlen($tag->getAttribute('href'))) {
 166                         $this->parse_file($info['url'], $tag->getAttribute('href'), $list);
 167                     }
 168                 }
 169             } else if (strstr($info['content_type'], 'css')) {
 170                 // parse as css
 171                 $csscontent = $curl->get($info['url']);
 172                 $csstoanalyze .= $csscontent."\n";
 173             } else if (strstr($info['content_type'], 'image/')) {
 174                 // download this file
 175                 $this->add_image_to_list($info['url'], $info['url'], $list);
 176             } else {
 177                 $list['error'] = get_string('validfiletype', 'repository_url');
 178             }
 179
 180             // parse all found css styles
 181             if (strlen($csstoanalyze)) {
 182                 $urls = extract_css_urls($csstoanalyze);
 183                 if (!empty($urls['property'])) {
 184                     foreach ($urls['property'] as $url) {
 185                         $this->add_image_to_list($info['url'], $url, $list);
 186                     }
 187                 }
 188                 if (!empty($urls['import'])) {
 189                     foreach ($urls['import'] as $cssurl) {
 190                         $this->parse_file($info['url'], $cssurl, $list);
 191                     }
 192                 }
 193             }
 194         }
 195     }
 196     protected function add_image_to_list($baseurl, $url, &$list) {
 197         if (empty($list['list'])) {
 198             $list['list'] = array();
 199         }
 200         $src = url_to_absolute($baseurl, htmlspecialchars_decode($url));
 201         foreach ($list['list'] as $image) {
 202             if ($image['source'] == $src) {
 203                 return;
 204             }
 205         }
 206         $list['list'][] = array(
 207             'title'=>$this->guess_filename($url, ''),
 208             'source'=>$src,
 209             'thumbnail'=>$src,
 210             'thumbnail_height'=>84,
 211             'thumbnail_width'=>84
 212         );
 213     }
 214     public function guess_filename($url, $type) {
 215         $pattern = '#\/([\w_\?\-.]+)$#';
 216         $matches = null;
 217         preg_match($pattern, $url, $matches);
 218         if (empty($matches[1])) {
 219             return $url;
 220         } else {
 221             return $matches[1];
 222         }
 223     }
 224
 225     /**
 226      * Escapes a url by replacing spaces with %20.
 227      *
 228      * Note: In general moodle does not automatically escape urls, but for the purposes of making this plugin more user friendly
 229      * and make it consistent with some other areas in moodle (such as mod_url), urls will automatically be escaped.
 230      *
 231      * If moodle_url or PARAM_URL is changed to clean characters that need to be escaped, then this function can be removed
 232      *
 233      * @param string $url An unescaped url.
 234      * @return string The escaped url
 235      */
 236     protected function escape_url($url) {
 237         $url = str_replace('"', '%22', $url);
 238         $url = str_replace('\'', '%27', $url);
 239         $url = str_replace(' ', '%20', $url);
 240         $url = str_replace('<', '%3C', $url);
 241         $url = str_replace('>', '%3E', $url);
 242         return $url;
 243     }
 244
 245     public function supported_returntypes() {
 246         return (FILE_INTERNAL | FILE_EXTERNAL);
 247     }
 248
 249     /**
 250      * Return the source information
 251      *
 252      * @param stdClass $url
 253      * @return string|null
 254      */
 255     public function get_file_source_info($url) {
 256         return $url;
 257     }
 258
 259     /**
 260      * file types supported by url downloader plugin
 261      *
 262      * @return array
 263      */
 264     public function supported_filetypes() {
 265         return array('web_image');
 266     }
 267
 268     /**
 269      * Is this repository accessing private data?
 270      *
 271      * @return bool
 272      */
 273     public function contains_private_data() {
 274         return false;
 275     }
 276 }