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 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(dirname(__FILE__).'/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 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         $this->file_url = optional_param('file', '', PARAM_RAW);
  50     }
  51
  52     public function check_login() {
  53         if (!empty($this->file_url)) {
  54             return true;
  55         } else {
  56             return false;
  57         }
  58     }
  59     /**
  60      * @return mixed
  61      */
  62     public function print_login() {
  63         $strdownload = get_string('download', 'repository');
  64         $strname     = get_string('rename', 'repository_url');
  65         $strurl      = get_string('url', 'repository_url');
  66         if ($this->options['ajax']) {
  67             $url = new stdClass();
  68             $url->label = $strurl.': ';
  69             $url->id   = 'fileurl';
  70             $url->type = 'text';
  71             $url->name = 'file';
  72
  73             $ret['login'] = array($url);
  74             $ret['login_btn_label'] = get_string('download', 'repository_url');
  75             $ret['allowcaching'] = true; // indicates that login form can be cached in filepicker.js
  76             return $ret;
  77         } else {
  78             echo <<<EOD
  79 <table>
  80 <tr>
  81 <td>{$strurl}: </td><td><input name="file" type="text" /></td>
  82 </tr>
  83 </table>
  84 <input type="submit" value="{$strdownload}" />
  85 EOD;
  86
  87         }
  88     }
  89
  90     /**
  91      * @param mixed $path
  92      * @param string $search
  93      * @return array
  94      */
  95     public function get_listing($path='', $page='') {
  96         global $CFG, $OUTPUT;
  97         $ret = array();
  98         $ret['list'] = array();
  99         $ret['nosearch'] = true;
 100         $ret['norefresh'] = true;
 101         $ret['nologin'] = true;
 102
 103         $this->parse_file(null, $this->file_url, $ret, true);
 104         return $ret;
 105     }
 106
 107     /**
 108      * Parses one file (either html or css)
 109      *
 110      * @param string $baseurl (optional) URL of the file where link to this file was found
 111      * @param string $relativeurl relative or absolute link to the file
 112      * @param array $list
 113      * @param bool $mainfile true only for main HTML false and false for all embedded/linked files
 114      */
 115     protected function parse_file($baseurl, $relativeurl, &$list, $mainfile = false) {
 116         if (preg_match('/([\'"])(.*)\1/', $relativeurl, $matches)) {
 117             $relativeurl = $matches[2];
 118         }
 119         if (empty($baseurl)) {
 120             $url = $relativeurl;
 121         } else {
 122             $url = htmlspecialchars_decode(url_to_absolute($baseurl, $relativeurl));
 123         }
 124         if (in_array($url, $this->processedfiles)) {
 125             // avoid endless recursion
 126             return;
 127         }
 128         $this->processedfiles[] = $url;
 129         $curl = new curl;
 130         $curl->setopt(array('CURLOPT_FOLLOWLOCATION' => true, 'CURLOPT_MAXREDIRS' => 3));
 131         $msg = $curl->head($url);
 132         $info = $curl->get_info();
 133         if ($info['http_code'] != 200) {
 134             if ($mainfile) {
 135                 $list['error'] = $msg;
 136             }
 137         } else {
 138             $csstoanalyze = '';
 139             if ($mainfile && (strstr($info['content_type'], 'text/html') || empty($info['content_type']))) {
 140                 // parse as html
 141                 $htmlcontent = $curl->get($info['url']);
 142                 $ddoc = new DOMDocument();
 143                 @$ddoc->loadHTML($htmlcontent);
 144                 // extract <img>
 145                 $tags = $ddoc->getElementsByTagName('img');
 146                 foreach ($tags as $tag) {
 147                     $url = $tag->getAttribute('src');
 148                     $this->add_image_to_list($info['url'], $url, $list);
 149                 }
 150                 // analyse embedded css (<style>)
 151                 $tags = $ddoc->getElementsByTagName('style');
 152                 foreach ($tags as $tag) {
 153                     if ($tag->getAttribute('type') == 'text/css') {
 154                         $csstoanalyze .= $tag->textContent."\n";
 155                     }
 156                 }
 157                 // analyse links to css (<link type='text/css' href='...'>)
 158                 $tags = $ddoc->getElementsByTagName('link');
 159                 foreach ($tags as $tag) {
 160                     if ($tag->getAttribute('type') == 'text/css' && strlen($tag->getAttribute('href'))) {
 161                         $this->parse_file($info['url'], $tag->getAttribute('href'), $list);
 162                     }
 163                 }
 164             } else if (strstr($info['content_type'], 'css')) {
 165                 // parse as css
 166                 $csscontent = $curl->get($info['url']);
 167                 $csstoanalyze .= $csscontent."\n";
 168             } else if (strstr($info['content_type'], 'image/')) {
 169                 // download this file
 170                 $this->add_image_to_list($info['url'], $info['url'], $list);
 171             }
 172
 173             // parse all found css styles
 174             if (strlen($csstoanalyze)) {
 175                 $urls = extract_css_urls($csstoanalyze);
 176                 if (!empty($urls['property'])) {
 177                     foreach ($urls['property'] as $url) {
 178                         $this->add_image_to_list($info['url'], $url, $list);
 179                     }
 180                 }
 181                 if (!empty($urls['import'])) {
 182                     foreach ($urls['import'] as $cssurl) {
 183                         $this->parse_file($info['url'], $cssurl, $list);
 184                     }
 185                 }
 186             }
 187         }
 188     }
 189     protected function add_image_to_list($baseurl, $url, &$list) {
 190         if (empty($list['list'])) {
 191             $list['list'] = array();
 192         }
 193         $src = url_to_absolute($baseurl, htmlspecialchars_decode($url));
 194         foreach ($list['list'] as $image) {
 195             if ($image['source'] == $src) {
 196                 return;
 197             }
 198         }
 199         $list['list'][] = array(
 200             'title'=>$this->guess_filename($url, ''),
 201             'source'=>$src,
 202             'thumbnail'=>$src,
 203             'thumbnail_height'=>84,
 204             'thumbnail_width'=>84
 205         );
 206     }
 207     public function guess_filename($url, $type) {
 208         $pattern = '#\/([\w_\?\-.]+)$#';
 209         $matches = null;
 210         preg_match($pattern, $url, $matches);
 211         if (empty($matches[1])) {
 212             return $url;
 213         } else {
 214             return $matches[1];
 215         }
 216     }
 217
 218     public function supported_returntypes() {
 219         return (FILE_INTERNAL | FILE_EXTERNAL);
 220     }
 221
 222     /**
 223      * Return the source information
 224      *
 225      * @param stdClass $url
 226      * @return string|null
 227      */
 228     public function get_file_source_info($url) {
 229         return $url;
 230     }
 231
 232     /**
 233      * file types supported by url downloader plugin
 234      *
 235      * @return array
 236      */
 237     public function supported_filetypes() {
 238         return array('web_image');
 239     }
 240 }