inc/3rdparty/makefulltextfeedHelpers.php

   1 <?php
   2
   3 // Autoloading of classes allows us to include files only when they're
   4 // needed. If we've got a cached copy, for example, only Zend_Cache is loaded.
   5 function autoload($class_name) {
   6         static $dir = null;
   7         if ($dir === null) $dir = dirname(__FILE__).'/libraries/';
   8         static $mapping = array(
   9                 // Include FeedCreator for RSS/Atom creation
  10                 'FeedWriter' => 'feedwriter/FeedWriter.php',
  11                 'FeedItem' => 'feedwriter/FeedItem.php',
  12                 // Include ContentExtractor and Readability for identifying and extracting content from URLs
  13                 'ContentExtractor' => 'content-extractor/ContentExtractor.php',
  14                 'SiteConfig' => 'content-extractor/SiteConfig.php',
  15                 'Readability' => 'readability/Readability.php',
  16                 // Include Humble HTTP Agent to allow parallel requests and response caching
  17                 'HumbleHttpAgent' => 'humble-http-agent/HumbleHttpAgent.php',
  18                 'SimplePie_HumbleHttpAgent' => 'humble-http-agent/SimplePie_HumbleHttpAgent.php',
  19                 'CookieJar' => 'humble-http-agent/CookieJar.php',
  20                 // Include Zend Cache to improve performance (cache results)
  21                 'Zend_Cache' => 'Zend/Cache.php',
  22                 // Language detect
  23                 'Text_LanguageDetect' => 'language-detect/LanguageDetect.php',
  24                 // HTML5 Lib
  25                 'HTML5_Parser' => 'html5/Parser.php',
  26                 // htmLawed - used if XSS filter is enabled (xss_filter)
  27                 'htmLawed' => 'htmLawed/htmLawed.php'
  28         );
  29         if (isset($mapping[$class_name])) {
  30                 debug("** Loading class $class_name ({$mapping[$class_name]})");
  31                 require $dir.$mapping[$class_name];
  32                 return true;
  33         } else {
  34                 return false;
  35         }
  36 }
  37 spl_autoload_register('autoload');
  38 require dirname(__FILE__).'/libraries/simplepie/autoloader.php';
  39
  40
  41 class DummySingleItemFeed {
  42         public $item;
  43         function __construct($url) { $this->item = new DummySingleItem($url); }
  44         public function get_title() { return ''; }
  45         public function get_description() { return 'Content extracted from '.$this->item->url; }
  46         public function get_link() { return $this->item->url; }
  47         public function get_language() { return false; }
  48         public function get_image_url() { return false; }
  49         public function get_items($start=0, $max=1) { return array(0=>$this->item); }
  50 }
  51 class DummySingleItem {
  52         public $url;
  53         function __construct($url) { $this->url = $url; }
  54         public function get_permalink() { return $this->url; }
  55         public function get_title() { return null; }
  56         public function get_date($format='') { return false; }
  57         public function get_author($key=0) { return null; }
  58         public function get_authors() { return null; }
  59         public function get_description() { return ''; }
  60         public function get_enclosure($key=0, $prefer=null) { return null; }
  61         public function get_enclosures() { return null; }
  62         public function get_categories() { return null; }
  63 }
  64
  65 ///////////////////////////////
  66 // HELPER FUNCTIONS
  67 ///////////////////////////////
  68
  69 function url_allowed($url) {
  70         global $options;
  71         if (!empty($options->allowed_urls)) {
  72                 $allowed = false;
  73                 foreach ($options->allowed_urls as $allowurl) {
  74                         if (stristr($url, $allowurl) !== false) {
  75                                 $allowed = true;
  76                                 break;
  77                         }
  78                 }
  79                 if (!$allowed) return false;
  80         } else {
  81                 foreach ($options->blocked_urls as $blockurl) {
  82                         if (stristr($url, $blockurl) !== false) {
  83                                 return false;
  84                         }
  85                 }
  86         }
  87         return true;
  88 }
  89
  90 //////////////////////////////////////////////
  91 // Convert $html to UTF8
  92 // (uses HTTP headers and HTML to find encoding)
  93 // adapted from http://stackoverflow.com/questions/910793/php-detect-encoding-and-make-everything-utf-8
  94 //////////////////////////////////////////////
  95 function convert_to_utf8($html, $header=null)
  96 {
  97         $encoding = null;
  98         if ($html || $header) {
  99                 if (is_array($header)) $header = implode("\n", $header);
 100                 if (!$header || !preg_match_all('/^Content-Type:\s+([^;]+)(?:;\s*charset=["\']?([^;"\'\n]*))?/im', $header, $match, PREG_SET_ORDER)) {
 101                         // error parsing the response
 102                         debug('Could not find Content-Type header in HTTP response');
 103                 } else {
 104                         $match = end($match); // get last matched element (in case of redirects)
 105                         if (isset($match[2])) $encoding = trim($match[2], "\"' \r\n\0\x0B\t");
 106                 }
 107                 // TODO: check to see if encoding is supported (can we convert it?)
 108                 // If it's not, result will be empty string.
 109                 // For now we'll check for invalid encoding types returned by some sites, e.g. 'none'
 110                 // Problem URL: http://facta.co.jp/blog/archives/20111026001026.html
 111                 if (!$encoding || $encoding == 'none') {
 112                         // search for encoding in HTML - only look at the first 50000 characters
 113                         // Why 50000? See, for example, http://www.lemonde.fr/festival-de-cannes/article/2012/05/23/deux-cretes-en-goguette-sur-la-croisette_1705732_766360.html
 114                         // TODO: improve this so it looks at smaller chunks first
 115                         $html_head = substr($html, 0, 50000);
 116                         if (preg_match('/^<\?xml\s+version=(?:"[^"]*"|\'[^\']*\')\s+encoding=("[^"]*"|\'[^\']*\')/s', $html_head, $match)) {
 117                                 $encoding = trim($match[1], '"\'');
 118                         } elseif (preg_match('/<meta\s+http-equiv=["\']?Content-Type["\']? content=["\'][^;]+;\s*charset=["\']?([^;"\'>]+)/i', $html_head, $match)) {
 119                                 $encoding = trim($match[1]);
 120                         } elseif (preg_match_all('/<meta\s+([^>]+)>/i', $html_head, $match)) {
 121                                 foreach ($match[1] as $_test) {
 122                                         if (preg_match('/charset=["\']?([^"\']+)/i', $_test, $_m)) {
 123                                                 $encoding = trim($_m[1]);
 124                                                 break;
 125                                         }
 126                                 }
 127                         }
 128                 }
 129                 if (isset($encoding)) $encoding = trim($encoding);
 130                 // trim is important here!
 131                 if (!$encoding || (strtolower($encoding) == 'iso-8859-1')) {
 132                         // replace MS Word smart qutoes
 133                         $trans = array();
 134                         $trans[chr(130)] = '&sbquo;';    // Single Low-9 Quotation Mark
 135                         $trans[chr(131)] = '&fnof;';    // Latin Small Letter F With Hook
 136                         $trans[chr(132)] = '&bdquo;';    // Double Low-9 Quotation Mark
 137                         $trans[chr(133)] = '&hellip;';    // Horizontal Ellipsis
 138                         $trans[chr(134)] = '&dagger;';    // Dagger
 139                         $trans[chr(135)] = '&Dagger;';    // Double Dagger
 140                         $trans[chr(136)] = '&circ;';    // Modifier Letter Circumflex Accent
 141                         $trans[chr(137)] = '&permil;';    // Per Mille Sign
 142                         $trans[chr(138)] = '&Scaron;';    // Latin Capital Letter S With Caron
 143                         $trans[chr(139)] = '&lsaquo;';    // Single Left-Pointing Angle Quotation Mark
 144                         $trans[chr(140)] = '&OElig;';    // Latin Capital Ligature OE
 145                         $trans[chr(145)] = '&lsquo;';    // Left Single Quotation Mark
 146                         $trans[chr(146)] = '&rsquo;';    // Right Single Quotation Mark
 147                         $trans[chr(147)] = '&ldquo;';    // Left Double Quotation Mark
 148                         $trans[chr(148)] = '&rdquo;';    // Right Double Quotation Mark
 149                         $trans[chr(149)] = '&bull;';    // Bullet
 150                         $trans[chr(150)] = '&ndash;';    // En Dash
 151                         $trans[chr(151)] = '&mdash;';    // Em Dash
 152                         $trans[chr(152)] = '&tilde;';    // Small Tilde
 153                         $trans[chr(153)] = '&trade;';    // Trade Mark Sign
 154                         $trans[chr(154)] = '&scaron;';    // Latin Small Letter S With Caron
 155                         $trans[chr(155)] = '&rsaquo;';    // Single Right-Pointing Angle Quotation Mark
 156                         $trans[chr(156)] = '&oelig;';    // Latin Small Ligature OE
 157                         $trans[chr(159)] = '&Yuml;';    // Latin Capital Letter Y With Diaeresis
 158                         $html = strtr($html, $trans);
 159                 }
 160                 if (!$encoding) {
 161                         debug('No character encoding found, so treating as UTF-8');
 162                         $encoding = 'utf-8';
 163                 } else {
 164                         debug('Character encoding: '.$encoding);
 165                         if (strtolower($encoding) != 'utf-8') {
 166                                 debug('Converting to UTF-8');
 167                                 $html = SimplePie_Misc::change_encoding($html, $encoding, 'utf-8');
 168                                 /*
 169                                 if (function_exists('iconv')) {
 170                                         // iconv appears to handle certain character encodings better than mb_convert_encoding
 171                                         $html = iconv($encoding, 'utf-8', $html);
 172                                 } else {
 173                                         $html = mb_convert_encoding($html, 'utf-8', $encoding);
 174                                 }
 175                                 */
 176                         }
 177                 }
 178         }
 179         return $html;
 180 }
 181
 182 function makeAbsolute($base, $elem) {
 183         $base = new SimplePie_IRI($base);
 184         // remove '//' in URL path (used to prevent URLs from resolving properly)
 185         // TODO: check if this is still the case
 186         if (isset($base->path)) $base->path = preg_replace('!//+!', '/', $base->path);
 187         foreach(array('a'=>'href', 'img'=>'src') as $tag => $attr) {
 188                 $elems = $elem->getElementsByTagName($tag);
 189                 for ($i = $elems->length-1; $i >= 0; $i--) {
 190                         $e = $elems->item($i);
 191                         //$e->parentNode->replaceChild($articleContent->ownerDocument->createTextNode($e->textContent), $e);
 192                         makeAbsoluteAttr($base, $e, $attr);
 193                 }
 194                 if (strtolower($elem->tagName) == $tag) makeAbsoluteAttr($base, $elem, $attr);
 195         }
 196 }
 197 function makeAbsoluteAttr($base, $e, $attr) {
 198         if ($e->hasAttribute($attr)) {
 199                 // Trim leading and trailing white space. I don't really like this but
 200                 // unfortunately it does appear on some sites. e.g.  <img src=" /path/to/image.jpg" />
 201                 $url = trim(str_replace('%20', ' ', $e->getAttribute($attr)));
 202                 $url = str_replace(' ', '%20', $url);
 203                 if (!preg_match('!https?://!i', $url)) {
 204                         if ($absolute = SimplePie_IRI::absolutize($base, $url)) {
 205                                 $e->setAttribute($attr, $absolute);
 206                         }
 207                 }
 208         }
 209 }
 210 function makeAbsoluteStr($base, $url) {
 211         $base = new SimplePie_IRI($base);
 212         // remove '//' in URL path (causes URLs not to resolve properly)
 213         if (isset($base->path)) $base->path = preg_replace('!//+!', '/', $base->path);
 214         if (preg_match('!^https?://!i', $url)) {
 215                 // already absolute
 216                 return $url;
 217         } else {
 218                 if ($absolute = SimplePie_IRI::absolutize($base, $url)) {
 219                         return $absolute;
 220                 }
 221                 return false;
 222         }
 223 }
 224 // returns single page response, or false if not found
 225 function getSinglePage($item, $html, $url) {
 226         global $http, $extractor;
 227         debug('Looking for site config files to see if single page link exists');
 228         $site_config = $extractor->buildSiteConfig($url, $html);
 229         $splink = null;
 230         if (!empty($site_config->single_page_link)) {
 231                 $splink = $site_config->single_page_link;
 232         } elseif (!empty($site_config->single_page_link_in_feed)) {
 233                 // single page link xpath is targeted at feed
 234                 $splink = $site_config->single_page_link_in_feed;
 235                 // so let's replace HTML with feed item description
 236                 $html = $item->get_description();
 237         }
 238         if (isset($splink)) {
 239                 // Build DOM tree from HTML
 240                 $readability = new Readability($html, $url);
 241                 $xpath = new DOMXPath($readability->dom);
 242                 // Loop through single_page_link xpath expressions
 243                 $single_page_url = null;
 244                 foreach ($splink as $pattern) {
 245                         $elems = @$xpath->evaluate($pattern, $readability->dom);
 246                         if (is_string($elems)) {
 247                                 $single_page_url = trim($elems);
 248                                 break;
 249                         } elseif ($elems instanceof DOMNodeList && $elems->length > 0) {
 250                                 foreach ($elems as $item) {
 251                                         if ($item instanceof DOMElement && $item->hasAttribute('href')) {
 252                                                 $single_page_url = $item->getAttribute('href');
 253                                                 break 2;
 254                                         } elseif ($item instanceof DOMAttr && $item->value) {
 255                                                 $single_page_url = $item->value;
 256                                                 break 2;
 257                                         }
 258                                 }
 259                         }
 260                 }
 261                 // If we've got URL, resolve against $url
 262                 if (isset($single_page_url) && ($single_page_url = makeAbsoluteStr($url, $single_page_url))) {
 263                         // check it's not what we have already!
 264                         if ($single_page_url != $url) {
 265                                 // it's not, so let's try to fetch it...
 266                                 $_prev_ref = $http->referer;
 267                                 $http->referer = $single_page_url;
 268                                 if (($response = $http->get($single_page_url, true)) && $response['status_code'] < 300) {
 269                                         $http->referer = $_prev_ref;
 270                                         return $response;
 271                                 }
 272                                 $http->referer = $_prev_ref;
 273                         }
 274                 }
 275         }
 276         return false;
 277 }
 278
 279 // based on content-type http header, decide what to do
 280 // param: HTTP headers string
 281 // return: array with keys: 'mime', 'type', 'subtype', 'action', 'name'
 282 // e.g. array('mime'=>'image/jpeg', 'type'=>'image', 'subtype'=>'jpeg', 'action'=>'link', 'name'=>'Image')
 283 function get_mime_action_info($headers) {
 284         global $options;
 285         // check if action defined for returned Content-Type
 286         $info = array();
 287         if (preg_match('!^Content-Type:\s*(([-\w]+)/([-\w\+]+))!im', $headers, $match)) {
 288                 // look for full mime type (e.g. image/jpeg) or just type (e.g. image)
 289                 // match[1] = full mime type, e.g. image/jpeg
 290                 // match[2] = first part, e.g. image
 291                 // match[3] = last part, e.g. jpeg
 292                 $info['mime'] = strtolower(trim($match[1]));
 293                 $info['type'] = strtolower(trim($match[2]));
 294                 $info['subtype'] = strtolower(trim($match[3]));
 295                 foreach (array($info['mime'], $info['type']) as $_mime) {
 296                         if (isset($options->content_type_exc[$_mime])) {
 297                                 $info['action'] = $options->content_type_exc[$_mime]['action'];
 298                                 $info['name'] = $options->content_type_exc[$_mime]['name'];
 299                                 break;
 300                         }
 301                 }
 302         }
 303         return $info;
 304 }
 305
 306 function remove_url_cruft($url) {
 307         // remove google analytics for the time being
 308         // regex adapted from http://navitronic.co.uk/2010/12/removing-google-analytics-cruft-from-urls/
 309         // https://gist.github.com/758177
 310         return preg_replace('/(\?|\&)utm_[a-z]+=[^\&]+/', '', $url);
 311 }
 312
 313 function make_substitutions($string) {
 314         if ($string == '') return $string;
 315         global $item, $effective_url;
 316         $string = str_replace('{url}', htmlspecialchars($item->get_permalink()), $string);
 317         $string = str_replace('{effective-url}', htmlspecialchars($effective_url), $string);
 318         return $string;
 319 }
 320
 321 function get_cache() {
 322         global $options, $valid_key;
 323         static $cache = null;
 324         if ($cache === null) {
 325                 $frontendOptions = array(
 326                         'lifetime' => 10*60, // cache lifetime of 10 minutes
 327                         'automatic_serialization' => false,
 328                         'write_control' => false,
 329                         'automatic_cleaning_factor' => $options->cache_cleanup,
 330                         'ignore_user_abort' => false
 331                 );
 332                 $backendOptions = array(
 333                         'cache_dir' => ($valid_key) ? $options->cache_dir.'/rss-with-key/' : $options->cache_dir.'/rss/', // directory where to put the cache files
 334                         'file_locking' => false,
 335                         'read_control' => true,
 336                         'read_control_type' => 'strlen',
 337                         'hashed_directory_level' => $options->cache_directory_level,
 338                         'hashed_directory_perm' => 0777,
 339                         'cache_file_perm' => 0664,
 340                         'file_name_prefix' => 'ff'
 341                 );
 342                 // getting a Zend_Cache_Core object
 343                 $cache = Zend_Cache::factory('Core', 'File', $frontendOptions, $backendOptions);
 344         }
 345         return $cache;
 346 }
 347
 348 function debug($msg) {
 349         global $debug_mode;
 350         if ($debug_mode) {
 351                 echo '* ',$msg,"\n";
 352                 ob_flush();
 353                 flush();
 354         }
 355 }