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 $mapping = array(
   9                 // Include FeedCreator for RSS/Atom creation
  10                 'FeedWriter' => 'feedwriter/FeedWriter.php',
  11                 'FeedItem' => 'feedwriter/FeedItem.php',
  12                 // Include ContentExtractor and Readability for identifying and extracting content from URLs
  13                 'ContentExtractor' => 'content-extractor/ContentExtractor.php',
  14                 'SiteConfig' => 'content-extractor/SiteConfig.php',
  15                 // Include Humble HTTP Agent to allow parallel requests and response caching
  16                 'HumbleHttpAgent' => 'humble-http-agent/HumbleHttpAgent.php',
  17                 'SimplePie_HumbleHttpAgent' => 'humble-http-agent/SimplePie_HumbleHttpAgent.php',
  18                 'CookieJar' => 'humble-http-agent/CookieJar.php',
  19                 // Include Zend Cache to improve performance (cache results)
  20                 'Zend_Cache' => 'Zend/Cache.php',
  21                 // Language detect
  22                 'Text_LanguageDetect' => 'language-detect/LanguageDetect.php',
  23                 // HTML5 Lib
  24                 'HTML5_Parser' => 'html5/Parser.php',
  25                 // htmLawed - used if XSS filter is enabled (xss_filter)
  26                 'htmLawed' => 'htmLawed/htmLawed.php'
  27         );
  28         if (isset($mapping[$class_name])) {
  29                 debug("** Loading class $class_name ({$mapping[$class_name]})");
  30                 require $dir.$mapping[$class_name];
  31                 return true;
  32         } else {
  33                 return false;
  34         }
  35 }
  36 spl_autoload_register('autoload');
  37
  38 class DummySingleItemFeed {
  39         public $item;
  40         function __construct($url) { $this->item = new DummySingleItem($url); }
  41         public function get_title() { return ''; }
  42         public function get_description() { return 'Content extracted from '.$this->item->url; }
  43         public function get_link() { return $this->item->url; }
  44         public function get_language() { return false; }
  45         public function get_image_url() { return false; }
  46         public function get_items($start=0, $max=1) { return array(0=>$this->item); }
  47 }
  48 class DummySingleItem {
  49         public $url;
  50         function __construct($url) { $this->url = $url; }
  51         public function get_permalink() { return $this->url; }
  52         public function get_title() { return null; }
  53         public function get_date($format='') { return false; }
  54         public function get_author($key=0) { return null; }
  55         public function get_authors() { return null; }
  56         public function get_description() { return ''; }
  57         public function get_enclosure($key=0, $prefer=null) { return null; }
  58         public function get_enclosures() { return null; }
  59         public function get_categories() { return null; }
  60 }
  61
  62 ///////////////////////////////
  63 // HELPER FUNCTIONS
  64 ///////////////////////////////
  65
  66 // Adapted from WordPress
  67 // http://core.trac.wordpress.org/browser/tags/3.5.1/wp-includes/formatting.php#L2173
  68 function get_excerpt($text, $num_words=55, $more=null) {
  69         if (null === $more) $more = '&hellip;';
  70         $text = strip_tags($text);
  71         //TODO: Check if word count is based on single characters (East Asian characters)
  72         /*
  73         if (1==2) {
  74         $text = trim(preg_replace("/[\n\r\t ]+/", ' ', $text), ' ');
  75         preg_match_all('/./u', $text, $words_array);
  76         $words_array = array_slice($words_array[0], 0, $num_words + 1);
  77         $sep = '';
  78         } else {
  79         $words_array = preg_split("/[\n\r\t ]+/", $text, $num_words + 1, PREG_SPLIT_NO_EMPTY);
  80         $sep = ' ';
  81         }
  82         */
  83         $words_array = preg_split("/[\n\r\t ]+/", $text, $num_words + 1, PREG_SPLIT_NO_EMPTY);
  84         $sep = ' ';
  85         if (count($words_array) > $num_words) {
  86                 array_pop($words_array);
  87                 $text = implode($sep, $words_array);
  88                 $text = $text.$more;
  89         } else {
  90                 $text = implode($sep, $words_array);
  91         }
  92         // trim whitespace at beginning or end of string
  93         // See: http://stackoverflow.com/questions/4166896/trim-unicode-whitespace-in-php-5-2
  94         $text = preg_replace('/^[\pZ\pC]+|[\pZ\pC]+$/u', '', $text);
  95         return $text;
  96 }
  97
  98 function url_allowed($url) {
  99         global $options;
 100         if (!empty($options->allowed_urls)) {
 101                 $allowed = false;
 102                 foreach ($options->allowed_urls as $allowurl) {
 103                         if (stristr($url, $allowurl) !== false) {
 104                                 $allowed = true;
 105                                 break;
 106                         }
 107                 }
 108                 if (!$allowed) return false;
 109         } else {
 110                 foreach ($options->blocked_urls as $blockurl) {
 111                         if (stristr($url, $blockurl) !== false) {
 112                                 return false;
 113                         }
 114                 }
 115         }
 116         return true;
 117 }
 118
 119 //////////////////////////////////////////////
 120 // Convert $html to UTF8
 121 // (uses HTTP headers and HTML to find encoding)
 122 // adapted from http://stackoverflow.com/questions/910793/php-detect-encoding-and-make-everything-utf-8
 123 //////////////////////////////////////////////
 124 function convert_to_utf8($html, $header=null)
 125 {
 126         $encoding = null;
 127         if ($html || $header) {
 128                 if (is_array($header)) $header = implode("\n", $header);
 129                 if (!$header || !preg_match_all('/^Content-Type:\s+([^;]+)(?:;\s*charset=["\']?([^;"\'\n]*))?/im', $header, $match, PREG_SET_ORDER)) {
 130                         // error parsing the response
 131                         debug('Could not find Content-Type header in HTTP response');
 132                 } else {
 133                         $match = end($match); // get last matched element (in case of redirects)
 134                         if (isset($match[2])) $encoding = trim($match[2], "\"' \r\n\0\x0B\t");
 135                 }
 136                 // TODO: check to see if encoding is supported (can we convert it?)
 137                 // If it's not, result will be empty string.
 138                 // For now we'll check for invalid encoding types returned by some sites, e.g. 'none'
 139                 // Problem URL: http://facta.co.jp/blog/archives/20111026001026.html
 140                 if (!$encoding || $encoding == 'none') {
 141                         // search for encoding in HTML - only look at the first 50000 characters
 142                         // 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
 143                         // TODO: improve this so it looks at smaller chunks first
 144                         $html_head = substr($html, 0, 50000);
 145                         if (preg_match('/^<\?xml\s+version=(?:"[^"]*"|\'[^\']*\')\s+encoding=("[^"]*"|\'[^\']*\')/s', $html_head, $match)) {
 146                                 $encoding = trim($match[1], '"\'');
 147                         } elseif (preg_match('/<meta\s+http-equiv=["\']?Content-Type["\']? content=["\'][^;]+;\s*charset=["\']?([^;"\'>]+)/i', $html_head, $match)) {
 148                                 $encoding = trim($match[1]);
 149                         } elseif (preg_match_all('/<meta\s+([^>]+)>/i', $html_head, $match)) {
 150                                 foreach ($match[1] as $_test) {
 151                                         if (preg_match('/charset=["\']?([^"\']+)/i', $_test, $_m)) {
 152                                                 $encoding = trim($_m[1]);
 153                                                 break;
 154                                         }
 155                                 }
 156                         }
 157                 }
 158                 if (isset($encoding)) $encoding = trim($encoding);
 159                 // trim is important here!
 160                 if (!$encoding || (strtolower($encoding) == 'iso-8859-1')) {
 161                         // replace MS Word smart qutoes
 162                         $trans = array();
 163                         $trans[chr(130)] = '&sbquo;';    // Single Low-9 Quotation Mark
 164                         $trans[chr(131)] = '&fnof;';    // Latin Small Letter F With Hook
 165                         $trans[chr(132)] = '&bdquo;';    // Double Low-9 Quotation Mark
 166                         $trans[chr(133)] = '&hellip;';    // Horizontal Ellipsis
 167                         $trans[chr(134)] = '&dagger;';    // Dagger
 168                         $trans[chr(135)] = '&Dagger;';    // Double Dagger
 169                         $trans[chr(136)] = '&circ;';    // Modifier Letter Circumflex Accent
 170                         $trans[chr(137)] = '&permil;';    // Per Mille Sign
 171                         $trans[chr(138)] = '&Scaron;';    // Latin Capital Letter S With Caron
 172                         $trans[chr(139)] = '&lsaquo;';    // Single Left-Pointing Angle Quotation Mark
 173                         $trans[chr(140)] = '&OElig;';    // Latin Capital Ligature OE
 174                         $trans[chr(145)] = '&lsquo;';    // Left Single Quotation Mark
 175                         $trans[chr(146)] = '&rsquo;';    // Right Single Quotation Mark
 176                         $trans[chr(147)] = '&ldquo;';    // Left Double Quotation Mark
 177                         $trans[chr(148)] = '&rdquo;';    // Right Double Quotation Mark
 178                         $trans[chr(149)] = '&bull;';    // Bullet
 179                         $trans[chr(150)] = '&ndash;';    // En Dash
 180                         $trans[chr(151)] = '&mdash;';    // Em Dash
 181                         $trans[chr(152)] = '&tilde;';    // Small Tilde
 182                         $trans[chr(153)] = '&trade;';    // Trade Mark Sign
 183                         $trans[chr(154)] = '&scaron;';    // Latin Small Letter S With Caron
 184                         $trans[chr(155)] = '&rsaquo;';    // Single Right-Pointing Angle Quotation Mark
 185                         $trans[chr(156)] = '&oelig;';    // Latin Small Ligature OE
 186                         $trans[chr(159)] = '&Yuml;';    // Latin Capital Letter Y With Diaeresis
 187                         $html = strtr($html, $trans);
 188                 }
 189                 if (!$encoding) {
 190                         debug('No character encoding found, so treating as UTF-8');
 191                         $encoding = 'utf-8';
 192                 } else {
 193                         debug('Character encoding: '.$encoding);
 194                         if (strtolower($encoding) != 'utf-8') {
 195                                 debug('Converting to UTF-8');
 196                                 $html = SimplePie_Misc::change_encoding($html, $encoding, 'utf-8');
 197                         }
 198                 }
 199         }
 200         return $html;
 201 }
 202
 203 function makeAbsolute($base, $elem) {
 204         $base = new SimplePie_IRI($base);
 205         // remove '//' in URL path (used to prevent URLs from resolving properly)
 206         // TODO: check if this is still the case
 207         if (isset($base->path)) $base->path = preg_replace('!//+!', '/', $base->path);
 208         foreach(array('a'=>'href', 'img'=>'src') as $tag => $attr) {
 209                 $elems = $elem->getElementsByTagName($tag);
 210                 for ($i = $elems->length-1; $i >= 0; $i--) {
 211                         $e = $elems->item($i);
 212                         //$e->parentNode->replaceChild($articleContent->ownerDocument->createTextNode($e->textContent), $e);
 213                         makeAbsoluteAttr($base, $e, $attr);
 214                 }
 215                 if (strtolower($elem->tagName) == $tag) makeAbsoluteAttr($base, $elem, $attr);
 216         }
 217 }
 218 function makeAbsoluteAttr($base, $e, $attr) {
 219         if ($e->hasAttribute($attr)) {
 220                 // Trim leading and trailing white space. I don't really like this but
 221                 // unfortunately it does appear on some sites. e.g.  <img src=" /path/to/image.jpg" />
 222                 $url = trim(str_replace('%20', ' ', $e->getAttribute($attr)));
 223                 $url = str_replace(' ', '%20', $url);
 224                 if (!preg_match('!https?://!i', $url)) {
 225                         if ($absolute = SimplePie_IRI::absolutize($base, $url)) {
 226                                 $e->setAttribute($attr, $absolute);
 227                         }
 228                 }
 229         }
 230 }
 231 function makeAbsoluteStr($base, $url) {
 232         $base = new SimplePie_IRI($base);
 233         // remove '//' in URL path (causes URLs not to resolve properly)
 234         if (isset($base->path)) $base->path = preg_replace('!//+!', '/', $base->path);
 235         if (preg_match('!^https?://!i', $url)) {
 236                 // already absolute
 237                 return $url;
 238         } else {
 239                 if ($absolute = SimplePie_IRI::absolutize($base, $url)) {
 240                         return $absolute;
 241                 }
 242                 return false;
 243         }
 244 }
 245 // returns single page response, or false if not found
 246 function getSinglePage($item, $html, $url) {
 247         global $http, $extractor;
 248         debug('Looking for site config files to see if single page link exists');
 249         $site_config = $extractor->buildSiteConfig($url, $html);
 250         $splink = null;
 251         if (!empty($site_config->single_page_link)) {
 252                 $splink = $site_config->single_page_link;
 253         } elseif (!empty($site_config->single_page_link_in_feed)) {
 254                 // single page link xpath is targeted at feed
 255                 $splink = $site_config->single_page_link_in_feed;
 256                 // so let's replace HTML with feed item description
 257                 $html = $item->get_description();
 258         }
 259         if (isset($splink)) {
 260                 // Build DOM tree from HTML
 261                 $readability = new Readability($html, $url);
 262                 $xpath = new DOMXPath($readability->dom);
 263                 // Loop through single_page_link xpath expressions
 264                 $single_page_url = null;
 265                 foreach ($splink as $pattern) {
 266                         $elems = @$xpath->evaluate($pattern, $readability->dom);
 267                         if (is_string($elems)) {
 268                                 $single_page_url = trim($elems);
 269                                 break;
 270                         } elseif ($elems instanceof DOMNodeList && $elems->length > 0) {
 271                                 foreach ($elems as $item) {
 272                                         if ($item instanceof DOMElement && $item->hasAttribute('href')) {
 273                                                 $single_page_url = $item->getAttribute('href');
 274                                                 break 2;
 275                                         } elseif ($item instanceof DOMAttr && $item->value) {
 276                                                 $single_page_url = $item->value;
 277                                                 break 2;
 278                                         }
 279                                 }
 280                         }
 281                 }
 282                 // If we've got URL, resolve against $url
 283                 if (isset($single_page_url) && ($single_page_url = makeAbsoluteStr($url, $single_page_url))) {
 284                         // check it's not what we have already!
 285                         if ($single_page_url != $url) {
 286                                 // it's not, so let's try to fetch it...
 287                                 $_prev_ref = $http->referer;
 288                                 $http->referer = $single_page_url;
 289                                 if (($response = $http->get($single_page_url, true)) && $response['status_code'] < 300) {
 290                                         $http->referer = $_prev_ref;
 291                                         return $response;
 292                                 }
 293                                 $http->referer = $_prev_ref;
 294                         }
 295                 }
 296         }
 297         return false;
 298 }
 299
 300 // based on content-type http header, decide what to do
 301 // param: HTTP headers string
 302 // return: array with keys: 'mime', 'type', 'subtype', 'action', 'name'
 303 // e.g. array('mime'=>'image/jpeg', 'type'=>'image', 'subtype'=>'jpeg', 'action'=>'link', 'name'=>'Image')
 304 function get_mime_action_info($headers) {
 305         global $options;
 306         // check if action defined for returned Content-Type
 307         $info = array();
 308         if (preg_match('!^Content-Type:\s*(([-\w]+)/([-\w\+]+))!im', $headers, $match)) {
 309                 // look for full mime type (e.g. image/jpeg) or just type (e.g. image)
 310                 // match[1] = full mime type, e.g. image/jpeg
 311                 // match[2] = first part, e.g. image
 312                 // match[3] = last part, e.g. jpeg
 313                 $info['mime'] = strtolower(trim($match[1]));
 314                 $info['type'] = strtolower(trim($match[2]));
 315                 $info['subtype'] = strtolower(trim($match[3]));
 316                 foreach (array($info['mime'], $info['type']) as $_mime) {
 317                         if (isset($options->content_type_exc[$_mime])) {
 318                                 $info['action'] = $options->content_type_exc[$_mime]['action'];
 319                                 $info['name'] = $options->content_type_exc[$_mime]['name'];
 320                                 break;
 321                         }
 322                 }
 323         }
 324         return $info;
 325 }
 326
 327 function remove_url_cruft($url) {
 328         // remove google analytics for the time being
 329         // regex adapted from http://navitronic.co.uk/2010/12/removing-google-analytics-cruft-from-urls/
 330         // https://gist.github.com/758177
 331         return preg_replace('/(\?|\&)utm_[a-z]+=[^\&]+/', '', $url);
 332 }
 333
 334 function make_substitutions($string) {
 335         if ($string == '') return $string;
 336         global $item, $effective_url;
 337         $string = str_replace('{url}', htmlspecialchars($item->get_permalink()), $string);
 338         $string = str_replace('{effective-url}', htmlspecialchars($effective_url), $string);
 339         return $string;
 340 }
 341
 342 function get_cache() {
 343         global $options, $valid_key;
 344         static $cache = null;
 345         if ($cache === null) {
 346                 $frontendOptions = array(
 347                         'lifetime' => 10*60, // cache lifetime of 10 minutes
 348                         'automatic_serialization' => false,
 349                         'write_control' => false,
 350                         'automatic_cleaning_factor' => $options->cache_cleanup,
 351                         'ignore_user_abort' => false
 352                 );
 353                 $backendOptions = array(
 354                         'cache_dir' => ($valid_key) ? $options->cache_dir.'/rss-with-key/' : $options->cache_dir.'/rss/', // directory where to put the cache files
 355                         'file_locking' => false,
 356                         'read_control' => true,
 357                         'read_control_type' => 'strlen',
 358                         'hashed_directory_level' => $options->cache_directory_level,
 359                         'hashed_directory_perm' => 0777,
 360                         'cache_file_perm' => 0664,
 361                         'file_name_prefix' => 'ff'
 362                 );
 363                 // getting a Zend_Cache_Core object
 364                 $cache = Zend_Cache::factory('Core', 'File', $frontendOptions, $backendOptions);
 365         }
 366         return $cache;
 367 }
 368
 369 function debug($msg) {
 370         global $debug_mode;
 371         if ($debug_mode) {
 372                 echo '* ',$msg,"\n";
 373                 ob_flush();
 374                 flush();
 375         }
 376 }
 377
 378 function get_base_url($dom) {
 379         $xpath = new DOMXPath($dom);
 380         $base_url = @$xpath->evaluate('string(//head/base/@href)', $dom);
 381         if ($base_url !== '') {
 382                 return $base_url;
 383         } else {
 384                 return false;
 385         }
 386 }