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, $prefer=null) { return null; }
  61         public function get_enclosures() { return null; }
  62         public function get_categories() { return null; }
  63 }
  64
  65 ///////////////////////////////
  66 // HELPER FUNCTIONS
  67 ///////////////////////////////
  68
  69 // Adapted from WordPress
  70 // http://core.trac.wordpress.org/browser/tags/3.5.1/wp-includes/formatting.php#L2173
  71 function get_excerpt($text, $num_words=55, $more=null) {
  72         if (null === $more) $more = '&hellip;';
  73         $text = strip_tags($text);
  74         //TODO: Check if word count is based on single characters (East Asian characters)
  75         /*
  76         if (1==2) {
  77         $text = trim(preg_replace("/[\n\r\t ]+/", ' ', $text), ' ');
  78         preg_match_all('/./u', $text, $words_array);
  79         $words_array = array_slice($words_array[0], 0, $num_words + 1);
  80         $sep = '';
  81         } else {
  82         $words_array = preg_split("/[\n\r\t ]+/", $text, $num_words + 1, PREG_SPLIT_NO_EMPTY);
  83         $sep = ' ';
  84         }
  85         */
  86         $words_array = preg_split("/[\n\r\t ]+/", $text, $num_words + 1, PREG_SPLIT_NO_EMPTY);
  87         $sep = ' ';
  88         if (count($words_array) > $num_words) {
  89                 array_pop($words_array);
  90                 $text = implode($sep, $words_array);
  91                 $text = $text.$more;
  92         } else {
  93                 $text = implode($sep, $words_array);
  94         }
  95         // trim whitespace at beginning or end of string
  96         // See: http://stackoverflow.com/questions/4166896/trim-unicode-whitespace-in-php-5-2
  97         $text = preg_replace('/^[\pZ\pC]+|[\pZ\pC]+$/u', '', $text);
  98         return $text;
  99 }
 100
 101 function url_allowed($url) {
 102         global $options;
 103         if (!empty($options->allowed_urls)) {
 104                 $allowed = false;
 105                 foreach ($options->allowed_urls as $allowurl) {
 106                         if (stristr($url, $allowurl) !== false) {
 107                                 $allowed = true;
 108                                 break;
 109                         }
 110                 }
 111                 if (!$allowed) return false;
 112         } else {
 113                 foreach ($options->blocked_urls as $blockurl) {
 114                         if (stristr($url, $blockurl) !== false) {
 115                                 return false;
 116                         }
 117                 }
 118         }
 119         return true;
 120 }
 121
 122 //////////////////////////////////////////////
 123 // Convert $html to UTF8
 124 // (uses HTTP headers and HTML to find encoding)
 125 // adapted from http://stackoverflow.com/questions/910793/php-detect-encoding-and-make-everything-utf-8
 126 //////////////////////////////////////////////
 127 function convert_to_utf8($html, $header=null)
 128 {
 129         $encoding = null;
 130         if ($html || $header) {
 131                 if (is_array($header)) $header = implode("\n", $header);
 132                 if (!$header || !preg_match_all('/^Content-Type:\s+([^;]+)(?:;\s*charset=["\']?([^;"\'\n]*))?/im', $header, $match, PREG_SET_ORDER)) {
 133                         // error parsing the response
 134                         debug('Could not find Content-Type header in HTTP response');
 135                 } else {
 136                         $match = end($match); // get last matched element (in case of redirects)
 137                         if (isset($match[2])) $encoding = trim($match[2], "\"' \r\n\0\x0B\t");
 138                 }
 139                 // TODO: check to see if encoding is supported (can we convert it?)
 140                 // If it's not, result will be empty string.
 141                 // For now we'll check for invalid encoding types returned by some sites, e.g. 'none'
 142                 // Problem URL: http://facta.co.jp/blog/archives/20111026001026.html
 143                 if (!$encoding || $encoding == 'none') {
 144                         // search for encoding in HTML - only look at the first 50000 characters
 145                         // 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
 146                         // TODO: improve this so it looks at smaller chunks first
 147                         $html_head = substr($html, 0, 50000);
 148                         if (preg_match('/^<\?xml\s+version=(?:"[^"]*"|\'[^\']*\')\s+encoding=("[^"]*"|\'[^\']*\')/s', $html_head, $match)) {
 149                                 $encoding = trim($match[1], '"\'');
 150                         } elseif (preg_match('/<meta\s+http-equiv=["\']?Content-Type["\']? content=["\'][^;]+;\s*charset=["\']?([^;"\'>]+)/i', $html_head, $match)) {
 151                                 $encoding = trim($match[1]);
 152                         } elseif (preg_match_all('/<meta\s+([^>]+)>/i', $html_head, $match)) {
 153                                 foreach ($match[1] as $_test) {
 154                                         if (preg_match('/charset=["\']?([^"\']+)/i', $_test, $_m)) {
 155                                                 $encoding = trim($_m[1]);
 156                                                 break;
 157                                         }
 158                                 }
 159                         }
 160                 }
 161                 if (isset($encoding)) $encoding = trim($encoding);
 162                 // trim is important here!
 163                 if (!$encoding || (strtolower($encoding) == 'iso-8859-1')) {
 164                         // replace MS Word smart qutoes
 165                         $trans = array();
 166                         $trans[chr(130)] = '&sbquo;';    // Single Low-9 Quotation Mark
 167                         $trans[chr(131)] = '&fnof;';    // Latin Small Letter F With Hook
 168                         $trans[chr(132)] = '&bdquo;';    // Double Low-9 Quotation Mark
 169                         $trans[chr(133)] = '&hellip;';    // Horizontal Ellipsis
 170                         $trans[chr(134)] = '&dagger;';    // Dagger
 171                         $trans[chr(135)] = '&Dagger;';    // Double Dagger
 172                         $trans[chr(136)] = '&circ;';    // Modifier Letter Circumflex Accent
 173                         $trans[chr(137)] = '&permil;';    // Per Mille Sign
 174                         $trans[chr(138)] = '&Scaron;';    // Latin Capital Letter S With Caron
 175                         $trans[chr(139)] = '&lsaquo;';    // Single Left-Pointing Angle Quotation Mark
 176                         $trans[chr(140)] = '&OElig;';    // Latin Capital Ligature OE
 177                         $trans[chr(145)] = '&lsquo;';    // Left Single Quotation Mark
 178                         $trans[chr(146)] = '&rsquo;';    // Right Single Quotation Mark
 179                         $trans[chr(147)] = '&ldquo;';    // Left Double Quotation Mark
 180                         $trans[chr(148)] = '&rdquo;';    // Right Double Quotation Mark
 181                         $trans[chr(149)] = '&bull;';    // Bullet
 182                         $trans[chr(150)] = '&ndash;';    // En Dash
 183                         $trans[chr(151)] = '&mdash;';    // Em Dash
 184                         $trans[chr(152)] = '&tilde;';    // Small Tilde
 185                         $trans[chr(153)] = '&trade;';    // Trade Mark Sign
 186                         $trans[chr(154)] = '&scaron;';    // Latin Small Letter S With Caron
 187                         $trans[chr(155)] = '&rsaquo;';    // Single Right-Pointing Angle Quotation Mark
 188                         $trans[chr(156)] = '&oelig;';    // Latin Small Ligature OE
 189                         $trans[chr(159)] = '&Yuml;';    // Latin Capital Letter Y With Diaeresis
 190                         $html = strtr($html, $trans);
 191                 }
 192                 if (!$encoding) {
 193                         debug('No character encoding found, so treating as UTF-8');
 194                         $encoding = 'utf-8';
 195                 } else {
 196                         debug('Character encoding: '.$encoding);
 197                         if (strtolower($encoding) != 'utf-8') {
 198                                 debug('Converting to UTF-8');
 199                                 $html = SimplePie_Misc::change_encoding($html, $encoding, 'utf-8');
 200                         }
 201                 }
 202         }
 203         return $html;
 204 }
 205
 206 function makeAbsolute($base, $elem) {
 207         $base = new SimplePie_IRI($base);
 208         // remove '//' in URL path (used to prevent URLs from resolving properly)
 209         // TODO: check if this is still the case
 210         if (isset($base->path)) $base->path = preg_replace('!//+!', '/', $base->path);
 211         foreach(array('a'=>'href', 'img'=>'src') as $tag => $attr) {
 212                 $elems = $elem->getElementsByTagName($tag);
 213                 for ($i = $elems->length-1; $i >= 0; $i--) {
 214                         $e = $elems->item($i);
 215                         //$e->parentNode->replaceChild($articleContent->ownerDocument->createTextNode($e->textContent), $e);
 216                         makeAbsoluteAttr($base, $e, $attr);
 217                 }
 218                 if (strtolower($elem->tagName) == $tag) makeAbsoluteAttr($base, $elem, $attr);
 219         }
 220 }
 221 function makeAbsoluteAttr($base, $e, $attr) {
 222         if ($e->hasAttribute($attr)) {
 223                 // Trim leading and trailing white space. I don't really like this but
 224                 // unfortunately it does appear on some sites. e.g.  <img src=" /path/to/image.jpg" />
 225                 $url = trim(str_replace('%20', ' ', $e->getAttribute($attr)));
 226                 $url = str_replace(' ', '%20', $url);
 227                 if (!preg_match('!https?://!i', $url)) {
 228                         if ($absolute = SimplePie_IRI::absolutize($base, $url)) {
 229                                 $e->setAttribute($attr, $absolute);
 230                         }
 231                 }
 232         }
 233 }
 234 function makeAbsoluteStr($base, $url) {
 235         $base = new SimplePie_IRI($base);
 236         // remove '//' in URL path (causes URLs not to resolve properly)
 237         if (isset($base->path)) $base->path = preg_replace('!//+!', '/', $base->path);
 238         if (preg_match('!^https?://!i', $url)) {
 239                 // already absolute
 240                 return $url;
 241         } else {
 242                 if ($absolute = SimplePie_IRI::absolutize($base, $url)) {
 243                         return $absolute;
 244                 }
 245                 return false;
 246         }
 247 }
 248 // returns single page response, or false if not found
 249 function getSinglePage($item, $html, $url) {
 250         global $http, $extractor;
 251         debug('Looking for site config files to see if single page link exists');
 252         $site_config = $extractor->buildSiteConfig($url, $html);
 253         $splink = null;
 254         if (!empty($site_config->single_page_link)) {
 255                 $splink = $site_config->single_page_link;
 256         } elseif (!empty($site_config->single_page_link_in_feed)) {
 257                 // single page link xpath is targeted at feed
 258                 $splink = $site_config->single_page_link_in_feed;
 259                 // so let's replace HTML with feed item description
 260                 $html = $item->get_description();
 261         }
 262         if (isset($splink)) {
 263                 // Build DOM tree from HTML
 264                 $readability = new Readability($html, $url);
 265                 $xpath = new DOMXPath($readability->dom);
 266                 // Loop through single_page_link xpath expressions
 267                 $single_page_url = null;
 268                 foreach ($splink as $pattern) {
 269                         $elems = @$xpath->evaluate($pattern, $readability->dom);
 270                         if (is_string($elems)) {
 271                                 $single_page_url = trim($elems);
 272                                 break;
 273                         } elseif ($elems instanceof DOMNodeList && $elems->length > 0) {
 274                                 foreach ($elems as $item) {
 275                                         if ($item instanceof DOMElement && $item->hasAttribute('href')) {
 276                                                 $single_page_url = $item->getAttribute('href');
 277                                                 break 2;
 278                                         } elseif ($item instanceof DOMAttr && $item->value) {
 279                                                 $single_page_url = $item->value;
 280                                                 break 2;
 281                                         }
 282                                 }
 283                         }
 284                 }
 285                 // If we've got URL, resolve against $url
 286                 if (isset($single_page_url) && ($single_page_url = makeAbsoluteStr($url, $single_page_url))) {
 287                         // check it's not what we have already!
 288                         if ($single_page_url != $url) {
 289                                 // it's not, so let's try to fetch it...
 290                                 $_prev_ref = $http->referer;
 291                                 $http->referer = $single_page_url;
 292                                 if (($response = $http->get($single_page_url, true)) && $response['status_code'] < 300) {
 293                                         $http->referer = $_prev_ref;
 294                                         return $response;
 295                                 }
 296                                 $http->referer = $_prev_ref;
 297                         }
 298                 }
 299         }
 300         return false;
 301 }
 302
 303 // based on content-type http header, decide what to do
 304 // param: HTTP headers string
 305 // return: array with keys: 'mime', 'type', 'subtype', 'action', 'name'
 306 // e.g. array('mime'=>'image/jpeg', 'type'=>'image', 'subtype'=>'jpeg', 'action'=>'link', 'name'=>'Image')
 307 function get_mime_action_info($headers) {
 308         global $options;
 309         // check if action defined for returned Content-Type
 310         $info = array();
 311         if (preg_match('!^Content-Type:\s*(([-\w]+)/([-\w\+]+))!im', $headers, $match)) {
 312                 // look for full mime type (e.g. image/jpeg) or just type (e.g. image)
 313                 // match[1] = full mime type, e.g. image/jpeg
 314                 // match[2] = first part, e.g. image
 315                 // match[3] = last part, e.g. jpeg
 316                 $info['mime'] = strtolower(trim($match[1]));
 317                 $info['type'] = strtolower(trim($match[2]));
 318                 $info['subtype'] = strtolower(trim($match[3]));
 319                 foreach (array($info['mime'], $info['type']) as $_mime) {
 320                         if (isset($options->content_type_exc[$_mime])) {
 321                                 $info['action'] = $options->content_type_exc[$_mime]['action'];
 322                                 $info['name'] = $options->content_type_exc[$_mime]['name'];
 323                                 break;
 324                         }
 325                 }
 326         }
 327         return $info;
 328 }
 329
 330 function remove_url_cruft($url) {
 331         // remove google analytics for the time being
 332         // regex adapted from http://navitronic.co.uk/2010/12/removing-google-analytics-cruft-from-urls/
 333         // https://gist.github.com/758177
 334         return preg_replace('/(\?|\&)utm_[a-z]+=[^\&]+/', '', $url);
 335 }
 336
 337 function make_substitutions($string) {
 338         if ($string == '') return $string;
 339         global $item, $effective_url;
 340         $string = str_replace('{url}', htmlspecialchars($item->get_permalink()), $string);
 341         $string = str_replace('{effective-url}', htmlspecialchars($effective_url), $string);
 342         return $string;
 343 }
 344
 345 function get_cache() {
 346         global $options, $valid_key;
 347         static $cache = null;
 348         if ($cache === null) {
 349                 $frontendOptions = array(
 350                         'lifetime' => 10*60, // cache lifetime of 10 minutes
 351                         'automatic_serialization' => false,
 352                         'write_control' => false,
 353                         'automatic_cleaning_factor' => $options->cache_cleanup,
 354                         'ignore_user_abort' => false
 355                 );
 356                 $backendOptions = array(
 357                         'cache_dir' => ($valid_key) ? $options->cache_dir.'/rss-with-key/' : $options->cache_dir.'/rss/', // directory where to put the cache files
 358                         'file_locking' => false,
 359                         'read_control' => true,
 360                         'read_control_type' => 'strlen',
 361                         'hashed_directory_level' => $options->cache_directory_level,
 362                         'hashed_directory_perm' => 0777,
 363                         'cache_file_perm' => 0664,
 364                         'file_name_prefix' => 'ff'
 365                 );
 366                 // getting a Zend_Cache_Core object
 367                 $cache = Zend_Cache::factory('Core', 'File', $frontendOptions, $backendOptions);
 368         }
 369         return $cache;
 370 }
 371
 372 function debug($msg) {
 373         global $debug_mode;
 374         if ($debug_mode) {
 375                 echo '* ',$msg,"\n";
 376                 ob_flush();
 377                 flush();
 378         }
 379 }
 380
 381 function get_base_url($dom) {
 382         $xpath = new DOMXPath($dom);
 383         $base_url = @$xpath->evaluate('string(//head/base/@href)', $dom);
 384         if ($base_url !== '') {
 385                 return $base_url;
 386         } else {
 387                 return false;
 388         }
 389 }