classes/feedparser.php

   1 <?php
   2 class FeedParser {
   3         private $doc;
   4         private $error;
   5         private $items;
   6         private $link;
   7         private $title;
   8         private $type;
   9         private $xpath;
  10
  11         const FEED_RDF = 0;
  12         const FEED_RSS = 1;
  13         const FEED_ATOM = 2;
  14
  15         function __construct($data) {
  16                 libxml_use_internal_errors(true);
  17                 libxml_clear_errors();
  18                 $this->doc = new DOMDocument();
  19                 $this->doc->loadXML($data);
  20                 $this->error = $this->format_error(libxml_get_last_error());
  21                 libxml_clear_errors();
  22
  23                 $this->items = array();
  24         }
  25
  26         function init() {
  27                 $root = $this->doc->firstChild;
  28                 $xpath = new DOMXPath($this->doc);
  29                 $xpath->registerNamespace('atom', 'http://www.w3.org/2005/Atom');
  30                 $xpath->registerNamespace('media', 'http://search.yahoo.com/mrss/');
  31                 $xpath->registerNamespace('rdf', 'http://www.w3.org/1999/02/22-rdf-syntax-ns#');
  32                 $xpath->registerNamespace('slash', 'http://purl.org/rss/1.0/modules/slash/');
  33                 $xpath->registerNamespace('dc', 'http://purl.org/dc/elements/1.1/');
  34                 $xpath->registerNamespace('content', 'http://purl.org/rss/1.0/modules/content/');
  35
  36                 $this->xpath = $xpath;
  37
  38                 $root = $xpath->query("(//atom:feed|//channel|//rdf:rdf|//rdf:RDF)")->item(0);
  39
  40                 if ($root) {
  41                         switch (mb_strtolower($root->tagName)) {
  42                         case "rdf:rdf":
  43                                 $this->type = $this::FEED_RDF;
  44                                 break;
  45                         case "channel":
  46                                 $this->type = $this::FEED_RSS;
  47                                 break;
  48                         case "feed":
  49                                 $this->type = $this::FEED_ATOM;
  50                                 break;
  51                         default:
  52                                 $this->error = "Unknown/unsupported feed type";
  53                                 return;
  54                         }
  55
  56                         switch ($this->type) {
  57                         case $this::FEED_ATOM:
  58
  59                                 $title = $xpath->query("//atom:feed/atom:title")->item(0);
  60
  61                                 if ($title) {
  62                                         $this->title = $title->nodeValue;
  63                                 }
  64
  65                                 $link = $xpath->query("//atom:feed/atom:link[not(@rel)]")->item(0);
  66
  67                                 if ($link && $link->hasAttributes()) {
  68                                         $this->link = $link->getAttribute("href");
  69                                 }
  70
  71                                 $articles = $xpath->query("//atom:entry");
  72
  73                                 foreach ($articles as $article) {
  74                                         array_push($this->items, new FeedItem_Atom($article, $this->doc, $this->xpath));
  75                                 }
  76
  77                                 break;
  78                         case $this::FEED_RSS:
  79
  80                                 $title = $xpath->query("//channel/title")->item(0);
  81
  82                                 if ($title) {
  83                                         $this->title = $title->nodeValue;
  84                                 }
  85
  86                                 $link = $xpath->query("//channel/link")->item(0);
  87
  88                                 if ($link && $link->hasAttributes()) {
  89                                         $this->link = $link->getAttribute("href");
  90                                 }
  91
  92                                 $articles = $xpath->query("//channel/item");
  93
  94                                 foreach ($articles as $article) {
  95                                         array_push($this->items, new FeedItem_RSS($article, $this->doc, $this->xpath));
  96                                 }
  97
  98                                 break;
  99                         case $this::FEED_RDF:
 100                                 $xpath->registerNamespace('rssfake', 'http://purl.org/rss/1.0/');
 101
 102                                 $title = $xpath->query("//rssfake:channel/rssfake:title")->item(0);
 103
 104                                 if ($title) {
 105                                         $this->title = $title->nodeValue;
 106                                 }
 107
 108                                 $link = $xpath->query("//rssfake:channel/rssfake:link")->item(0);
 109
 110                                 if ($link) {
 111                                         $this->link = $link->nodeValue;
 112                                 }
 113
 114                                 $articles = $xpath->query("//rssfake:item");
 115
 116                                 foreach ($articles as $article) {
 117                                         array_push($this->items, new FeedItem_RSS($article, $this->doc, $this->xpath));
 118                                 }
 119
 120                                 break;
 121
 122                         }
 123                 } else {
 124                         $this->error = "Unknown/unsupported feed type";
 125                         return;
 126                 }
 127         }
 128
 129         function format_error($error) {
 130                 if ($error) {
 131                         return sprintf("LibXML error %s at line %d (column %d): %s",
 132                                 $error->code, $error->line, $error->column,
 133                                 $error->message);
 134                 } else {
 135                         return "";
 136                 }
 137         }
 138
 139         function error() {
 140                 return $this->error;
 141         }
 142
 143         function get_link() {
 144                 return $this->link;
 145         }
 146
 147         function get_title() {
 148                 return $this->title;
 149         }
 150
 151         function get_items() {
 152                 return $this->items;
 153         }
 154
 155         function get_links($rel) {
 156                 $rv = array();
 157
 158                 switch ($this->type) {
 159                 case $this::FEED_ATOM:
 160                         $links = $this->xpath->query("//atom:feed/atom:link");
 161
 162                         foreach ($links as $link) {
 163                                 if (!$rel || $link->hasAttribute('rel') && $link->getAttribute('rel') == $rel) {
 164                                         array_push($rv, $link->getAttribute('href'));
 165                                 }
 166                         }
 167                         break;
 168                 case $this::FEED_RSS:
 169                         $links = $this->xpath->query("//channel/link");
 170                         foreach ($links as $link) {
 171                                 if (!$rel || $link->hasAttribute('rel') && $link->getAttribute('rel') == $rel) {
 172                                         array_push($rv, $link->getAttribute('href'));
 173                                 }
 174                         }
 175                         break;
 176                 }
 177
 178                 return $rv;
 179         }
 180 } ?>