Magpie with feedparser
Summit - September 19, 2007 - 07:42
| Project: | Feedparser |
| Version: | 5.x-1.x-dev |
| Component: | feedaggregator_node |
| Category: | feature request |
| Priority: | normal |
| Assigned: | Unassigned |
| Status: | active |
Description
Hi,
I would like to use Magpie instead of Simplepie, because I would like to use productfeeds which are not rss 2.0, atom or other newsfeeds.
I found this code that can help may be:
<?php
/*
* Project: XML for MagpieRSS: a simple RSS integration tool
* File: xml_fetch.inc, a simple functional interface
to fetching and parsing XML
* Author: Bram Brambring <bram@brambring.nl>
* License: GPL
* Script requires an installation of MagpieRSS
* Tested with magpie 1.72
* The lastest version of MagpieRSS can be obtained from:
* <a href="http://magpierss.sourceforge.net
" title="http://magpierss.sourceforge.net
" rel="nofollow">http://magpierss.sourceforge.net
</a> *
*/
require_once( MAGPIE_DIR . 'rss_fetch.inc' );
require_once( MAGPIE_DIR . 'rss_cache.inc' );
/*=======================================================================*\
Function: xml_fetch:
Purpose: returns object for the give url
maintain the cache
Input: url of RSS file
XML class
cache age
Output: parsed RSS object (see rss_parse.inc)
NOTES:
see rss_fetch.inc
\*=======================================================================*/
function xml_fetch ($url,$class='MagpieRSS',$cache_age=MAGPIE_CACHE_AGE) {
// initialize constants
init();
if ( !isset($url) ) {
error("fetch_rss called without a url");
return false;
}
// if cache is disabled
if ( !MAGPIE_CACHE_ON ) {
// fetch file, and parse it
$resp = _fetch_remote_file( $url );
if ( is_success( $resp->status ) ) {
return _response( $resp,$class );
}
else {
error("Failed to fetch $url and cache is off");
return false;
}
}
// else cache is ON
else {
// Flow
// 1. check cache
// 2. if there is a hit, make sure its fresh
// 3. if cached obj fails freshness check, fetch remote
// 4. if remote fails, return stale object, or error
$cache = new RSSCache( MAGPIE_CACHE_DIR, $cache_age );
if (MAGPIE_DEBUG and $cache->ERROR) {
debug($cache->ERROR, E_USER_WARNING);
}
$cache_status = 0; // response of check_cache
$request_headers = array(); // HTTP headers to send with fetch
$rss = 0; // parsed RSS object
$errormsg = 0; // errors, if any
// store parsed XML by desired output encoding
// as character munging happens at parse time
$cache_key = $url . MAGPIE_OUTPUT_ENCODING;
if (!$cache->ERROR) {
// return cache HIT, MISS, or STALE
$cache_status = $cache->check_cache( $cache_key);
}
// if object cached, and cache is fresh, return cached obj
if ( $cache_status == 'HIT' ) {
$rss = $cache->get( $cache_key );
if ( isset($rss) and $rss ) {
// should be cache age
$rss->from_cache = 1;
if ( MAGPIE_DEBUG > 1) {
debug("MagpieRSS: Cache HIT", E_USER_NOTICE);
}
return $rss;
}
}
// else attempt a conditional get
// setup headers
if ( $cache_status == 'STALE' ) {
$rss = $cache->get( $cache_key );
if ( $rss and $rss->etag and $rss->last_modified ) {
$request_headers['If-None-Match'] = $rss->etag;
$request_headers['If-Last-Modified'] = $rss->last_modified;
}
}
$resp = _fetch_remote_file( $url, $request_headers );
if (isset($resp) and $resp) {
if ($resp->status == '304' ) {
// we have the most current copy
if ( MAGPIE_DEBUG > 1) {
debug("Got 304 for $url");
}
// reset cache on 304 (at minutillo insistent prodding)
$cache->set($cache_key, $rss);
return $rss;
}
elseif ( is_success( $resp->status ) ) {
$rss = _response( $resp,$class );
if ( $rss ) {
if (MAGPIE_DEBUG > 1) {
debug("Fetch successful");
}
// add object to cache
$cache->set( $cache_key, $rss );
return $rss;
}
}
else {
$errormsg = "Failed to fetch $url ";
if ( $resp->status == '-100' ) {
$errormsg .= "(Request timed out after " . MAGPIE_FETCH_TIME_OUT . " seconds)";
}
elseif ( $resp->error ) {
# compensate for Snoopy's annoying habbit to tacking
# on '\n'
$http_error = substr($resp->error, 0, -2);
$errormsg .= "(HTTP Error: $http_error)";
}
else {
$errormsg .= "(HTTP Response: " . $resp->response_code .')';
}
}
}
else {
$errormsg = "Unable to retrieve RSS file for unknown reasons.";
}
// else fetch failed
// attempt to return cached object
if ($rss) {
if ( MAGPIE_DEBUG ) {
debug("Returning STALE object for $url");
}
return $rss;
}
// else we totally failed
error( $errormsg );
return false;
} // end if ( !MAGPIE_CACHE_ON )
} // end fetch_rss()
/*=======================================================================*\
Function: _response
Purpose: parse an HTTP response object into an RSS object
Input: an HTTP response object (see Snoopy)
Output: parsed RSS object (see rss_parse)
\*=======================================================================*/
function _response ($resp,$class) {
$rss = new $class( $resp->results, MAGPIE_OUTPUT_ENCODING, MAGPIE_INPUT_ENCODING, MAGPIE_DETECT_ENCODING );
// if RSS parsed successfully
if ( $rss and !$rss->ERROR) {
// find Etag, and Last-Modified
foreach($resp->headers as $h) {
// 2003-03-02 - Nicola Asuni (<a href="http://www.tecnick.com" title="www.tecnick.com" rel="nofollow">www.tecnick.com</a>) - fixed bug "Undefined offset: 1"
if (strpos($h, ": ")) {
list($field, $val) = explode(": ", $h, 2);
}
else {
$field = $h;
$val = "";
}
if ( $field == 'ETag' ) {
$rss->etag = $val;
}
if ( $field == 'Last-Modified' ) {
$rss->last_modified = $val;
}
}
return $rss;
} // else construct error message
else {
$errormsg = "Failed to parse RSS file.";
if ($rss) {
$errormsg .= " (" . $rss->ERROR . ")";
}
error($errormsg);
return false;
} // end if ($rss and !$rss->error)
}
#########################
# Affilaites .nl parser
/*
* Project: MagpieRSS: a simple RSS integration tool
* File: tradetracker.inc, a simple functional interface
to fetching and parsing the XML from tradetracker
* Author: Bram Brambring <bram@brambring.nl>
* License: GPL
* Script requires an installation of MagpieRSS
* The lastest version of MagpieRSS can be obtained from:
* <a href="http://magpierss.sourceforge.net
" title="http://magpierss.sourceforge.net
" rel="nofollow">http://magpierss.sourceforge.net
</a> *
* For questions, help, comments, discussion, etc., please join the
* Magpie mailing list:
* magpierss-general@lists.sourceforge.net
*
*/
class AffiliateRSS extends MagpieRSS{
function AffiliateRSS ($source, $output_encoding='ISO-8859-1',
$input_encoding=null, $detect_encoding=true)
{
# if PHP xml isn't compiled in, die
#
if (!function_exists('xml_parser_create')) {
$this->error( "Failed to load PHP's XML Extension. " .
"http://www.php.net/manual/en/ref.xml.php",
E_USER_ERROR );
}
list($parser, $source) = $this->create_parser($source,
$output_encoding, $input_encoding, $detect_encoding);
if (!is_resource($parser)) {
$this->error( "Failed to create an instance of PHP's XML parser. " .
"http://www.php.net/manual/en/ref.xml.php",
E_USER_ERROR );
}
$this->parser = $parser;
# pass in parser, and a reference to this object
# setup handlers
#
xml_set_object( $this->parser, $this );
xml_set_element_handler($this->parser,
'feed_start_element', 'feed_end_element' );
xml_set_character_data_handler( $this->parser, 'feed_cdata' );
$status = xml_parse( $this->parser, $source );
if (! $status ) {
$errorcode = xml_get_error_code( $this->parser );
if ( $errorcode != XML_ERROR_NONE ) {
$xml_error = xml_error_string( $errorcode );
$error_line = xml_get_current_line_number($this->parser);
$error_col = xml_get_current_column_number($this->parser);
$errormsg = "$xml_error at line $error_line, column $error_col";
$this->error( $errormsg );
}
}
xml_parser_free( $this->parser );
}
function feed_start_element($p, $element, &$attrs) {
$el = $element = strtolower($element);
$attrs = array_change_key_case($attrs, CASE_LOWER);
// check for a namespace, and split if found
$ns = false;
if ( strpos( $element, ':' ) ) {
list($ns, $el) = split( ':', $element, 2);
}
if ( $ns and $ns != 'rdf' ) {
$this->current_namespace = $ns;
}
# if feed type isn't set, then this is first element of feed
# identify feed from root element
#
if ( $el == 'items' )
{
}
elseif ($el == 'item' )
{
$this->initem = true;
}
// set stack[0] to current element
else {
array_unshift($this->stack, $el);
}
}
function feed_end_element ($p, $el) {
$el = strtolower($el);
if ( $el == 'item' )
{
$this->items[] = $this->current_item;
$this->current_item = array();
$this->current_categories = array();
$this->current_fields = array();
$this->initem = false;
}
elseif ($el == 'items' )
{
}
else {
array_shift( $this->stack);
}
$this->current_namespace = false;
}
} // end class RSS
class TradetrackerRSS extends MagpieRSS{
var $current_categories = array(); // item currently being parsed
var $current_fields = array(); // item currently being parsed
var $incategories = false;
var $incategory = false;
var $inadditonal = false;
var $infield = false;
function TradetrackerRSS ($source, $output_encoding='ISO-8859-1',
$input_encoding=null, $detect_encoding=true)
{
# if PHP xml isn't compiled in, die
#
if (!function_exists('xml_parser_create')) {
$this->error( "Failed to load PHP's XML Extension. " .
"http://www.php.net/manual/en/ref.xml.php",
E_USER_ERROR );
}
list($parser, $source) = $this->create_parser($source,
$output_encoding, $input_encoding, $detect_encoding);
if (!is_resource($parser)) {
$this->error( "Failed to create an instance of PHP's XML parser. " .
"http://www.php.net/manual/en/ref.xml.php",
E_USER_ERROR );
}
$this->parser = $parser;
# pass in parser, and a reference to this object
# setup handlers
#
xml_set_object( $this->parser, $this );
xml_set_element_handler($this->parser,
'feed_start_element', 'feed_end_element' );
xml_set_character_data_handler( $this->parser, 'feed_cdata' );
$status = xml_parse( $this->parser, $source );
if (! $status ) {
$errorcode = xml_get_error_code( $this->parser );
if ( $errorcode != XML_ERROR_NONE ) {
$xml_error = xml_error_string( $errorcode );
$error_line = xml_get_current_line_number($this->parser);
$error_col = xml_get_current_column_number($this->parser);
$errormsg = "$xml_error at line $error_line, column $error_col";
$this->error( $errormsg );
}
}
xml_parser_free( $this->parser );
}
function feed_start_element($p, $element, &$attrs) {
$el = $element = strtolower($element);
$attrs = array_change_key_case($attrs, CASE_LOWER);
// check for a namespace, and split if found
$ns = false;
if ( strpos( $element, ':' ) ) {
list($ns, $el) = split( ':', $element, 2);
}
if ( $ns and $ns != 'rdf' ) {
$this->current_namespace = $ns;
}
# if feed type isn't set, then this is first element of feed
# identify feed from root element
#
if ( $el == 'products' )
{
}
elseif ($el == 'product' )
{
$this->initem = true;
}
elseif ($el == 'categories') {
$this->incategories = true;
}
elseif ($el == 'category') {
$this->incategory = true;
$this->concat( $this->current_item['_category'], $attrs['path'] );
$this->concat($this->current_categories[$attrs['name']],$attrs['path']);
}
elseif ($el == 'additional') {
$this->inadditional = true;
}
elseif ($el == 'field') {
$this->infield = true;
if ( isset($attrs['name']) && isset($attrs['value']) ) {
$this->concat($this->current_fields[$attrs['name']],$attrs['value']);
}
}
// set stack[0] to current element
else {
array_unshift($this->stack, $el);
}
}
function feed_end_element ($p, $el) {
$el = strtolower($el);
if ( $el == 'product' )
{
$this->current_item['categories']=$this->current_categories;
$this->current_item['fields']=$this->current_fields;
$this->items[] = $this->current_item;
$this->current_item = array();
$this->current_categories = array();
$this->current_fields = array();
$this->initem = false;
}
elseif ($el == 'additional') {
$this->inadditional = false;
}
elseif ($el == 'field') {
$this->infield = false;
}
elseif ($el == 'categories') {
$this->incategories = false;
}
elseif ($el == 'category') {
$this->incategory = false;
}
elseif ($el == 'products' )
{
}
else {
array_shift( $this->stack);
}
$this->current_namespace = false;
}
} // end class RSS
?>Is it possible to integrate this code in feedparser?
How can this be done?
Greetings,
Martijn