[pLog-svn] r1066 - in plog/trunk/class/xml: rss_parser rssparser
rssparser/magpierss
oscar at devel.plogworld.net
oscar at devel.plogworld.net
Mon Feb 14 15:13:23 GMT 2005
Author: oscar
Date: 2005-02-14 15:13:22 +0000 (Mon, 14 Feb 2005)
New Revision: 1066
Added:
plog/trunk/class/xml/rssparser/magpierdd
plog/trunk/class/xml/rssparser/magpierss/
plog/trunk/class/xml/rssparser/magpierss/README
plog/trunk/class/xml/rssparser/magpierss/extlib/
plog/trunk/class/xml/rssparser/magpierss/rss_fetch.inc
plog/trunk/class/xml/rssparser/magpierss/rss_parse.inc
plog/trunk/class/xml/rssparser/magpierss/rss_utils.inc
Removed:
plog/trunk/class/xml/rss_parser/README
plog/trunk/class/xml/rss_parser/extlib/
plog/trunk/class/xml/rss_parser/rss_cache.inc
plog/trunk/class/xml/rss_parser/rss_fetch.inc
plog/trunk/class/xml/rss_parser/rss_parse.inc
plog/trunk/class/xml/rss_parser/rss_utils.inc
Log:
moved everything to its final new location
Deleted: plog/trunk/class/xml/rss_parser/README
===================================================================
--- plog/trunk/class/xml/rss_parser/README 2005-02-14 15:05:51 UTC (rev 1065)
+++ plog/trunk/class/xml/rss_parser/README 2005-02-14 15:13:22 UTC (rev 1066)
@@ -1,3 +0,0 @@
-Version 0.5.1 of the MagpieRSS library
-
-Please visit http://magpierss.sourceforge.net/
Deleted: plog/trunk/class/xml/rss_parser/rss_cache.inc
===================================================================
--- plog/trunk/class/xml/rss_parser/rss_cache.inc 2005-02-14 15:05:51 UTC (rev 1065)
+++ plog/trunk/class/xml/rss_parser/rss_cache.inc 2005-02-14 15:13:22 UTC (rev 1066)
@@ -1,184 +0,0 @@
-<?php
-/*
- * Project: MagpieRSS: a simple RSS integration tool
- * File: rss_cache.inc, a simple, rolling(no GC), cache
- * for RSS objects, keyed on URL.
- * Author: Kellan Elliott-McCrea <kellan at protest.net>
- * Version: 0.51
- * License: GPL
- *
- * The lastest version of MagpieRSS can be obtained from:
- * http://magpierss.sourceforge.net
- *
- * For questions, help, comments, discussion, etc., please join the
- * Magpie mailing list:
- * http://lists.sourceforge.net/lists/listinfo/magpierss-general
- *
- */
-
-class RSSCache {
- var $BASE_CACHE = './cache'; // where the cache files are stored
- var $MAX_AGE = 3600; // when are files stale, default one hour
- var $ERROR = ""; // accumulate error messages
-
- function RSSCache ($base='', $age='') {
- if ( $base ) {
- $this->BASE_CACHE = $base;
- }
- if ( $age ) {
- $this->MAX_AGE = $age;
- }
-
- // attempt to make the cache directory
- if ( ! file_exists( $this->BASE_CACHE ) ) {
- $status = @mkdir( $this->BASE_CACHE, 0755 );
-
- // if make failed
- if ( ! $status ) {
- $this->error(
- "Cache couldn't make dir '" . $this->BASE_CACHE . "'."
- );
- }
- }
- }
-
-/*=======================================================================*\
- Function: set
- Purpose: add an item to the cache, keyed on url
- Input: url from wich the rss file was fetched
- Output: true on sucess
-\*=======================================================================*/
- function set ($url, $rss) {
- $this->ERROR = "";
- $cache_file = $this->file_name( $url );
- $fp = @fopen( $cache_file, 'w' );
-
- if ( ! $fp ) {
- $this->error(
- "Cache unable to open file for writing: $cache_file"
- );
- return 0;
- }
-
-
- $data = $this->serialize( $rss );
- fwrite( $fp, $data );
- fclose( $fp );
-
- return $cache_file;
- }
-
-/*=======================================================================*\
- Function: get
- Purpose: fetch an item from the cache
- Input: url from wich the rss file was fetched
- Output: cached object on HIT, false on MISS
-\*=======================================================================*/
- function get ($url) {
- $this->ERROR = "";
- $cache_file = $this->file_name( $url );
-
- if ( ! file_exists( $cache_file ) ) {
- $this->debug(
- "Cache doesn't contain: $url (cache file: $cache_file)"
- );
- return 0;
- }
-
- $fp = @fopen($cache_file, 'r');
- if ( ! $fp ) {
- $this->error(
- "Failed to open cache file for reading: $cache_file"
- );
- return 0;
- }
-
- $data = fread( $fp, filesize($cache_file) );
- $rss = $this->unserialize( $data );
-
- return $rss;
- }
-
-/*=======================================================================*\
- Function: check_cache
- Purpose: check a url for membership in the cache
- and whether the object is older then MAX_AGE (ie. STALE)
- Input: url from wich the rss file was fetched
- Output: cached object on HIT, false on MISS
-\*=======================================================================*/
- function check_cache ( $url ) {
- $this->ERROR = "";
- $filename = $this->file_name( $url );
-
- if ( file_exists( $filename ) ) {
- // find how long ago the file was added to the cache
- // and whether that is longer then MAX_AGE
- $mtime = filemtime( $filename );
- $age = time() - $mtime;
- if ( $this->MAX_AGE > $age ) {
- // object exists and is current
- return 'HIT';
- }
- else {
- // object exists but is old
- return 'STALE';
- }
- }
- else {
- // object does not exist
- return 'MISS';
- }
- }
-
-/*=======================================================================*\
- Function: serialize
-\*=======================================================================*/
- function serialize ( $rss ) {
- return serialize( $rss );
- }
-
-/*=======================================================================*\
- Function: unserialize
-\*=======================================================================*/
- function unserialize ( $data ) {
- return unserialize( $data );
- }
-
-/*=======================================================================*\
- Function: file_name
- Purpose: map url to location in cache
- Input: url from wich the rss file was fetched
- Output: a file name
-\*=======================================================================*/
- function file_name ($url) {
- $filename = md5( $url );
- return join( DIRECTORY_SEPARATOR, array( $this->BASE_CACHE, $filename ) );
- }
-
-/*=======================================================================*\
- Function: error
- Purpose: register error
-\*=======================================================================*/
- function error ($errormsg, $lvl=E_USER_WARNING) {
- // append PHP's error message if track_errors enabled
- if ( isset($php_errormsg) ) {
- $errormsg .= " ($php_errormsg)";
- }
- $this->ERROR = $errormsg;
- if ( MAGPIE_DEBUG ) {
- trigger_error( $errormsg, $lvl);
- }
- else {
- error_log( $errormsg, 0);
- }
- }
-
- function debug ($debugmsg, $lvl=E_USER_NOTICE) {
- if ( MAGPIE_DEBUG ) {
- $this->error("MagpieRSS [debug] $debugmsg", $lvl);
- }
- }
-
-}
-
-?>
Deleted: plog/trunk/class/xml/rss_parser/rss_fetch.inc
===================================================================
--- plog/trunk/class/xml/rss_parser/rss_fetch.inc 2005-02-14 15:05:51 UTC (rev 1065)
+++ plog/trunk/class/xml/rss_parser/rss_fetch.inc 2005-02-14 15:13:22 UTC (rev 1066)
@@ -1,438 +0,0 @@
-<?php
-/*
- * Project: MagpieRSS: a simple RSS integration tool
- * File: rss_fetch.inc, a simple functional interface
- to fetching and parsing RSS files, via the
- function fetch_rss()
- * Author: Kellan Elliott-McCrea <kellan at protest.net>
- * License: GPL
- *
- * The lastest version of MagpieRSS can be obtained from:
- * http://magpierss.sourceforge.net
- *
- * For questions, help, comments, discussion, etc., please join the
- * Magpie mailing list:
- * magpierss-general at lists.sourceforge.net
- *
- */
-
-// Setup MAGPIE_DIR for use on hosts that don't include
-// the current path in include_path.
-// with thanks to rajiv and smarty
-if (!defined('DIR_SEP')) {
- define('DIR_SEP', DIRECTORY_SEPARATOR);
-}
-
-if (!defined('MAGPIE_DIR')) {
- define('MAGPIE_DIR', dirname(__FILE__) . DIR_SEP);
-}
-
-require_once( MAGPIE_DIR . 'rss_parse.inc' );
-require_once( MAGPIE_DIR . 'rss_cache.inc' );
-
-// for including 3rd party libraries
-define('MAGPIE_EXTLIB', MAGPIE_DIR . 'extlib' . DIR_SEP);
-require_once( MAGPIE_EXTLIB . 'Snoopy.class.inc');
-
-
-/*
- * CONSTANTS - redefine these in your script to change the
- * behaviour of fetch_rss() currently, most options effect the cache
- *
- * MAGPIE_CACHE_ON - Should Magpie cache parsed RSS objects?
- * For me a built in cache was essential to creating a "PHP-like"
- * feel to Magpie, see rss_cache.inc for rationale
- *
- *
- * MAGPIE_CACHE_DIR - Where should Magpie cache parsed RSS objects?
- * This should be a location that the webserver can write to. If this
- * directory does not already exist Mapie will try to be smart and create
- * it. This will often fail for permissions reasons.
- *
- *
- * MAGPIE_CACHE_AGE - How long to store cached RSS objects? In seconds.
- *
- *
- * MAGPIE_CACHE_FRESH_ONLY - If remote fetch fails, throw error
- * instead of returning stale object?
- *
- * MAGPIE_DEBUG - Display debugging notices?
- *
-*/
-
-
-/*=======================================================================*\
- Function: fetch_rss:
- Purpose: return RSS object for the give url
- maintain the cache
- Input: url of RSS file
- Output: parsed RSS object (see rss_parse.inc)
-
- NOTES ON CACHEING:
- If caching is on (MAGPIE_CACHE_ON) fetch_rss will first check the cache.
-
- NOTES ON RETRIEVING REMOTE FILES:
- If conditional gets are on (MAGPIE_CONDITIONAL_GET_ON) fetch_rss will
- return a cached object, and touch the cache object upon recieving a
- 304.
-
- NOTES ON FAILED REQUESTS:
- If there is an HTTP error while fetching an RSS object, the cached
- version will be return, if it exists (and if MAGPIE_CACHE_FRESH_ONLY is off)
-\*=======================================================================*/
-
-define('MAGPIE_VERSION', '0.61');
-
-$MAGPIE_ERROR = "";
-
-function fetch_rss ($url) {
- // initialize constants
- init();
-
- if ( !isset($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_to_rss( $resp );
- }
- 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, MAGPIE_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
- $errormsg = 0; // errors, if any
-
- if (!$cache->ERROR) {
- // return cache HIT, MISS, or STALE
- $cache_status = $cache->check_cache( $url );
- }
-
- // if object cached, and cache is fresh, return cached obj
- if ( $cache_status == 'HIT' ) {
- $rss = $cache->get( $url );
- if ( isset($rss) and $rss ) {
- $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( $url );
- if ( $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)
- $cache->set($url, $rss);
- return $rss;
- }
- elseif ( is_success( $resp->status ) ) {
- $rss = _response_to_rss( $resp );
- if ( $rss ) {
- if (MAGPIE_DEBUG > 1) {
- debug("Fetch successful");
- }
- // add object to cache
- $cache->set( $url, $rss );
- return $rss;
- }
- }
- else {
- $errormsg = "Failed to fetch $url. ";
- if ( $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( $errormsg );
-
- return false;
-
- } // end if ( !MAGPIE_CACHE_ON ) {
-} // end fetch_rss()
-
-/*=======================================================================*\
- Function: error
- Purpose: set MAGPIE_ERROR, and trigger error
-\*=======================================================================*/
-
-function error ($errormsg, $lvl=E_USER_WARNING) {
- global $MAGPIE_ERROR;
-
- // append PHP's error message if track_errors enabled
- if ( isset($php_errormsg) ) {
- $errormsg .= " ($php_errormsg)";
- }
- if ( $errormsg ) {
- $errormsg = "MagpieRSS: $errormsg";
- $MAGPIE_ERROR = $errormsg;
- trigger_error( $errormsg, $lvl);
- }
-}
-
-function debug ($debugmsg, $lvl=E_USER_NOTICE) {
- trigger_error("MagpieRSS [debug] $debugmsg", $lvl);
-}
-
-/*=======================================================================*\
- Function: magpie_error
- Purpose: accessor for the magpie error variable
-\*=======================================================================*/
-function magpie_error ($errormsg="") {
- global $MAGPIE_ERROR;
-
- if ( isset($errormsg) and $errormsg ) {
- $MAGPIE_ERROR = $errormsg;
- }
-
- return $MAGPIE_ERROR;
-}
-
-/*=======================================================================*\
- Function: _fetch_remote_file
- Purpose: retrieve an arbitrary remote file
- Input: url of the remote file
- headers to send along with the request (optional)
- Output: an HTTP response object (see Snoopy.class.inc)
-\*=======================================================================*/
-function _fetch_remote_file ($url, $headers = "" ) {
- // Snoopy is an HTTP client in PHP
- $client = new Snoopy();
- $client->agent = MAGPIE_USER_AGENT;
- $client->read_timeout = MAGPIE_FETCH_TIME_OUT;
- $client->use_gzip = MAGPIE_USE_GZIP;
- if (is_array($headers) ) {
- $client->rawheaders = $headers;
- }
-
- @$client->fetch($url);
- return $client;
-
-}
-
-/*=======================================================================*\
- Function: _response_to_rss
- Purpose: parse an HTTP response object into an RSS object
- Input: an HTTP response object (see Snoopy)
- Output: parsed RSS object (see rss_parse)
-\*=======================================================================*/
-function _response_to_rss ($resp) {
- $rss = new MagpieRSS( $resp->results );
-
- // if RSS parsed successfully
- if ( $rss and !$rss->ERROR) {
-
- // find Etag, and Last-Modified
- foreach($resp->headers as $h) {
- // 2003-03-02 - Nicola Asuni (www.tecnick.com) - fixed bug "Undefined offset: 1"
- if (strpos($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)
-}
-
-/*=======================================================================*\
- Function: init
- Purpose: setup constants with default values
- check for user overrides
-\*=======================================================================*/
-function init () {
- if ( defined('MAGPIE_INITALIZED') ) {
- return;
- }
- else {
- define('MAGPIE_INITALIZED', 1);
- }
-
- if ( !defined('MAGPIE_CACHE_ON') ) {
- define('MAGPIE_CACHE_ON', 1);
- }
-
- if ( !defined('MAGPIE_CACHE_DIR') ) {
- define('MAGPIE_CACHE_DIR', './cache');
- }
-
- if ( !defined('MAGPIE_CACHE_AGE') ) {
- define('MAGPIE_CACHE_AGE', 60*60); // one hour
- }
-
- if ( !defined('MAGPIE_CACHE_FRESH_ONLY') ) {
- define('MAGPIE_CACHE_FRESH_ONLY', 0);
- }
-
- if ( !defined('MAGPIE_DEBUG') ) {
- define('MAGPIE_DEBUG', 0);
- }
-
- if ( !defined('MAGPIE_USER_AGENT') ) {
- $ua = 'MagpieRSS/'. MAGPIE_VERSION . ' (+http://magpierss.sf.net';
-
- if ( MAGPIE_CACHE_ON ) {
- $ua = $ua . ')';
- }
- else {
- $ua = $ua . '; No cache)';
- }
-
- define('MAGPIE_USER_AGENT', $ua);
- }
-
- if ( !defined('MAGPIE_FETCH_TIME_OUT') ) {
- define('MAGPIE_FETCH_TIME_OUT', 5); // 5 second timeout
- }
-
- // use gzip encoding to fetch rss files if supported?
- if ( !defined('MAGPIE_USE_GZIP') ) {
- define('MAGPIE_USE_GZIP', true);
- }
-}
-
-// NOTE: the following code should really be in Snoopy, or at least
-// somewhere other then rss_fetch!
-
-/*=======================================================================*\
- HTTP STATUS CODE PREDICATES
- These functions attempt to classify an HTTP status code
- based on RFC 2616 and RFC 2518.
-
- All of them take an HTTP status code as input, and return true or false
-
- All this code is adapted from LWP's HTTP::Status.
-\*=======================================================================*/
-
-
-/*=======================================================================*\
- Function: is_info
- Purpose: return true if Informational status code
-\*=======================================================================*/
-function is_info ($sc) {
- return $sc >= 100 && $sc < 200;
-}
-
-/*=======================================================================*\
- Function: is_success
- Purpose: return true if Successful status code
-\*=======================================================================*/
-function is_success ($sc) {
- return $sc >= 200 && $sc < 300;
-}
-
-/*=======================================================================*\
- Function: is_redirect
- Purpose: return true if Redirection status code
-\*=======================================================================*/
-function is_redirect ($sc) {
- return $sc >= 300 && $sc < 400;
-}
-
-/*=======================================================================*\
- Function: is_error
- Purpose: return true if Error status code
-\*=======================================================================*/
-function is_error ($sc) {
- return $sc >= 400 && $sc < 600;
-}
-
-/*=======================================================================*\
- Function: is_client_error
- Purpose: return true if Error status code, and its a client error
-\*=======================================================================*/
-function is_client_error ($sc) {
- return $sc >= 400 && $sc < 500;
-}
-
-/*=======================================================================*\
- Function: is_client_error
- Purpose: return true if Error status code, and its a server error
-\*=======================================================================*/
-function is_server_error ($sc) {
- return $sc >= 500 && $sc < 600;
-}
-
-?>
Deleted: plog/trunk/class/xml/rss_parser/rss_parse.inc
===================================================================
--- plog/trunk/class/xml/rss_parser/rss_parse.inc 2005-02-14 15:05:51 UTC (rev 1065)
+++ plog/trunk/class/xml/rss_parser/rss_parse.inc 2005-02-14 15:13:22 UTC (rev 1066)
@@ -1,547 +0,0 @@
-<?php
-/*
- * Project: MagpieRSS: a simple RSS integration tool
- * File: rss_parse.inc - parse an RSS or Atom feed
- * return as a simple object.
- *
- * Handles RSS 0.9x, RSS 2.0, RSS 1.0, and Atom 0.3
- *
- * The lastest version of MagpieRSS can be obtained from:
- * http://magpierss.sourceforge.net
- *
- * For questions, help, comments, discussion, etc., please join the
- * Magpie mailing list:
- * magpierss-general at lists.sourceforge.net
- *
- * Author: Kellan Elliott-McCrea <kellan at protest.net>
- * Version: 0.6a
- * License: GPL
- *
- *
- * ABOUT MAGPIE's APPROACH TO PARSING:
- * - Magpie is based on expat, an XML parser, and therefore will only parse
- * valid XML files. This includes all properly constructed RSS or Atom.
- *
- * - Magpie is an inclusive parser. It will include any elements that
- * it can turn into a key value pair in the parsed feed object it returns.
- *
- * - Magpie supports namespaces, and will return any elements found in a
- * namespace in a sub-array, with the key point to that array being the
- * namespace prefix.
- * (e.g. if an item contains a <dc:date> element, then that date can
- * be accessed at $item['dc']['date']
- *
- * - Magpie supports nested elements by combining the names. If an item
- * includes XML like:
- * <author>
- * <name>Kellan</name>
- * </author>
- *
- * The name field is accessible at $item['author_name']
- *
- * - Magpie makes no attempt validate a feed beyond insuring that it
- * is valid XML.
- * RSS validators are readily available on the web at:
- * http://feeds.archive.org/validator/
- * http://www.ldodds.com/rss_validator/1.0/validator.html
- *
- *
- * EXAMPLE PARSED RSS ITEM:
- *
- * Magpie tries to parse RSS into easy to use PHP datastructures.
- *
- * For example, Magpie on encountering (a rather complex) RSS 1.0 item entry:
- *
- * <item rdf:about="http://protest.net/NorthEast/calendrome.cgi?span=event&ID=210257">
- * <title>Weekly Peace Vigil</title>
- * <link>http://protest.net/NorthEast/calendrome.cgi?span=event&ID=210257</link>
- * <description>Wear a white ribbon</description>
- * <dc:subject>Peace</dc:subject>
- * <ev:startdate>2002-06-01T11:00:00</ev:startdate>
- * <ev:location>Northampton, MA</ev:location>
- * <ev:type>Protest</ev:type>
- * </item>
- *
- * Would transform it into the following associative array, and push it
- * onto the array $rss-items
- *
- * array(
- * title => 'Weekly Peace Vigil',
- * link => 'http://protest.net/NorthEast/calendrome.cgi?span=event&ID=210257',
- * description => 'Wear a white ribbon',
- * dc => array (
- * subject => 'Peace'
- * ),
- * ev => array (
- * startdate => '2002-06-01T11:00:00',
- * enddate => '2002-06-01T12:00:00',
- * type => 'Protest',
- * location => 'Northampton, MA'
- * )
- * )
- *
- *
- *
- * A FEW NOTES ON PARSING Atom FEEDS
- *
- * Atom support is considered alpha. Atom elements will be often be available
- * as their RSS equivalent, summary is available as description for example.
- *
- * Elements of mode=xml, as flattened into a single string, just as if they
- * had been wrapped in a CDATA container.
- *
- * See: http://laughingmeme.org/archives/001676.html
- *
- */
-
-define('RSS', 'RSS');
-define('ATOM', 'Atom');
-
-
-class MagpieRSS {
- /*
- * Hybrid parser, and object. (probably a bad idea! :)
- *
- * Useage Example:
- *
- * $some_rss = "<?xml version="1.0"......
- *
- * $rss = new MagpieRSS( $some_rss );
- *
- * // print rss chanel title
- * echo $rss->channel['title'];
- *
- * // print the title of each item
- * foreach ($rss->items as $item ) {
- * echo $item[title];
- * }
- *
- * see: rss_fetch.inc for a simpler interface
- */
-
- var $parser;
-
- var $current_item = array(); // item currently being parsed
- var $items = array(); // collection of parsed items
- var $channel = array(); // hash of channel fields
- var $textinput = array();
- var $image = array();
- var $feed_type;
- var $feed_version;
-
- // parser variables
- var $stack = array(); // parser stack
- var $inchannel = false;
- var $initem = false;
- var $incontent = false; // if in Atom <content mode="xml"> field
- var $intextinput = false;
- var $inimage = false;
- var $current_field = '';
- var $current_namespace = false;
-
- var $ERROR = "";
-
- var $_CONTENT_CONSTRUCTS = array('content', 'summary', 'info', 'title', 'tagline', 'copyright');
-/*======================================================================*\
- Function: MagpieRSS
- Purpose: Constructor, sets up XML parser,parses source,
- and populates object..
- Input: String containing the RSS to be parsed
-\*======================================================================*/
- function MagpieRSS ($source) {
-
- # 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 );
- }
-
- $parser = @xml_parser_create();
-
- if (!is_resource($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 );
-
- $this->normalize();
- }
-
- function feed_start_element($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 = $ns;
- }
-
- # if feed type isn't set, then this is first element of feed
- # identify feed from root element
- #
- if (!isset($this->feed_type) ) {
- if ( $el == 'rdf' ) {
- $this->feed_type = RSS;
- $this->feed_version = '1.0';
- }
- elseif ( $el == 'rss' ) {
- $this->feed_type = RSS;
- $this->feed_version = $attrs['version'];
- }
- elseif ( $el == 'feed' ) {
- $this->feed_type = ATOM;
- $this->feed_version = $attrs['version'];
- $this->inchannel = true;
- }
- return;
- }
-
- if ( $el == 'channel' )
- {
- $this->inchannel = true;
- }
- elseif ($el == 'item' or $el == 'entry' )
- {
- $this->initem = true;
- if ( isset($attrs['rdf:about']) ) {
- $this->current_item['about'] = $attrs['rdf:about'];
- }
- }
-
- // if we're in the default namespace of an RSS feed,
- // record textinput or image fields
- elseif (
- $this->feed_type == RSS and
- $this->current_namespace == '' and
- $el == 'textinput' )
- {
- $this->intextinput = true;
- }
-
- elseif (
- $this->feed_type == RSS and
- $this->current_namespace == '' and
- $el == 'image' )
- {
- $this->inimage = true;
- }
-
- # handle atom content constructs
- elseif ( $this->feed_type == ATOM and in_array($el, $this->_CONTENT_CONSTRUCTS) )
- {
- // avoid clashing w/ RSS mod_content
- if ($el == 'content' ) {
- $el = 'atom_content';
- }
-
- $this->incontent = $el;
-
-
- }
-
- // if inside an Atom content construct (e.g. content or summary) field treat tags as text
- elseif ($this->feed_type == ATOM and $this->incontent )
- {
- // if tags are inlined, then flatten
- $attrs_str = join(' ',
- array_map('map_attrs',
- array_keys($attrs),
- array_values($attrs) ) );
-
- $this->append_content( "<$element $attrs_str>" );
-
- array_unshift( $this->stack, $el );
- }
-
- // Atom support many links per containging element.
- // Magpie treats link elements of type rel='alternate'
- // as being equivalent to RSS's simple link element.
- //
- elseif ($this->feed_type == ATOM and $el == 'link' )
- {
- if ( isset($attrs['rel']) and $attrs['rel'] == 'alternate' )
- {
- $link_el = 'link';
- }
- else {
- $link_el = 'link_' . $attrs['rel'];
- }
-
- $this->append($link_el, $attrs['href']);
- }
- // set stack[0] to current element
- else {
- array_unshift($this->stack, $el);
- }
- }
-
-
-
- function feed_cdata ($p, $text) {
-
- if ($this->feed_type == ATOM and $this->incontent)
- {
- $this->append_content( $text );
- }
- else {
- $current_el = join('_', array_reverse($this->stack));
- $this->append($current_el, $text);
- }
- }
-
- function feed_end_element ($p, $el) {
- $el = strtolower($el);
-
- if ( $el == 'item' or $el == 'entry' )
- {
- $this->items[] = $this->current_item;
- $this->current_item = array();
- $this->initem = false;
- }
- elseif ($this->feed_type == RSS and $this->current_namespace == '' and $el == 'textinput' )
- {
- $this->intextinput = false;
- }
- elseif ($this->feed_type == RSS and $this->current_namespace == '' and $el == 'image' )
- {
- $this->inimage = false;
- }
- elseif ($this->feed_type == ATOM and in_array($el, $this->_CONTENT_CONSTRUCTS) )
- {
- $this->incontent = false;
- }
- elseif ($el == 'channel' or $el == 'feed' )
- {
- $this->inchannel = false;
- }
- elseif ($this->feed_type == ATOM and $this->incontent ) {
- // balance tags properly
- // note: i don't think this is actually neccessary
- if ( $this->stack[0] == $el )
- {
- $this->append_content("</$el>");
- }
- else {
- $this->append_content("<$el />");
- }
-
- array_shift( $this->stack );
- }
- else {
- array_shift( $this->stack );
- }
-
- $this->current_namespace = false;
- }
-
- function concat (&$str1, $str2="") {
- if (!isset($str1) ) {
- $str1="";
- }
- $str1 .= $str2;
- }
-
-
-
- function append_content($text) {
- if ( $this->initem ) {
- $this->concat( $this->current_item[ $this->incontent ], $text );
- }
- elseif ( $this->inchannel ) {
- $this->concat( $this->channel[ $this->incontent ], $text );
- }
- }
-
- // smart append - field and namespace aware
- function append($el, $text) {
- if (!$el) {
- return;
- }
- if ( $this->current_namespace )
- {
- if ( $this->initem ) {
- $this->concat(
- $this->current_item[ $this->current_namespace ][ $el ], $text);
- }
- elseif ($this->inchannel) {
- $this->concat(
- $this->channel[ $this->current_namespace][ $el ], $text );
- }
- elseif ($this->intextinput) {
- $this->concat(
- $this->textinput[ $this->current_namespace][ $el ], $text );
- }
- elseif ($this->inimage) {
- $this->concat(
- $this->image[ $this->current_namespace ][ $el ], $text );
- }
- }
- else {
- if ( $this->initem ) {
- $this->concat(
- $this->current_item[ $el ], $text);
- }
- elseif ($this->intextinput) {
- $this->concat(
- $this->textinput[ $el ], $text );
- }
- elseif ($this->inimage) {
- $this->concat(
- $this->image[ $el ], $text );
- }
- elseif ($this->inchannel) {
- $this->concat(
- $this->channel[ $el ], $text );
- }
-
- }
- }
-
- function normalize () {
- // if atom populate rss fields
- if ( $this->is_atom() ) {
- $this->channel['descripton'] = $this->channel['tagline'];
- for ( $i = 0; $i < count($this->items); $i++) {
- $item = $this->items[$i];
- if ( isset($item['summary']) )
- $item['description'] = $item['summary'];
- if ( isset($item['atom_content']))
- $item['content']['encoded'] = $item['atom_content'];
-
- $this->items[$i] = $item;
- }
- }
- elseif ( $this->is_rss() ) {
- $this->channel['tagline'] = $this->channel['description'];
- for ( $i = 0; $i < count($this->items); $i++) {
- $item = $this->items[$i];
- if ( isset($item['description']))
- $item['summary'] = $item['description'];
- if ( isset($item['content']['encoded'] ) )
- $item['atom_content'] = $item['content']['encoded'];
-
- $this->items[$i] = $item;
- }
- }
- }
-
- function error ($errormsg, $lvl=E_USER_WARNING) {
- // append PHP's error message if track_errors enabled
- if ( $php_errormsg ) {
- $errormsg .= " ($php_errormsg)";
- }
- $this->ERROR = $errormsg;
- if ( MAGPIE_DEBUG ) {
- trigger_error( $errormsg, $lvl);
- }
- else {
- error_log( $errormsg, 0);
- }
- }
-
- function is_rss () {
- if ( $this->feed_type == RSS ) {
- return $this->feed_version;
- }
- else {
- return false;
- }
- }
-
- function is_atom() {
- if ( $this->feed_type == ATOM ) {
- return $this->feed_version;
- }
- else {
- return false;
- }
- }
-
-/*======================================================================*\
- EVERYTHING BELOW HERE IS FOR DEBUGGING PURPOSES
-\*======================================================================*/
- function show_list () {
- echo "<ol>\n";
- foreach ($this->items as $item) {
- echo "<li>", $this->show_item( $item );
- }
- echo "</ol>";
- }
-
- function show_channel () {
- echo "channel:<br>";
- echo "<ul>";
- while ( list($key, $value) = each( $this->channel ) ) {
- echo "<li> $key: $value";
- }
- echo "</ul>";
- }
-
- function show_item ($item) {
- echo "item: $item[title]";
- echo "<ul>";
- while ( list($key, $value) = each($item) ) {
- if ( is_array($value) ) {
- echo "<br><b>$key</b>";
- echo "<ul>";
- while ( list( $ns_key, $ns_value) = each( $value ) ) {
- echo "<li>$ns_key: $ns_value";
- }
- echo "</ul>";
- }
- else {
- echo "<li> $key: $value";
- }
- }
- echo "</ul>";
- }
-
-/*======================================================================*\
- END DEBUGGING FUNCTIONS
-\*======================================================================*/
-
-
-
-} # end class RSS
-
-function map_attrs($k, $v) {
- return "$k=\"$v\"";
-}
-
-
-?>
Deleted: plog/trunk/class/xml/rss_parser/rss_utils.inc
===================================================================
--- plog/trunk/class/xml/rss_parser/rss_utils.inc 2005-02-14 15:05:51 UTC (rev 1065)
+++ plog/trunk/class/xml/rss_parser/rss_utils.inc 2005-02-14 15:13:22 UTC (rev 1066)
@@ -1,67 +0,0 @@
-<?php
-/*
- * Project: MagpieRSS: a simple RSS integration tool
- * File: rss_utils.inc, utility methods for working with RSS
- * Author: Kellan Elliott-McCrea <kellan at protest.net>
- * Version: 0.51
- * License: GPL
- *
- * The lastest version of MagpieRSS can be obtained from:
- * http://magpierss.sourceforge.net
- *
- * For questions, help, comments, discussion, etc., please join the
- * Magpie mailing list:
- * magpierss-general at lists.sourceforge.net
- */
-
-
-/*======================================================================*\
- Function: parse_w3cdtf
- Purpose: parse a W3CDTF date into unix epoch
-
- NOTE: http://www.w3.org/TR/NOTE-datetime
-\*======================================================================*/
-
-function parse_w3cdtf ( $date_str ) {
-
- # regex to match wc3dtf
- $pat = "/(\d{4})-(\d{2})-(\d{2})T(\d{2}):(\d{2})(:(\d{2}))?(?:([-+])(\d{2}):?(\d{2})|(Z))?/";
-
- if ( preg_match( $pat, $date_str, $match ) ) {
- list( $year, $month, $day, $hours, $minutes, $seconds) =
- array( $match[1], $match[2], $match[3], $match[4], $match[5], $match[6]);
-
- # calc epoch for current date assuming GMT
- $epoch = gmmktime( $hours, $minutes, $seconds, $month, $day, $year);
-
- $offset = 0;
- if ( $match[10] == 'Z' ) {
- # zulu time, aka GMT
- }
- else {
- list( $tz_mod, $tz_hour, $tz_min ) =
- array( $match[8], $match[9], $match[10]);
-
- # zero out the variables
- if ( ! $tz_hour ) { $tz_hour = 0; }
- if ( ! $tz_min ) { $tz_min = 0; }
-
- $offset_secs = (($tz_hour*60)+$tz_min)*60;
-
- # is timezone ahead of GMT? then subtract offset
- #
- if ( $tz_mod == '+' ) {
- $offset_secs = $offset_secs * -1;
- }
-
- $offset = $offset_secs;
- }
- $epoch = $epoch + $offset;
- return $epoch;
- }
- else {
- return -1;
- }
-}
-
-?>
Copied: plog/trunk/class/xml/rssparser/magpierdd (from rev 1065, plog/trunk/class/xml/rss_parser/rss_cache.inc)
Copied: plog/trunk/class/xml/rssparser/magpierss/README (from rev 1065, plog/trunk/class/xml/rss_parser/README)
Copied: plog/trunk/class/xml/rssparser/magpierss/extlib (from rev 1065, plog/trunk/class/xml/rss_parser/extlib)
Copied: plog/trunk/class/xml/rssparser/magpierss/rss_fetch.inc (from rev 1065, plog/trunk/class/xml/rss_parser/rss_fetch.inc)
Copied: plog/trunk/class/xml/rssparser/magpierss/rss_parse.inc (from rev 1065, plog/trunk/class/xml/rss_parser/rss_parse.inc)
Copied: plog/trunk/class/xml/rssparser/magpierss/rss_utils.inc (from rev 1065, plog/trunk/class/xml/rss_parser/rss_utils.inc)
More information about the pLog-svn
mailing list