atomparser.php

<?php

/* ---

    A really simple Atom/XML parser with caching.

    Usage:

    try
    {
        $p = new AtomParser('fluxbb.atom');

        print_r($p->parse());
    }
    catch (AtomParserException $e)
    {
        echo $e->getMessage(), "\n\n";

        print_r($e->get_libxml_errors());
    }

    Changelog:

     - 15 May, 2010: Replaced filectime() with filemtime() (Fixed by Franz/lie2815).

--- */

class AtomParser
{
    const DS = DIRECTORY_SEPARATOR;
    private $parser, $xml_file, $simplexml_obj, $cache, $cache_file,
            $cache_file_exists, $cache_duraction;

    public function __construct($xml_file, $cache=false, $cache_path='/tmp',
                                $cache_duraction=3600)
    {
        $this->xml_file = $xml_file;

        if ($this->cache = $cache)
        {
            $this->cache_file = $cache_path.self::DS.'atom-parser-'.sha1($this->xml_file).'.php';
            $this->cache_duraction = $cache_duraction;
            $this->cache_file_exists = file_exists($this->cache_file);
        }
    }

    public function parse()
    {
        // Read cache if cache file exists and not expired
        if ($this->cache && $this->cache_file_exists &&
            (time() - filemtime($this->cache_file)) < $this->cache_duraction)
        {
            return require $this->cache_file;
        }

        // Enable this so libxml_get_errors() can be used to get all the errors
        libxml_use_internal_errors(true);

        // LIBXML_NOCDATA: Merge CDATA as text nodes
        // http://www.php.net/manual/en/libxml.constants.php
        $this->simplexml_obj = @simplexml_load_file($this->xml_file, null, LIBXML_NOCDATA);

        if (!$this->simplexml_obj)
        {
            $libxml_errors = libxml_get_errors();
            libxml_clear_errors();

            throw new AtomParserException('Failed to load XML file: '.$this->xml_file,
                                          $libxml_errors);
        }

        $array = $this->object2array($this->simplexml_obj);

        // Write array to the cache
        if ($this->cache)
        {
            if (!@file_put_contents($this->cache_file, '<?php'."\n\n".'return '.var_export($array, true).';'))
                throw new AtomParserException('Failed to write cache file: '.$this->cache_file);
        }

        return $array;
    }

    // A modified version of:
    // http://nl.php.net/manual/en/function.simplexml-load-file.php#56691
    private function object2array($object)
    {
        $return = null;

        if (is_array($object))
            foreach ($object as $key => $value)
                $return[$key] = $this->object2array($value);
        else
        {
            if (is_object($object))
            {
                $var = get_object_vars($object);
                foreach ($var as $key => $value)
                    $return[$key] = ($key && !$value) ? null : $this->object2array($value);
            }
            else
                return $object;
        }

        return $return;
    }
}

class AtomParserException extends Exception
{
    private $libxml_errors;

    public function __construct($message, $libxml_errors, $code = 0,
                                Exception $previous = null)
    {
        $this->libxml_errors = $libxml_errors;

        if (version_compare(PHP_VERSION, '5.3.0', '>='))
            parent::__construct($message, $code, $previous);
        else
            parent::__construct($message, $code);
    }

    public function get_libxml_errors()
    {
        return $this->libxml_errors;
    }
}