Magento CLI import script

dataflow.php

#!/usr/bin/php
<?php

/**
 * @file
 * Based on http://www.magentocommerce.com/boards/viewthread/113901/P30/#t277875
 * 
 * @todo
 * - fix importing
 * - fix logging
 */

// Disallow running via web server.
if (PHP_SAPI != 'cli') {
    ob_clean();
    header('HTTP/1.0 403 Forbidden');
    print '<h1>You are forbidden from accessing this resource.</h1>';
    exit();
}

$script = basename(__FILE__);
$usage_example = "Usage: {$script} profile_id\n";

// Syntax check of first argument: profile_id.
if (!isset($argv[1]) || !is_numeric($argv[1]) || (int) $argv[1] != $argv[1]) {
    exit($usage_example);
}
else {
    $profile_id = $argv[1];
}

// Log file used to report results.
$logfile = 'dataflow_cli.log';

// Load the main Magento file.
require './app/Mage.php';

Mage::app()->setCurrentStore(Mage_Core_Model_App::ADMIN_STORE_ID);

// Instantiate a session for the "root" user.
$userModel = Mage::getModel('admin/user');
$userModel->setUserId(0);
Mage::getSingleton('admin/session')->setUser($userModel);

// Load the dataflow profile.
$profile = Mage::getModel('dataflow/profile');
$profile->load($profile_id);
if (!$profile->getId()) {
    exit("Profile with id #{$profile_id} does not exist.");
}

Mage::log("$direction {$profile_id} started.");

//Mage::register('current_convert_profile', $profile);

// Run the dataflow profile.
$profile->run();
$recordCount = 0;
$batchModel = Mage::getSingleton('dataflow/batch');

// Reporting.
$direction = ucwords($profile->getDirection());
$success = "{$direction} with id #{$batchModel->getId()} completed succesfully.\n";
Mage::log($success);
print $success;

This script was based on a script originally written for Magento 1.4. Exports are tested and work properly on 1.5, but importing doesn't seem to work. Later versions not tested.

From the looks of it, imports only setup a batch system which splits the import up into chunks. Those chunks still need to be run separately. Stackoverflow has a way to do that, but I don't like the fact it's actually using curl (thus the webserver) for this purpose.