fixes #692 - plugin (deactivated by default) to anonymize visitor IP...

fixes #692 - plugin (deactivated by default) to anonymize visitor IP addresses; the number of octets to mask is configurable; let me know if I've missed any edge cases in the unit tests


git-svn-id: http://dev.piwik.org/svn/trunk@1877 59fd770c-687e-43c8-a1e3-f5a4ff64c105

config/global.ini.php

@@ -199,6 +199,9 @@ campaign_keyword_var_name	= piwik_kwd
 ; maximum length of a Page Title or a Page URL recorded in the log_action.name table
 page_maximum_length = 1024;
 
+; number of octets in IP address to mask, in order to anonymize a visitor's IP address; if the AnonymizeIP plugin is deactivated, this value is ignored; for IPv4 addresses, valid values are 0..4
+ip_address_mask_length = 1
+
 [log]
 ;possible values for log: screen, database, file
 ; normal messages

lang/en.php

@@ -128,6 +128,7 @@ $translations = array(
 	'Actions_ColumnPageURL' => 'Page URL',
 	'Actions_ColumnClickedURL' => 'Clicked URL',
 	'Actions_ColumnDownloadURL' => 'Download URL',
+	'AnonymizeIP_PluginDescription' => 'Anonymize visitor IP addresses to comply with your local privacy laws/guidelines.',
 	'API_PluginDescription' => 'All the data in Piwik is available through simple APIs. This plugin is the web service entry point, that you can call to get your Web Analytics data in xml, json, php, csv, etc.',
 	'API_QuickDocumentation' => '<h2>API quick documentation</h2><p>If you don\'t have data for today you can first <a href=\'misc/generateVisits.php\' target=_blank>generate some data</a> using the Visits Generator script.</p><p>You can try the different formats available for every method. It is very easy to extract any data you want from Piwik!</p><p><b>For more information have a look at the <a href=\'http://dev.piwik.org/trac/wiki/API\'>official API Documentation</a> or the <a href=\'http://dev.piwik.org/trac/wiki/API/Reference\'>API Reference</a>.</b></P><h2>User authentication</h2><p>If you want to <b>request the data in your scripts, in a crontab, etc. </b> you need to add the parameter <code><u>&amp;token_auth=%s</u></code> to the API calls URLs that require authentication.</p><p>This token_auth is as secret as your login and password, <b>do not share it!</p>',
 	'API_LoadedAPIs' => 'Loaded successfully %s APIs',

plugins/AnonymizeIP/AnonymizeIP.php

+<?php
+/**
+ * Piwik - Open source web analytics
+ *
+ * @link http://piwik.org
+ * @license http://www.gnu.org/licenses/gpl-3.0.html Gpl v3 or later
+ * @version $Id$
+ *
+ * @category Piwik_Plugins
+ * @package Piwik_AnonymizeIP
+ */
+
+/**
+ * Anonymize visitor IP addresses to comply with the privacy laws/guidelines in countries, such as Germany.
+ *
+ * @package Piwik_AnonymizeIP
+ */
+class Piwik_AnonymizeIP extends Piwik_Plugin
+{
+	/**
+	 * Get plugin information
+	 */
+	public function getInformation()
+	{
+		return array(
+			'name' => 'AnonymizeIP',
+			'description' => Piwik_Translate('AnonymizeIP_PluginDescription'),
+			'author' => 'Piwik',
+			'author_homepage' => 'http://piwik.org/',
+			'version' => Piwik_Version::VERSION,
+			'TrackerPlugin' => true,
+		);
+	}
+
+	/**
+	 * Get list of hooks to register
+	 */
+	public function getListHooksRegistered()
+	{
+		return array(
+			'Tracker.saveVisitorInformation' => 'anonymizeVisitorIpAddress',
+		);
+	}
+
+	/**
+	 * Internal function to mask portions of the visitor IP address
+	 *
+	 * @param $ip Unsigned long representation of IP address
+	 * @param $maskLength Number of octets to reset
+	 */
+	static public function applyIPMask($ip, $maskLength)
+	{
+		$maskedIP = pack('V', (float)$ip);
+
+		switch($maskLength) {
+			case 4:
+				$maskedIP[3] = "\0";
+			case 3:
+				$maskedIP[2] = "\0";
+			case 2:
+				$maskedIP[1] = "\0";
+			case 1:
+				$maskedIP[0] = "\0";
+			case 0:
+			default:
+		}
+
+		$res = unpack('V', $maskedIP);
+		return sprintf("%u", $res[1]);
+	}
+
+	/**
+	 * Hook on Tracker.saveVisitorInformation to anonymize visitor IP addresses
+	 */
+	function anonymizeVisitorIpAddress($notification)
+	{
+		$visitorInfo =& $notification->getNotificationObject();
+		$visitorInfo['location_ip'] = self::applyIPMask($visitorInfo['location_ip'], Piwik_Tracker_Config::getInstance()->Tracker['ip_address_mask_length']);
+	}
+}

plugins/AnonymizeIP/tests/AnonymizeIP.test.php

+<?php
+if(!defined("PIWIK_PATH_TEST_TO_ROOT")) {
+	define('PIWIK_PATH_TEST_TO_ROOT', getcwd().'/../../..');
+}
+if(!defined('PIWIK_CONFIG_TEST_INCLUDED'))
+{
+	require_once PIWIK_PATH_TEST_TO_ROOT . "/tests/config_test.php";
+}
+
+if(!class_exists('Piwik_AnonymizeIP', false))
+{
+	require_once dirname(__FILE__) . '/../AnonymizeIP.php';
+}
+
+class Test_Piwik_AnonymizeIP extends UnitTestCase
+{
+	// IP addresses and expected results
+	protected $ipAddresses = array(
+		// long => array( expected0, expected1, expected2, expected3, expected4 ),
+		'0'   => array( 0, 0, 0, 0, 0 ),													// 00 00 00 00
+		'1'   => array( 1, 0, 0, 0, 0 ),													// 00 00 00 01
+		'255' => array( 255, 0, 0, 0, 0 ),													// 00 00 00 FF
+		'256' => array( 256, 256, 0, 0, 0 ),												// 00 00 01 00
+		'257' => array( 257, 256, 0, 0, 0 ),												// 00 00 01 01
+		'65535' => array( 65535, 65280, 0, 0, 0),											// 00 00 FF FF
+		'65536' => array( 65536, 65536, 65536, 0, 0),										// 00 01 00 00
+		'65793' => array( 65793, 65792, 65536, 0, 0),										// 00 01 01 01
+		'16777215' => array( 16777215, 16776960, 16711680, 0, 0),							// 00 FF FF FF
+		'16777216' => array( 16777216, 16777216, 16777216, 16777216, 0),					// 01 00 00 00
+		'2147483647' => array( 2147483647, 2147483392, 2147418112, 2130706432, 0),			// 7F FF FF FF
+		'2147483648' => array( '2147483648', '2147483648', '2147483648', '2147483648', 0),	// 80 00 00 00
+		'4294967295' => array( '4294967295', '4294967040', '4294901760', '4278190080', 0),	// FF FF FF FF
+	);
+
+	public function test_applyIPMask()
+	{
+		foreach($this->ipAddresses as $ip => $expected)
+		{
+			// each IP is tested with 0 to 4 octets masked
+			for($maskLength = 0; $maskLength <= 4; $maskLength++)
+			{
+				$this->assertTrue( Piwik_AnonymizeIP::applyIPMask($ip, $maskLength) == $expected[$maskLength] );
+			}
+		}
+	}
+}