public function test_getPurgeDataSettings_shouldAlsoUseOptionValuesIfUIisEnabled()
{
$this->setUIEnabled(true);
$settings = $this->manager->getPurgeDataSettings();
$expected = $this->getDefaultPurgeSettings();
$expected['delete_logs_enable'] = 1;
$expected['delete_logs_older_than'] = 270;
$expected['delete_reports_keep_week_reports'] = 1;
$this->assertEquals($expected, $settings);
}
/**
* Tests that purgeData works correctly when the 'keep segment reports' setting is set to true.
*
* @group Integration
*/
public function testPurgeDataDeleteReportsKeepSegmentsReports()
{
PrivacyManager::savePurgeDataSettings(array('delete_reports_keep_day_reports' => 1, 'delete_reports_keep_segment_reports' => 1));
// get purge data prediction
$prediction = PrivacyManager::getPurgeEstimate();
// perform checks on prediction
$events = 3;
// only the event action for the three purged day, dayAgo=x are purged (others are still in use)
$expectedPrediction = array(Common::prefixTable('log_conversion') => 6, Common::prefixTable('log_link_visit_action') => 6 + $events, Common::prefixTable('log_visit') => 3, Common::prefixTable('log_conversion_item') => 3, Common::prefixTable('archive_numeric_2012_01') => -1, Common::prefixTable('archive_blob_2012_01') => 9);
$this->assertEquals($expectedPrediction, $prediction);
// purge data
$this->_setTimeToRun();
$this->assertTrue($this->instance->deleteLogData());
$this->assertTrue($this->instance->deleteReportData());
// perform checks
$this->checkLogDataPurged();
$this->_checkReportsAndMetricsPurged($janBlobsRemaining = 6, $janNumericRemaining = 70);
// 1 segmented blob + 5 day blobs
}
public function redirectDashboardToWelcomePage(&$module, &$action)
{
if ($module !== 'CoreHome' || $action !== 'index') {
return;
}
$siteId = Common::getRequestVar('idSite', false, 'int');
if (!$siteId) {
return;
}
// Skip the screen if purging logs is enabled
$settings = PrivacyManager::getPurgeDataSettings();
if ($settings['delete_logs_enable'] == 1) {
return;
}
$trackerModel = new TrackerModel();
if ($trackerModel->isSiteEmpty($siteId)) {
$module = 'SitesManager';
$action = 'siteWithoutData';
}
}
/**
* When tracking data in the past (using Tracking API), this function
* can be used to invalidate reports for the idSites and dates where new data
* was added.
* DEV: If you call this API, the UI should display the data correctly, but will process
* in real time, which could be very slow after large data imports.
* After calling this function via REST, you can manually force all data
* to be reprocessed by visiting the script as the Super User:
* http://example.net/piwik/misc/cron/archive.php?token_auth=$SUPER_USER_TOKEN_AUTH_HERE
* REQUIREMENTS: On large piwik setups, you will need in PHP configuration: max_execution_time = 0
* We recommend to use an hourly schedule of the script.
* More information: http://piwik.org/setup-auto-archiving/
*
* @param string $idSites Comma separated list of idSite that have had data imported for the specified dates
* @param string $dates Comma separated list of dates to invalidate for all these websites
* @throws Exception
* @return array
*/
public function invalidateArchivedReports($idSites, $dates)
{
$idSites = Site::getIdSitesFromIdSitesString($idSites);
if (empty($idSites)) {
throw new Exception("Specify a value for &idSites= as a comma separated list of website IDs, for which your token_auth has 'admin' permission");
}
Piwik::checkUserHasAdminAccess($idSites);
// Ensure the specified dates are valid
$toInvalidate = $invalidDates = array();
$dates = explode(',', trim($dates));
$dates = array_unique($dates);
foreach ($dates as $theDate) {
$theDate = trim($theDate);
try {
$date = Date::factory($theDate);
} catch (Exception $e) {
$invalidDates[] = $theDate;
continue;
}
if ($date->toString() == $theDate) {
$toInvalidate[] = $date;
} else {
$invalidDates[] = $theDate;
}
}
// If using the feature "Delete logs older than N days"...
$purgeDataSettings = PrivacyManager::getPurgeDataSettings();
$logsAreDeletedBeforeThisDate = $purgeDataSettings['delete_logs_schedule_lowest_interval'];
$logsDeleteEnabled = $purgeDataSettings['delete_logs_enable'];
$minimumDateWithLogs = false;
if ($logsDeleteEnabled && $logsAreDeletedBeforeThisDate) {
$minimumDateWithLogs = Date::factory('today')->subDay($logsAreDeletedBeforeThisDate);
}
// Given the list of dates, process which tables they should be deleted from
$minDate = false;
$warningDates = $processedDates = array();
/* @var $date Date */
foreach ($toInvalidate as $date) {
// we should only delete reports for dates that are more recent than N days
if ($minimumDateWithLogs && $date->isEarlier($minimumDateWithLogs)) {
$warningDates[] = $date->toString();
} else {
$processedDates[] = $date->toString();
}
$month = $date->toString('Y_m');
// For a given date, we must invalidate in the monthly archive table
$datesByMonth[$month][] = $date->toString();
// But also the year stored in January
$year = $date->toString('Y_01');
$datesByMonth[$year][] = $date->toString();
// but also weeks overlapping several months stored in the month where the week is starting
/* @var $week Week */
$week = Period\Factory::build('week', $date);
$weekAsString = $week->getDateStart()->toString('Y_m');
$datesByMonth[$weekAsString][] = $date->toString();
// Keep track of the minimum date for each website
if ($minDate === false || $date->isEarlier($minDate)) {
$minDate = $date;
}
}
if (empty($minDate)) {
throw new Exception("Check the 'dates' parameter is a valid date.");
}
// In each table, invalidate day/week/month/year containing this date
$archiveTables = ArchiveTableCreator::getTablesArchivesInstalled();
foreach ($archiveTables as $table) {
// Extract Y_m from table name
$suffix = ArchiveTableCreator::getDateFromTableName($table);
if (!isset($datesByMonth[$suffix])) {
continue;
}
// Dates which are to be deleted from this table
$datesToDeleteInTable = $datesByMonth[$suffix];
// Build one statement to delete all dates from the given table
$sql = $bind = array();
$datesToDeleteInTable = array_unique($datesToDeleteInTable);
foreach ($datesToDeleteInTable as $dateToDelete) {
$sql[] = '(date1 <= ? AND ? <= date2)';
$bind[] = $dateToDelete;
$bind[] = $dateToDelete;
}
$sql = implode(" OR ", $sql);
$query = "DELETE FROM {$table} " . " WHERE ( {$sql} ) " . " AND idsite IN (" . implode(",", $idSites) . ")";
Db::query($query, $bind);
}
\Piwik\Plugins\SitesManager\API::getInstance()->updateSiteCreatedTime($idSites, $minDate);
// Force to re-process data for these websites in the next cron core:archive command run
$invalidatedIdSites = self::getWebsiteIdsToInvalidate();
$invalidatedIdSites = array_merge($invalidatedIdSites, $idSites);
$invalidatedIdSites = array_unique($invalidatedIdSites);
$invalidatedIdSites = array_values($invalidatedIdSites);
Option::set(self::OPTION_INVALIDATED_IDSITES, serialize($invalidatedIdSites));
Site::clearCache();
$output = array();
// output logs
if ($warningDates) {
$output[] = 'Warning: the following Dates have not been invalidated, because they are earlier than your Log Deletion limit: ' . implode(", ", $warningDates) . "\n The last day with logs is " . $minimumDateWithLogs . ". " . "\n Please disable 'Delete old Logs' or set it to a higher deletion threshold (eg. 180 days or 365 years).'.";
}
$output[] = "Success. The following dates were invalidated successfully: " . implode(", ", $processedDates);
return $output;
}
/**
* Returns true if it is likely that the data for this report has been purged and if the
* user should be told about that.
*
* In order for this function to return true, the following must also be true:
* - The data table for this report must either be empty or not have been fetched.
* - The period of this report is not a multiple period.
* - The date of this report must be older than the delete_reports_older_than config option.
* @return bool
*/
private function hasReportBeenPurged()
{
if (!$this->isPluginActivated('PrivacyManager')) {
return false;
}
return PrivacyManager::hasReportBeenPurged($this->dataTable);
}
protected function getDeleteDataInfo()
{
Piwik::checkUserHasSuperUserAccess();
$deleteDataInfos = array();
$taskScheduler = new TaskScheduler();
$deleteDataInfos["config"] = PrivacyManager::getPurgeDataSettings();
$deleteDataInfos["deleteTables"] = "<br/>" . implode(", ", LogDataPurger::getDeleteTableLogTables());
$scheduleTimetable = $taskScheduler->getScheduledTimeForMethod("PrivacyManager", "deleteLogTables");
$optionTable = Option::get(self::OPTION_LAST_DELETE_PIWIK_LOGS);
//If task was already rescheduled, read time from taskTimetable. Else, calculate next possible runtime.
if (!empty($scheduleTimetable) && $scheduleTimetable - time() > 0) {
$nextPossibleSchedule = (int) $scheduleTimetable;
} else {
$date = Date::factory("today");
$nextPossibleSchedule = $date->addDay(1)->getTimestamp();
}
//deletion schedule did not run before
if (empty($optionTable)) {
$deleteDataInfos["lastRun"] = false;
//next run ASAP (with next schedule run)
$date = Date::factory("today");
$deleteDataInfos["nextScheduleTime"] = $nextPossibleSchedule;
} else {
$deleteDataInfos["lastRun"] = $optionTable;
$deleteDataInfos["lastRunPretty"] = Date::factory((int) $optionTable)->getLocalized('%day% %shortMonth% %longYear%');
//Calculate next run based on last run + interval
$nextScheduleRun = (int) ($deleteDataInfos["lastRun"] + $deleteDataInfos["config"]["delete_logs_schedule_lowest_interval"] * 24 * 60 * 60);
//is the calculated next run in the past? (e.g. plugin was disabled in the meantime or something) -> run ASAP
if ($nextScheduleRun - time() <= 0) {
$deleteDataInfos["nextScheduleTime"] = $nextPossibleSchedule;
} else {
$deleteDataInfos["nextScheduleTime"] = $nextScheduleRun;
}
}
$deleteDataInfos["nextRunPretty"] = MetricsFormatter::getPrettyTimeFromSeconds($deleteDataInfos["nextScheduleTime"] - time());
return $deleteDataInfos;
}
/**
* Returns true if it is likely that the data for this report has been purged and if the
* user should be told about that.
*
* In order for this function to return true, the following must also be true:
* - The data table for this report must either be empty or not have been fetched.
* - The period of this report is not a multiple period.
* - The date of this report must be older than the delete_reports_older_than config option.
* @return bool
*/
private function hasReportBeenPurged()
{
if (!\Piwik\Plugin\Manager::getInstance()->isPluginActivated('PrivacyManager')) {
return false;
}
return PrivacyManager::hasReportBeenPurged($this->dataTable);
}
public function test_markArchivesAsInvalidated_DoesNotInvalidateDatesBeforePurgeThreshold()
{
PrivacyManager::savePurgeDataSettings(array('delete_logs_enable' => 1, 'delete_logs_older_than' => 180));
$dateBeforeThreshold = Date::factory('today')->subDay(190);
$thresholdDate = Date::factory('today')->subDay(180);
$dateAfterThreshold = Date::factory('today')->subDay(170);
// can't test more than day since today will change, causing the test to fail w/ other periods randomly
$this->insertArchiveRow(1, $dateBeforeThreshold, 'day');
$this->insertArchiveRow(1, $dateAfterThreshold, 'day');
/** @var ArchiveInvalidator $archiveInvalidator */
$archiveInvalidator = self::$fixture->piwikEnvironment->getContainer()->get('Piwik\\Archive\\ArchiveInvalidator');
$result = $archiveInvalidator->markArchivesAsInvalidated(array(1), array($dateBeforeThreshold, $dateAfterThreshold), 'day');
$this->assertEquals($thresholdDate->toString(), $result->minimumDateWithLogs);
$expectedProcessedDates = array($dateAfterThreshold->toString());
$this->assertEquals($expectedProcessedDates, $result->processedDates);
$expectedWarningDates = array($dateBeforeThreshold->toString());
$this->assertEquals($expectedWarningDates, $result->warningDates);
$invalidatedArchives = $this->getInvalidatedIdArchives();
$countInvalidatedArchives = 0;
foreach ($invalidatedArchives as $idarchives) {
$countInvalidatedArchives += count($idarchives);
}
$this->assertEquals(1, $countInvalidatedArchives);
}
private function findOlderDateWithLogs()
{
// If using the feature "Delete logs older than N days"...
$purgeDataSettings = PrivacyManager::getPurgeDataSettings();
$logsDeletedWhenOlderThanDays = $purgeDataSettings['delete_logs_older_than'];
$logsDeleteEnabled = $purgeDataSettings['delete_logs_enable'];
if ($logsDeleteEnabled && $logsDeletedWhenOlderThanDays) {
$this->minimumDateWithLogs = Date::factory('today')->subDay($logsDeletedWhenOlderThanDays);
}
}
private function purgeData($deleteReportsOlderThan, $reportPeriodsToKeep, $keepBasicMetrics)
{
$metricsToKeep = PrivacyManager::getAllMetricsToKeep();
$maxRowsToDeletePerQuery = 100000;
$keepSegmentReports = false;
$purger = new ReportsPurger($deleteReportsOlderThan, $keepBasicMetrics, $reportPeriodsToKeep, $keepSegmentReports, $metricsToKeep, $maxRowsToDeletePerQuery);
$purger->purgeData();
}
private function savePurgeDataSettings($settings)
{
Piwik::checkUserHasSuperUserAccess();
$this->checkDataPurgeAdminSettingsIsEnabled();
PrivacyManager::savePurgeDataSettings($settings);
return true;
}
