moodle/moodle
3
0
Fork 0
mirror of https://github.com/moodle/moodle.git synced 2025-08-08 10:26:40 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'MDL-59630_master' of git://github.com/dmonllao/moodle

Browse source
This commit is contained in:
Andrew Nicols 2017-10-05 14:56:57 +08:00
commit 1afd3c72aa
17 changed files with 649 additions and 139 deletions
Download patch file Download diff file Expand all files Collapse all files

68
analytics/classes/local/analyser/base.php
View file

@ -112,6 +112,16 @@ abstract class base {
$this->log = array();
}
/**
* Returns the list of analysable elements available on the site.
*
* \core_analytics\local\analyser\by_course and \core_analytics\local\analyser\sitewide are implementing
* this method returning site courses (by_course) and the whole system (sitewide) as analysables.
*
* @return \core_analytics\analysable[]
*/
abstract public function get_analysables();
/**
* This function returns this analysable list of samples.
*
@ -141,7 +151,7 @@ abstract class base {
*
* @return string
*/
abstract protected function get_samples_origin();
abstract public function get_samples_origin();
/**
* Returns the context of a sample.
@ -166,15 +176,29 @@ abstract class base {
/**
* Main analyser method which processes the site analysables.
*
* \core_analytics\local\analyser\by_course and \core_analytics\local\analyser\sitewide are implementing
* this method returning site courses (by_course) and the whole system (sitewide) as analysables.
* In most of the cases you should have enough extending from one of these classes so you don't need
* to reimplement this method.
*
* @param bool $includetarget
* @return \stored_file[]
*/
abstract public function get_analysable_data($includetarget);
public function get_analysable_data($includetarget) {
$filesbytimesplitting = array();
$analysables = $this->get_analysables();
foreach ($analysables as $analysable) {
$files = $this->process_analysable($analysable, $includetarget);
// Later we will need to aggregate data by time splitting method.
foreach ($files as $timesplittingid => $file) {
$filesbytimesplitting[$timesplittingid][$analysable->get_id()] = $file;
}
}
// We join the datasets by time splitting method.
$timesplittingfiles = $this->merge_analysable_files($filesbytimesplitting, $includetarget);
return $timesplittingfiles;
}
/**
* Samples data this analyser provides.
@ -220,6 +244,36 @@ abstract class base {
}
}
/**
* Merges analysable dataset files into 1.
*
* @param array $filesbytimesplitting
* @param bool $includetarget
* @return \stored_file[]
*/
protected function merge_analysable_files($filesbytimesplitting, $includetarget) {
$timesplittingfiles = array();
foreach ($filesbytimesplitting as $timesplittingid => $files) {
if ($this->options['evaluation'] === true) {
// Delete the previous copy. Only when evaluating.
\core_analytics\dataset_manager::delete_previous_evaluation_file($this->modelid, $timesplittingid);
}
// Merge all course files into one.
if ($includetarget) {
$filearea = \core_analytics\dataset_manager::LABELLED_FILEAREA;
} else {
$filearea = \core_analytics\dataset_manager::UNLABELLED_FILEAREA;
}
$timesplittingfiles[$timesplittingid] = \core_analytics\dataset_manager::merge_datasets($files,
$this->modelid, $timesplittingid, $filearea, $this->options['evaluation']);
}
return $timesplittingfiles;
}
/**
* Checks that this analyser satisfies the provided indicator requirements.
*

64
analytics/classes/local/analyser/by_course.php
View file

@ -40,14 +40,14 @@ abstract class by_course extends base {
*
* @return \core_analytics\course[]
*/
public function get_courses() {
public function get_analysables() {
// Default to all system courses.
if (!empty($this->options['filter'])) {
$courses = $this->options['filter'];
} else {
// Iterate through all potentially valid courses.
$courses = get_courses();
$courses = get_courses('all', 'c.sortorder ASC');
}
unset($courses[SITEID]);
@ -55,7 +55,7 @@ abstract class by_course extends base {
foreach ($courses as $course) {
// Skip the frontpage course.
$analysable = \core_analytics\course::instance($course);
$analysables[$analysable->get_id()] = $analysable;
$analysables[] = $analysable;
}
if (empty($analysables)) {
@ -64,62 +64,4 @@ abstract class by_course extends base {
return $analysables;
}
/**
* Returns the analysed data
*
* @param bool $includetarget
* @return \stored_file[]
*/
public function get_analysable_data($includetarget) {
$filesbytimesplitting = array();
// This class and all children will iterate through a list of courses (\core_analytics\course).
$analysables = $this->get_courses('all', 'c.sortorder ASC');
foreach ($analysables as $analysableid => $analysable) {
$files = $this->process_analysable($analysable, $includetarget);
// Later we will need to aggregate data by time splitting method.
foreach ($files as $timesplittingid => $file) {
$filesbytimesplitting[$timesplittingid][$analysableid] = $file;
}
}
// We join the datasets by time splitting method.
$timesplittingfiles = $this->merge_analysable_files($filesbytimesplitting, $includetarget);
return $timesplittingfiles;
}
/**
* Merges analysable dataset files into 1.
*
* @param array $filesbytimesplitting
* @param bool $includetarget
* @return \stored_file[]
*/
protected function merge_analysable_files($filesbytimesplitting, $includetarget) {
$timesplittingfiles = array();
foreach ($filesbytimesplitting as $timesplittingid => $files) {
if ($this->options['evaluation'] === true) {
// Delete the previous copy. Only when evaluating.
\core_analytics\dataset_manager::delete_previous_evaluation_file($this->modelid, $timesplittingid);
}
// Merge all course files into one.
if ($includetarget) {
$filearea = \core_analytics\dataset_manager::LABELLED_FILEAREA;
} else {
$filearea = \core_analytics\dataset_manager::UNLABELLED_FILEAREA;
}
$timesplittingfiles[$timesplittingid] = \core_analytics\dataset_manager::merge_datasets($files,
$this->modelid, $timesplittingid, $filearea, $this->options['evaluation']);
}
return $timesplittingfiles;
}
}

32
analytics/classes/local/analyser/sitewide.php
View file

@ -36,36 +36,12 @@ defined('MOODLE_INTERNAL') || die();
abstract class sitewide extends base {
/**
* Returns the analysable data.
* Returns one single analysable element, the site.
*
* @param bool $includetarget
* @return \stored_file[] One file for each time splitting method.
* @return \core_analytics\analysable[]
*/
public function get_analysable_data($includetarget) {
// Here there is a single analysable and it is the system.
public function get_analysables() {
$analysable = new \core_analytics\site();
$files = $this->process_analysable($analysable, $includetarget);
// Copy to range files as there is just one analysable.
foreach ($files as $timesplittingid => $file) {
if ($this->options['evaluation'] === true) {
// Delete the previous copy. Only when evaluating.
\core_analytics\dataset_manager::delete_previous_evaluation_file($this->modelid, $timesplittingid);
}
// We use merge but it is just a copy.
if ($includetarget) {
$filearea = \core_analytics\dataset_manager::LABELLED_FILEAREA;
} else {
$filearea = \core_analytics\dataset_manager::UNLABELLED_FILEAREA;
}
$files[$timesplittingid] = \core_analytics\dataset_manager::merge_datasets(array($file), $this->modelid,
$timesplittingid, $filearea, $this->options['evaluation']);
}
return $files;
return array($analysable);
}
}

48
analytics/classes/manager.php
View file

@ -489,6 +489,54 @@ class manager {
}
}
/**
* Cleans up analytics db tables that do not directly depend on analysables that may have been deleted.
*/
public static function cleanup() {
global $DB;
// Clean up stuff that depends on contexts that do not exist anymore.
$sql = "SELECT DISTINCT ap.contextid FROM {analytics_predictions} ap
LEFT JOIN {context} ctx ON ap.contextid = ctx.id
WHERE ctx.id IS NULL";
$apcontexts = $DB->get_records_sql($sql);
$sql = "SELECT DISTINCT aic.contextid FROM {analytics_indicator_calc} aic
LEFT JOIN {context} ctx ON aic.contextid = ctx.id
WHERE ctx.id IS NULL";
$indcalccontexts = $DB->get_records_sql($sql);
$contexts = $apcontexts + $indcalccontexts;
if ($contexts) {
list($sql, $params) = $DB->get_in_or_equal(array_keys($contexts));
$DB->execute("DELETE FROM {analytics_prediction_actions} apa WHERE apa.predictionid IN
(SELECT ap.id FROM {analytics_predictions} ap WHERE ap.contextid $sql)", $params);
$DB->delete_records_select('analytics_predictions', "contextid $sql", $params);
$DB->delete_records_select('analytics_indicator_calc', "contextid $sql", $params);
}
// Clean up stuff that depends on analysable ids that do not exist anymore.
$models = self::get_all_models();
foreach ($models as $model) {
$analyser = $model->get_analyser(array('notimesplitting' => true));
$analysables = $analyser->get_analysables();
if (!$analysables) {
continue;
}
$analysableids = array_map(function($analysable) {
return $analysable->get_id();
}, $analysables);
list($notinsql, $params) = $DB->get_in_or_equal($analysableids, SQL_PARAMS_NAMED, 'param', false);
$params['modelid'] = $model->get_id();
$DB->delete_records_select('analytics_predict_samples', "modelid = :modelid AND analysableid $notinsql", $params);
$DB->delete_records_select('analytics_train_samples', "modelid = :modelid AND analysableid $notinsql", $params);
}
}
/**
* Returns the provided element classes in the site.
*

109
analytics/classes/model.php
View file

@ -247,15 +247,15 @@ class model {
/**
* Returns the model analyser (defined by the model target).
*
* @param array $options Default initialisation with no options.
* @return \core_analytics\local\analyser\base
*/
public function get_analyser() {
public function get_analyser($options = array()) {
if ($this->analyser !== null) {
return $this->analyser;
}
// Default initialisation with no options.
$this->init_analyser();
$this->init_analyser($options);
return $this->analyser;
}
@ -276,26 +276,29 @@ class model {
throw new \moodle_exception('errornotarget', 'analytics');
}
if (!empty($options['evaluation'])) {
// The evaluation process will run using all available time splitting methods unless one is specified.
if (!empty($options['timesplitting'])) {
$timesplitting = \core_analytics\manager::get_time_splitting($options['timesplitting']);
$timesplittings = array($timesplitting->get_id() => $timesplitting);
$timesplittings = array();
if (empty($options['notimesplitting'])) {
if (!empty($options['evaluation'])) {
// The evaluation process will run using all available time splitting methods unless one is specified.
if (!empty($options['timesplitting'])) {
$timesplitting = \core_analytics\manager::get_time_splitting($options['timesplitting']);
$timesplittings = array($timesplitting->get_id() => $timesplitting);
} else {
$timesplittings = \core_analytics\manager::get_enabled_time_splitting_methods();
}
} else {
$timesplittings = \core_analytics\manager::get_enabled_time_splitting_methods();
}
} else {
if (empty($this->model->timesplitting)) {
throw new \moodle_exception('invalidtimesplitting', 'analytics', '', $this->model->id);
if (empty($this->model->timesplitting)) {
throw new \moodle_exception('invalidtimesplitting', 'analytics', '', $this->model->id);
}
// Returned as an array as all actions (evaluation, training and prediction) go through the same process.
$timesplittings = array($this->model->timesplitting => $this->get_time_splitting());
}
// Returned as an array as all actions (evaluation, training and prediction) go through the same process.
$timesplittings = array($this->model->timesplitting => $this->get_time_splitting());
}
if (empty($timesplittings)) {
throw new \moodle_exception('errornotimesplittings', 'analytics');
if (empty($timesplittings)) {
throw new \moodle_exception('errornotimesplittings', 'analytics');
}
}
if (!empty($options['evaluation'])) {
@ -432,17 +435,20 @@ class model {
if ($this->model->timesplitting !== $timesplittingid ||
$this->model->indicators !== $indicatorsstr) {
// Delete generated predictions before changing the model version.
$this->clear_model();
// It needs to be reset as the version changes.
$this->uniqueid = null;
// We update the version of the model so different time splittings are not mixed up.
$this->model->version = $now;
// Delete generated predictions.
$this->clear_model();
// Purge all generated files.
\core_analytics\dataset_manager::clear_model_files($this->model->id);
// Reset trained flag.
$this->model->trained = 0;
if (!$this->is_static()) {
$this->model->trained = 0;
}
} else if ($this->model->enabled != $enabled) {
// We purge the cached contexts with insights as some will not be visible anymore.
@ -456,9 +462,6 @@ class model {
$this->model->usermodified = $USER->id;
$DB->update_record('analytics_models', $this->model);
// It needs to be reset (just in case, we may already used it).
$this->uniqueid = null;
}
/**
@ -472,7 +475,13 @@ class model {
\core_analytics\manager::check_can_manage_models();
$this->clear_model();
// Method self::clear_model is already clearing the current model version.
$predictor = \core_analytics\manager::get_predictions_processor();
$predictor->delete_output_dir($this->get_output_dir(array(), true));
$DB->delete_records('analytics_models', array('id' => $this->model->id));
$DB->delete_records('analytics_models_log', array('modelid' => $this->model->id));
}
/**
@ -973,13 +982,23 @@ class model {
throw new \moodle_exception('errorinvalidtimesplitting', 'analytics');
}
// Delete generated predictions before changing the model version.
$this->clear_model();
// It needs to be reset as the version changes.
$this->uniqueid = null;
$this->model->timesplitting = $timesplittingid;
$this->model->version = $now;
// Reset trained flag.
if (!$this->is_static()) {
$this->model->trained = 0;
}
}
// Purge pages with insights as this may change things.
if ($timesplittingid && $timesplittingid !== $this->model->timesplitting ||
$this->model->enabled != 1) {
if ($this->model->enabled != 1) {
$this->purge_insights_cache();
}
@ -988,9 +1007,6 @@ class model {
// We don't always update timemodified intentionally as we reserve it for target, indicators or timesplitting updates.
$DB->update_record('analytics_models', $this->model);
// It needs to be reset (just in case, we may already used it).
$this->uniqueid = null;
}
/**
@ -1228,9 +1244,10 @@ class model {
* models/$model->id/$model->version/execution
*
* @param array $subdirs
* @param bool $onlymodelid Preference over $subdirs
* @return string
*/
protected function get_output_dir($subdirs = array()) {
protected function get_output_dir($subdirs = array(), $onlymodelid = false) {
global $CFG;
$subdirstr = '';
@ -1244,8 +1261,12 @@ class model {
$outputdir = rtrim($CFG->dataroot, '/') . DIRECTORY_SEPARATOR . 'models';
}
// Append model id and version + subdirs.
$outputdir .= DIRECTORY_SEPARATOR . $this->model->id . DIRECTORY_SEPARATOR . $this->model->version . $subdirstr;
// Append model id
$outputdir .= DIRECTORY_SEPARATOR . $this->model->id;
if (!$onlymodelid) {
// Append version + subdirs.
$outputdir .= DIRECTORY_SEPARATOR . $this->model->version . $subdirstr;
}
make_writable_directory($outputdir);
@ -1410,11 +1431,25 @@ class model {
private function clear_model() {
global $DB;
// Delete current model version stored stuff.
$predictor = \core_analytics\manager::get_predictions_processor();
$predictor->clear_model($this->get_unique_id(), $this->get_output_dir());
$predictionids = $DB->get_fieldset_select('analytics_predictions', 'id', 'modelid = :modelid',
array('modelid' => $this->get_id()));
if ($predictionids) {
list($sql, $params) = $DB->get_in_or_equal($predictionids);
$DB->delete_records_select('analytics_prediction_actions', "predictionid $sql", $params);
}
$DB->delete_records('analytics_predictions', array('modelid' => $this->model->id));
$DB->delete_records('analytics_predict_samples', array('modelid' => $this->model->id));
$DB->delete_records('analytics_train_samples', array('modelid' => $this->model->id));
$DB->delete_records('analytics_used_files', array('modelid' => $this->model->id));
// Purge all generated files.
\core_analytics\dataset_manager::clear_model_files($this->model->id);
// We don't expect people to clear models regularly and the cost of filling the cache is
// 1 db read per context.
$this->purge_insights_cache();

2
analytics/classes/prediction.php
View file

@ -68,7 +68,7 @@ class prediction {
/**
* Constructor
*
* @param \stdClass $prediction
* @param \stdClass|int $prediction
* @param array $sampledata
* @return void
*/

33
analytics/classes/predictor.php
View file

@ -41,4 +41,37 @@ interface predictor {
* @return bool
*/
public function is_ready();
/**
* Delete all stored information of the current model id.
*
* This method is called when there are important changes to a model,
* all previous training algorithms using that version of the model
* should be deleted.
*
* In case you want to perform extra security measures before deleting
* a directory you can check that $modelversionoutputdir subdirectories
* can only be named 'execution', 'evaluation' or 'testing'.
*
* @param string $uniqueid The site model unique id string
* @param string $modelversionoutputdir The output dir of this model version
* @return null
*/
public function clear_model($uniqueid, $modelversionoutputdir);
/**
* Delete the output directory.
*
* This method is called when a model is completely deleted.
*
* In case you want to perform extra security measures before deleting
* a directory you can check that the subdirectories are timestamps
* (the model version) and each of this subdirectories' subdirectories
* can only be named 'execution', 'evaluation' or 'testing'.
*
* @param string $modeloutputdir The model directory id (parent of all model versions subdirectories).
* @return null
*/
public function delete_output_dir($modeloutputdir);
}