David Monllao
0690a271c3
This was supposed to be split into multiple commits to make it easier to understand but I failed to do it properly. So this is the list of changes: - New analytics_indicator_calc db table to store indicators calculations - Reuse previous calculations during prediction/training; other models previous calculations should also be reused as long as they belong to the same sample (sampleid depends on sampleorigin), time range and indicator - Allow bulk inserting of these calculations as this can hurt database performance - Block the same analysable to be analysed for training and for prediction - Use a new instance of the target and use it for is_valid_* functions as using ::is_valid_sample can lead to problems if people uses it to cache stuff
457 lines
20 KiB
PHP
457 lines
20 KiB
PHP
<?php
|
|
// This file is part of Moodle - http://moodle.org/
|
|
//
|
|
// Moodle is free software: you can redistribute it and/or modify
|
|
// it under the terms of the GNU General Public License as published by
|
|
// the Free Software Foundation, either version 3 of the License, or
|
|
// (at your option) any later version.
|
|
//
|
|
// Moodle is distributed in the hope that it will be useful,
|
|
// but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
// GNU General Public License for more details.
|
|
//
|
|
// You should have received a copy of the GNU General Public License
|
|
// along with Moodle. If not, see <http://www.gnu.org/licenses/>.
|
|
|
|
/**
|
|
* Unit tests for evaluation, training and prediction.
|
|
*
|
|
* @package core_analytics
|
|
* @copyright 2017 David Monllaó {@link http://www.davidmonllao.com}
|
|
* @license http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
|
|
*/
|
|
|
|
defined('MOODLE_INTERNAL') || die();
|
|
|
|
require_once(__DIR__ . '/fixtures/test_indicator_max.php');
|
|
require_once(__DIR__ . '/fixtures/test_indicator_min.php');
|
|
require_once(__DIR__ . '/fixtures/test_indicator_fullname.php');
|
|
require_once(__DIR__ . '/fixtures/test_indicator_random.php');
|
|
require_once(__DIR__ . '/fixtures/test_target_shortname.php');
|
|
require_once(__DIR__ . '/fixtures/test_static_target_shortname.php');
|
|
|
|
require_once(__DIR__ . '/../../course/lib.php');
|
|
|
|
/**
|
|
* Unit tests for evaluation, training and prediction.
|
|
*
|
|
* @package core_analytics
|
|
* @copyright 2017 David Monllaó {@link http://www.davidmonllao.com}
|
|
* @license http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
|
|
*/
|
|
class core_analytics_prediction_testcase extends advanced_testcase {
|
|
|
|
/**
|
|
* test_static_prediction
|
|
*
|
|
* @return void
|
|
*/
|
|
public function test_static_prediction() {
|
|
global $DB;
|
|
|
|
$this->resetAfterTest(true);
|
|
$this->setAdminuser();
|
|
|
|
$model = $this->add_perfect_model('test_static_target_shortname');
|
|
$model->enable('\core\analytics\time_splitting\no_splitting');
|
|
$this->assertEquals(1, $model->is_enabled());
|
|
$this->assertEquals(1, $model->is_trained());
|
|
|
|
// No training for static models.
|
|
$results = $model->train();
|
|
$trainedsamples = $DB->get_records('analytics_train_samples', array('modelid' => $model->get_id()));
|
|
$this->assertEmpty($trainedsamples);
|
|
$this->assertEmpty($DB->count_records('analytics_used_files',
|
|
array('modelid' => $model->get_id(), 'action' => 'trained')));
|
|
|
|
// Now we create 2 hidden courses (only hidden courses are getting predictions).
|
|
$courseparams = array('shortname' => 'aaaaaa', 'fullname' => 'aaaaaa', 'visible' => 0);
|
|
$course1 = $this->getDataGenerator()->create_course($courseparams);
|
|
$courseparams = array('shortname' => 'bbbbbb', 'fullname' => 'bbbbbb', 'visible' => 0);
|
|
$course2 = $this->getDataGenerator()->create_course($courseparams);
|
|
|
|
$result = $model->predict();
|
|
|
|
// Var $course1 predictions should be 1 == 'a', $course2 predictions should be 0 == 'b'.
|
|
$correct = array($course1->id => 1, $course2->id => 0);
|
|
foreach ($result->predictions as $uniquesampleid => $predictiondata) {
|
|
list($sampleid, $rangeindex) = $model->get_time_splitting()->infer_sample_info($uniquesampleid);
|
|
|
|
// The range index is not important here, both ranges prediction will be the same.
|
|
$this->assertEquals($correct[$sampleid], $predictiondata->prediction);
|
|
}
|
|
|
|
// 1 range for each analysable.
|
|
$predictedranges = $DB->get_records('analytics_predict_samples', array('modelid' => $model->get_id()));
|
|
$this->assertCount(2, $predictedranges);
|
|
$this->assertEquals(1, $DB->count_records('analytics_used_files',
|
|
array('modelid' => $model->get_id(), 'action' => 'predicted')));
|
|
// 2 predictions for each range.
|
|
$this->assertEquals(2, $DB->count_records('analytics_predictions',
|
|
array('modelid' => $model->get_id())));
|
|
|
|
// No new generated files nor records as there are no new courses available.
|
|
$model->predict();
|
|
$predictedranges = $DB->get_records('analytics_predict_samples', array('modelid' => $model->get_id()));
|
|
$this->assertCount(2, $predictedranges);
|
|
$this->assertEquals(1, $DB->count_records('analytics_used_files',
|
|
array('modelid' => $model->get_id(), 'action' => 'predicted')));
|
|
$this->assertEquals(2, $DB->count_records('analytics_predictions',
|
|
array('modelid' => $model->get_id())));
|
|
}
|
|
|
|
/**
|
|
* test_ml_training_and_prediction
|
|
*
|
|
* @dataProvider provider_ml_training_and_prediction
|
|
* @param string $timesplittingid
|
|
* @param int $predictedrangeindex
|
|
* @param int $nranges
|
|
* @param string $predictionsprocessorclass
|
|
* @return void
|
|
*/
|
|
public function test_ml_training_and_prediction($timesplittingid, $predictedrangeindex, $nranges, $predictionsprocessorclass) {
|
|
global $DB;
|
|
|
|
$this->resetAfterTest(true);
|
|
$this->setAdminuser();
|
|
set_config('enabled_stores', 'logstore_standard', 'tool_log');
|
|
|
|
$ncourses = 10;
|
|
|
|
// Generate training data.
|
|
$params = array(
|
|
'startdate' => mktime(0, 0, 0, 10, 24, 2015),
|
|
'enddate' => mktime(0, 0, 0, 2, 24, 2016),
|
|
);
|
|
for ($i = 0; $i < $ncourses; $i++) {
|
|
$name = 'a' . random_string(10);
|
|
$courseparams = array('shortname' => $name, 'fullname' => $name) + $params;
|
|
$this->getDataGenerator()->create_course($courseparams);
|
|
}
|
|
for ($i = 0; $i < $ncourses; $i++) {
|
|
$name = 'b' . random_string(10);
|
|
$courseparams = array('shortname' => $name, 'fullname' => $name) + $params;
|
|
$this->getDataGenerator()->create_course($courseparams);
|
|
}
|
|
|
|
// We repeat the test for all prediction processors.
|
|
$predictionsprocessor = \core_analytics\manager::get_predictions_processor($predictionsprocessorclass, false);
|
|
if ($predictionsprocessor->is_ready() !== true) {
|
|
$this->markTestSkipped('Skipping ' . $predictionsprocessorclass . ' as the predictor is not ready.');
|
|
}
|
|
|
|
set_config('predictionsprocessor', $predictionsprocessorclass, 'analytics');
|
|
|
|
$model = $this->add_perfect_model();
|
|
$model->enable($timesplittingid);
|
|
|
|
// No samples trained yet.
|
|
$this->assertEquals(0, $DB->count_records('analytics_train_samples', array('modelid' => $model->get_id())));
|
|
|
|
$results = $model->train();
|
|
$this->assertEquals(1, $model->is_enabled());
|
|
$this->assertEquals(1, $model->is_trained());
|
|
|
|
// 20 courses * the 3 model indicators * the number of time ranges of this time splitting method.
|
|
$indicatorcalc = 20 * 3 * $nranges;
|
|
$this->assertEquals($indicatorcalc, $DB->count_records('analytics_indicator_calc'));
|
|
|
|
// 1 training file was created.
|
|
$trainedsamples = $DB->get_records('analytics_train_samples', array('modelid' => $model->get_id()));
|
|
$this->assertCount(1, $trainedsamples);
|
|
$samples = json_decode(reset($trainedsamples)->sampleids, true);
|
|
$this->assertCount($ncourses * 2, $samples);
|
|
$this->assertEquals(1, $DB->count_records('analytics_used_files',
|
|
array('modelid' => $model->get_id(), 'action' => 'trained')));
|
|
|
|
$courseparams = $params + array('shortname' => 'aaaaaa', 'fullname' => 'aaaaaa', 'visible' => 0);
|
|
$course1 = $this->getDataGenerator()->create_course($courseparams);
|
|
$courseparams = $params + array('shortname' => 'bbbbbb', 'fullname' => 'bbbbbb', 'visible' => 0);
|
|
$course2 = $this->getDataGenerator()->create_course($courseparams);
|
|
|
|
// They will not be skipped for prediction though.
|
|
$result = $model->predict();
|
|
|
|
// Var $course1 predictions should be 1 == 'a', $course2 predictions should be 0 == 'b'.
|
|
$correct = array($course1->id => 1, $course2->id => 0);
|
|
foreach ($result->predictions as $uniquesampleid => $predictiondata) {
|
|
list($sampleid, $rangeindex) = $model->get_time_splitting()->infer_sample_info($uniquesampleid);
|
|
|
|
// The range index is not important here, both ranges prediction will be the same.
|
|
$this->assertEquals($correct[$sampleid], $predictiondata->prediction);
|
|
}
|
|
|
|
// 1 range will be predicted.
|
|
$predictedranges = $DB->get_records('analytics_predict_samples', array('modelid' => $model->get_id()));
|
|
$this->assertCount(1, $predictedranges);
|
|
foreach ($predictedranges as $predictedrange) {
|
|
$this->assertEquals($predictedrangeindex, $predictedrange->rangeindex);
|
|
$sampleids = json_decode($predictedrange->sampleids, true);
|
|
$this->assertCount(2, $sampleids);
|
|
$this->assertContains($course1->id, $sampleids);
|
|
$this->assertContains($course2->id, $sampleids);
|
|
}
|
|
$this->assertEquals(1, $DB->count_records('analytics_used_files',
|
|
array('modelid' => $model->get_id(), 'action' => 'predicted')));
|
|
// 2 predictions.
|
|
$this->assertEquals(2, $DB->count_records('analytics_predictions',
|
|
array('modelid' => $model->get_id())));
|
|
|
|
// No new generated files nor records as there are no new courses available.
|
|
$model->predict();
|
|
$predictedranges = $DB->get_records('analytics_predict_samples', array('modelid' => $model->get_id()));
|
|
$this->assertCount(1, $predictedranges);
|
|
foreach ($predictedranges as $predictedrange) {
|
|
$this->assertEquals($predictedrangeindex, $predictedrange->rangeindex);
|
|
}
|
|
$this->assertEquals(1, $DB->count_records('analytics_used_files',
|
|
array('modelid' => $model->get_id(), 'action' => 'predicted')));
|
|
$this->assertEquals(2, $DB->count_records('analytics_predictions',
|
|
array('modelid' => $model->get_id())));
|
|
|
|
// New samples that can be used for prediction.
|
|
$courseparams = $params + array('shortname' => 'cccccc', 'fullname' => 'cccccc', 'visible' => 0);
|
|
$course3 = $this->getDataGenerator()->create_course($courseparams);
|
|
$courseparams = $params + array('shortname' => 'dddddd', 'fullname' => 'dddddd', 'visible' => 0);
|
|
$course4 = $this->getDataGenerator()->create_course($courseparams);
|
|
|
|
$result = $model->predict();
|
|
|
|
$predictedranges = $DB->get_records('analytics_predict_samples', array('modelid' => $model->get_id()));
|
|
$this->assertCount(1, $predictedranges);
|
|
foreach ($predictedranges as $predictedrange) {
|
|
$this->assertEquals($predictedrangeindex, $predictedrange->rangeindex);
|
|
$sampleids = json_decode($predictedrange->sampleids, true);
|
|
$this->assertCount(4, $sampleids);
|
|
$this->assertContains($course1->id, $sampleids);
|
|
$this->assertContains($course2->id, $sampleids);
|
|
$this->assertContains($course3->id, $sampleids);
|
|
$this->assertContains($course4->id, $sampleids);
|
|
}
|
|
$this->assertEquals(2, $DB->count_records('analytics_used_files',
|
|
array('modelid' => $model->get_id(), 'action' => 'predicted')));
|
|
$this->assertEquals(4, $DB->count_records('analytics_predictions',
|
|
array('modelid' => $model->get_id())));
|
|
|
|
// New visible course (for training).
|
|
$course5 = $this->getDataGenerator()->create_course(array('shortname' => 'aaa', 'fullname' => 'aa'));
|
|
$course6 = $this->getDataGenerator()->create_course();
|
|
$result = $model->train();
|
|
$this->assertEquals(2, $DB->count_records('analytics_used_files',
|
|
array('modelid' => $model->get_id(), 'action' => 'trained')));
|
|
|
|
// Update one of the courses to not visible, it should be used again for prediction.
|
|
$course5->visible = 0;
|
|
update_course($course5);
|
|
|
|
$model->predict();
|
|
$this->assertEquals(1, $DB->count_records('analytics_predict_samples',
|
|
array('modelid' => $model->get_id())));
|
|
$this->assertEquals(2, $DB->count_records('analytics_used_files',
|
|
array('modelid' => $model->get_id(), 'action' => 'predicted')));
|
|
$this->assertEquals(4, $DB->count_records('analytics_predictions',
|
|
array('modelid' => $model->get_id())));
|
|
|
|
set_config('enabled_stores', '', 'tool_log');
|
|
get_log_manager(true);
|
|
}
|
|
|
|
/**
|
|
* provider_ml_training_and_prediction
|
|
*
|
|
* @return array
|
|
*/
|
|
public function provider_ml_training_and_prediction() {
|
|
$cases = array(
|
|
'no_splitting' => array('\core\analytics\time_splitting\no_splitting', 0, 1),
|
|
'quarters' => array('\core\analytics\time_splitting\quarters', 3, 4)
|
|
);
|
|
|
|
// We need to test all system prediction processors.
|
|
return $this->add_prediction_processors($cases);
|
|
}
|
|
|
|
|
|
/**
|
|
* Basic test to check that prediction processors work as expected.
|
|
*
|
|
* @dataProvider provider_ml_test_evaluation
|
|
* @param string $modelquality
|
|
* @param int $ncourses
|
|
* @param array $expected
|
|
* @param string $predictionsprocessorclass
|
|
* @return void
|
|
*/
|
|
public function test_ml_evaluation($modelquality, $ncourses, $expected, $predictionsprocessorclass) {
|
|
$this->resetAfterTest(true);
|
|
$this->setAdminuser();
|
|
set_config('enabled_stores', 'logstore_standard', 'tool_log');
|
|
|
|
$sometimesplittings = '\core\analytics\time_splitting\weekly,' .
|
|
'\core\analytics\time_splitting\single_range,' .
|
|
'\core\analytics\time_splitting\quarters';
|
|
set_config('timesplittings', $sometimesplittings, 'analytics');
|
|
|
|
if ($modelquality === 'perfect') {
|
|
$model = $this->add_perfect_model();
|
|
} else if ($modelquality === 'random') {
|
|
$model = $this->add_random_model();
|
|
} else {
|
|
throw new \coding_exception('Only perfect and random accepted as $modelquality values');
|
|
}
|
|
|
|
// Generate training data.
|
|
$params = array(
|
|
'startdate' => mktime(0, 0, 0, 10, 24, 2015),
|
|
'enddate' => mktime(0, 0, 0, 2, 24, 2016),
|
|
);
|
|
for ($i = 0; $i < $ncourses; $i++) {
|
|
$name = 'a' . random_string(10);
|
|
$params = array('shortname' => $name, 'fullname' => $name) + $params;
|
|
$this->getDataGenerator()->create_course($params);
|
|
}
|
|
for ($i = 0; $i < $ncourses; $i++) {
|
|
$name = 'b' . random_string(10);
|
|
$params = array('shortname' => $name, 'fullname' => $name) + $params;
|
|
$this->getDataGenerator()->create_course($params);
|
|
}
|
|
|
|
// We repeat the test for all prediction processors.
|
|
$predictionsprocessor = \core_analytics\manager::get_predictions_processor($predictionsprocessorclass, false);
|
|
if ($predictionsprocessor->is_ready() !== true) {
|
|
$this->markTestSkipped('Skipping ' . $predictionsprocessorclass . ' as the predictor is not ready.');
|
|
}
|
|
|
|
set_config('predictionsprocessor', $predictionsprocessorclass, 'analytics');
|
|
|
|
$results = $model->evaluate();
|
|
|
|
// We check that the returned status includes at least $expectedcode code.
|
|
foreach ($results as $timesplitting => $result) {
|
|
$message = 'The returned status code ' . $result->status . ' should include ' . $expected[$timesplitting];
|
|
$filtered = $result->status & $expected[$timesplitting];
|
|
$this->assertEquals($expected[$timesplitting], $filtered, $message);
|
|
}
|
|
|
|
set_config('enabled_stores', '', 'tool_log');
|
|
get_log_manager(true);
|
|
}
|
|
|
|
/**
|
|
* test_read_indicator_calculations
|
|
*
|
|
* @return void
|
|
*/
|
|
public function test_read_indicator_calculations() {
|
|
global $DB;
|
|
|
|
$this->resetAfterTest(true);
|
|
|
|
$starttime = 123;
|
|
$endtime = 321;
|
|
$sampleorigin = 'whatever';
|
|
|
|
$indicator = $this->getMockBuilder('test_indicator_max')->setMethods(['calculate_sample'])->getMock();
|
|
$indicator->expects($this->never())->method('calculate_sample');
|
|
|
|
$existingcalcs = array(111 => 1, 222 => 0.5);
|
|
$sampleids = array(111 => 111, 222 => 222);
|
|
list($values, $unused) = $indicator->calculate($sampleids, $sampleorigin, $starttime, $endtime, $existingcalcs);
|
|
}
|
|
|
|
/**
|
|
* provider_ml_test_evaluation
|
|
*
|
|
* @return array
|
|
*/
|
|
public function provider_ml_test_evaluation() {
|
|
|
|
$cases = array(
|
|
'bad' => array(
|
|
'modelquality' => 'random',
|
|
'ncourses' => 50,
|
|
'expectedresults' => array(
|
|
// The course duration is too much to be processed by in weekly basis.
|
|
'\core\analytics\time_splitting\weekly' => \core_analytics\model::NO_DATASET,
|
|
'\core\analytics\time_splitting\single_range' => \core_analytics\model::EVALUATE_LOW_SCORE,
|
|
'\core\analytics\time_splitting\quarters' => \core_analytics\model::EVALUATE_LOW_SCORE,
|
|
)
|
|
),
|
|
'good' => array(
|
|
'modelquality' => 'perfect',
|
|
'ncourses' => 50,
|
|
'expectedresults' => array(
|
|
// The course duration is too much to be processed by in weekly basis.
|
|
'\core\analytics\time_splitting\weekly' => \core_analytics\model::NO_DATASET,
|
|
'\core\analytics\time_splitting\single_range' => \core_analytics\model::OK,
|
|
'\core\analytics\time_splitting\quarters' => \core_analytics\model::OK,
|
|
)
|
|
)
|
|
);
|
|
return $this->add_prediction_processors($cases);
|
|
}
|
|
|
|
/**
|
|
* add_random_model
|
|
*
|
|
* @return \core_analytics\model
|
|
*/
|
|
protected function add_random_model() {
|
|
|
|
$target = \core_analytics\manager::get_target('test_target_shortname');
|
|
$indicators = array('test_indicator_max', 'test_indicator_min', 'test_indicator_random');
|
|
foreach ($indicators as $key => $indicator) {
|
|
$indicators[$key] = \core_analytics\manager::get_indicator($indicator);
|
|
}
|
|
|
|
$model = \core_analytics\model::create($target, $indicators);
|
|
|
|
// To load db defaults as well.
|
|
return new \core_analytics\model($model->get_id());
|
|
}
|
|
|
|
/**
|
|
* add_perfect_model
|
|
*
|
|
* @param string $targetclass
|
|
* @return \core_analytics\model
|
|
*/
|
|
protected function add_perfect_model($targetclass = 'test_target_shortname') {
|
|
|
|
$target = \core_analytics\manager::get_target($targetclass);
|
|
$indicators = array('test_indicator_max', 'test_indicator_min', 'test_indicator_fullname');
|
|
foreach ($indicators as $key => $indicator) {
|
|
$indicators[$key] = \core_analytics\manager::get_indicator($indicator);
|
|
}
|
|
|
|
$model = \core_analytics\model::create($target, $indicators);
|
|
|
|
// To load db defaults as well.
|
|
return new \core_analytics\model($model->get_id());
|
|
}
|
|
|
|
/**
|
|
* add_prediction_processors
|
|
*
|
|
* @param array $cases
|
|
* @return array
|
|
*/
|
|
protected function add_prediction_processors($cases) {
|
|
|
|
$return = array();
|
|
|
|
// We need to test all system prediction processors.
|
|
$predictionprocessors = \core_analytics\manager::get_all_prediction_processors();
|
|
foreach ($predictionprocessors as $classfullname => $unused) {
|
|
foreach ($cases as $key => $case) {
|
|
$newkey = $key . '-' . $classfullname;
|
|
$return[$newkey] = $case + array('predictionsprocessorclass' => $classfullname);
|
|
}
|
|
}
|
|
|
|
return $return;
|
|
}
|
|
}
|