Unit testing

For Unit Testing PHPUnit is used.

Run tests

Execute src/PersonalizedSearchBundle/vendor/bin/simple-phpunit

Test files

Test files need to be in the following directory inside the PersonalizedSearchBundle: tests/PersonalizedSearchBundle
The file name needs to match *Test.php (example: SampleTest.php)

Unit test guidelines

Taken from semantictesting.org.

The goal of unit tests is to prove that the smallest unit of code behaves exactly as you expect in isolation.

Objectives

Implement one or more test cases that causes the unit of code to fail
Implement one or more test cases that causes the unit of code to pass
Deliver production code that has at least 70% code coverage

Strategies

Assert external resources are not used
Assert test cases can be executed independent of other test cases
Assert the result produced by the unit of code
Assert the state of stateful objects
Assert interaction between the system under test and its collaborators

SampleTest

Following example shows what a very simple test file for the SegmentBasedAdapter could look like. $queryRed represents the original Elasticsearch query whereas $expectedSegmentBoostedQueryRed represents the query with the expected boosting. For more informations about mocking classes take a look at the documentation.

namespace Pimcore\Bundle\PersonalizedSearchBundle\Tests\Adapter;

use CustomerManagementFrameworkBundle\Targeting\SegmentTracker;
use PHPUnit\Framework\TestCase;
use Pimcore\Bundle\PersonalizedSearchBundle\Adapter\SegmentAdapter;
use Pimcore\Targeting\VisitorInfoStorage;

class SegmentAdapterTest extends TestCase
{
    public function testSegmentBasedAdapter()
    {
        $queryRed = array ( 'multi_match' => array ( 'query' => 'red', 'type' => 'cross_fields', 'operator' => 'and', 'fields' => array ( 0 => 'attributes.name^4', 1 => 'attributes.name.analyzed', 2 => 'attributes.name.analyzed_ngram', 3 => 'attributes.manufacturer_name^3', 4 => 'attributes.manufacturer_name.analyzed', 5 => 'attributes.manufacturer_name.analyzed_ngram', 6 => 'attributes.color', 7 => 'attributes.carClass', ), ), );

        $expectedSegmentBoostedQueryRed = array ( 'function_score' => array ( 'query' => array ( 'multi_match' => array ( 'query' => 'red', 'type' => 'cross_fields', 'operator' => 'and', 'fields' => array ( 0 => 'attributes.name^4', 1 => 'attributes.name.analyzed', 2 => 'attributes.name.analyzed_ngram', 3 => 'attributes.manufacturer_name^3', 4 => 'attributes.manufacturer_name.analyzed', 5 => 'attributes.manufacturer_name.analyzed_ngram', 6 => 'attributes.color', 7 => 'attributes.carClass', ), ), ), 'functions' => array ( 0 => array ( 'filter' => array ( 'match' => array ( 'relations.segments' => 860, ), ), 'weight' => 1.0, ), 1 => array ( 'filter' => array ( 'match' => array ( 'relations.segments' => 966, ), ), 'weight' => 6.0, ), 2 => array ( 'filter' => array ( 'match' => array ( 'relations.segments' => 967, ), ), 'weight' => 6.0, ), 3 => array ( 'filter' => array ( 'match' => array ( 'relations.segments' => 968, ), ), 'weight' => 6.0, ), ), 'boost_mode' => 'multiply', ), );

        $segmentAdapter = $this->constructSegmentAdapter();
        $actualSegmentBoostedQueryRed = $segmentAdapter->addPersonalization($queryRed);

        self::assertEquals($expectedSegmentBoostedQueryRed, $actualSegmentBoostedQueryRed);
    }

    private function constructSegmentAdapter() : SegmentAdapter
    {
        $visitorInfoStorage = $this
            ->getMockBuilder(VisitorInfoStorage::class)
            ->setMethods(['getVisitorInfo'])
            ->getMock();
        $visitorInfoStorage->method('getVisitorInfo')
            ->willReturn(null);

        $segmentTracker = $this
            ->getMockBuilder(SegmentTracker::class)
            ->setMethods(['getAssignments'])
            ->getMock();
        $segmentTracker->method('getAssignments')
            ->willReturn(array ( 860 => 1, 966 => 6, 967 => 6, 968 => 6, ));

        return new SegmentAdapter($visitorInfoStorage, $segmentTracker);
    }
}

Using PHP Stan

Run php stan using src/PersonalizedSearchBundle/vendor/bin/phpstan analyse src tests to analyze code in the src and tests folder.
You can currently choose from 9 levels (0 is the loosest and 8 is the strictest) by passing -l|--level to the analyse command: vendor/bin/phpstan analyse -l 2 src tests. Keep in mind that we want to reach PHP Stan level 2.

For further information visit the PHP Stan documentation.

Code Coverage Report

In order to generate a code coverage report when running PHP Unit, you can use the following command to generate a static HTML report saved to the directory /tmp/codeCoverage/.

vendor/bin/simple-phpunit --prepend build/xdebug-filter.php --coverage-html /tmp/codeCoverage/

You can then view the report using a standard browser. If you want to view the report directly from the development VM, you can start a simple Python-powered web server to serve the static HTML pages. The commands for achieving that are listed below.

# change into the directory in which your code coverage report is saved in
cd /tmp/codeCoverage/

python3 -m http.server 8080

You can now access the code coverage report by navigating your browser to http://<host-ip>:8080.

personalized-product-search/doc/UnitTesting.md