CI.Edu: Jenkins

The project aims to deliver a solution for a grading system that uses a continuous integration pipeline to grade the repositories with test automation of computer science students in different software engineering courses using different server deployments and programming languages.

1. Intro to continues integration in education

1.1 Problem statement

• Deployment and configuration of software projects in software engineering courses is mainly a manual task, since applications need to be built, tested, deployed and updated frequently, after every exercise where changes are made in the source code and features are updated.
• Since applications need dependencies, such as build tools, software development kits and web containers, testing for desired features is a time-consuming process that needs a lot of human interaction.
• The possibility that students change unit tests to their advantage is given when tests are manually executed.

1.2 Solution conzept

• Supervisors of software engineering courses need pipeline automation to get the grading of these development environments, so that student's repositories can be tested and checked for completion.
• The manual task of configuration pipelines that grade computer science students should be as coinvented as possible, so that supervisors are able to automatically grade their students programming skills.

1.3 Evaluated technologies

• Jenkins pipelines + Github (public/private Repo) -> http://10.25.2.215:32234
• Gitlab pipelines
• Github Workflows pipelines

2. SwengX coding projects

Over the different software engenieering courses, sample projects with sample unit, integration and end-to-end where covered and are used in this project as sample data to test pipeline automation of the deployment.

2.1 Java Enterprise - Gradle

gradle:6.8.3-jdk-15.

• swenga-calculator ✓ -> unit test fails case, gradle test
• swenga-guestbook -> n students, unit test passes, gradle test - todo morgen

build.gradle {
    test {
        useJUnitPlatform()
        ignoreFailures = true
    }
}
// must be added to build.gralde so the pipeline does not exit on failed tests

2.2 Spring Boot - Gradle

gradle:6.8.3-jdk-15.

• swenga-springboot-employee-manager ✓ -> spring boot, integration test, mysql db, gradle

build.gradle.kts {
    tasks.withType<Test> {
        useJUnitPlatform()
        ignoreFailures = true
        reports {
            junitXml.isEnabled = true
            html.isEnabled = false
        }
    }
}
// must be added to build.gralde so the pipeline does not exit on failed tests

2.3 Python

the testing tool pytest is used to generate XML reports.

• swengs-exercise1 ✓ -> for n students, simple unit test
• swengs-exercise2 -> n students, unit test passes, pytest - todo morgen

2.4 Django

• swengs-exercise3 ✓ -> django db, django-pytest
• swengs-exercise4 ✓ -> regression testing, django db, django-pytest

pytest.ini {
    [pytest]
    DJANGO_SETTINGS_MODULE = movie_site.settings
}
# pytest configuration

2.5 Angular

• swengs-ima-employees

https://github.com/karma-runner/karma-junit-reporter

https://testing-angular.com/angular-testing-principles/

3. Jenkins & Kubernetes

3.1 Installing jenkins

• Installation of Jenkins in an Kubernetes cluster

• Jenkins namespace

• Deployment per .yaml file or HELM

• Jenkins IP: http://10.25.2.215:32234

• credentials: username: admin / pw: def563d1312043b1824722a70fb7965b

3.2 Kubernetes plugin

• config.yml kubernetes admin credentials

• createing pods with docker images

4. Pipelines

4.1 Groovy sytnax + pod tempaltes

https://www.tutorialspoint.com/groovy/index.htm

4.2 Teacher student conzept

4.3 How to test

• Bash Script Pipelines für verschiedene Technologien & Programmiersprachen
• Builden und Testen der Studenten-Projekte in einzelnen Pods im Kubernetes
• Generieren von JUnit XML reports über die Testergebnisse
• Zusammenführen aller reports der Studierenen pro Hausarbeit

https://hub.docker.com/

BÜPA

Aufgabenstellung:

Schon jetzt wird in der Informatik- bzw. Programmierausbildung am Studiengang IMA mit Jupyter/Nbgrader eine Tool-Chain zur automatisierten Beurteilung von Hausarbeiten verwendet. Sobald man aber die Jupyter-Umgebung verlässt, wird eine automatische Beurteilung deutlich schwieriger. All dies kann mittels Bash-Script automatisiert werden!

In diesem BÜPA Projekt geht es nun darum, das Validieren der Abgaben mittels klassischer CI-Pipeline zu ersetzen.

• Das heißt am Ende des Abgabefensters wird ein automatisierter CI-Build gestartet, dessen Ergebnis zur Beurteilung einer Heimarbeit führt.
• Wichtiger Aspekt ist die Erstellung von Zugriffsberechtigungen (Hausübungen sollen nur für die betroffenen Studierenden bzw. Studierendenteams sowie Lehrende zugänglich sein).
• Ebenso soll es nicht möglich sein, vorgegebene Tests abzuändern (in der CI-Pipline werden die originalen Tests verwendet).

Docs

First Meeting

to evaulate:

• Database support,
• Access rights
• Multiple repos in pipeline
• Cost of solution,
• Language Support,
• Integration Test ...
• public private repo
• local online