Introduction

Our task was to build an evaluation system for the Computer Science Department at PES University. This evaluation system is to be used by faculties to evaluate final year projects as per the evaluation procedure.

Evaluation procedure

• Students are required to form teams and select a faculty member as a guide and register with the project coordinator of the Computer Science Department. If no guide is chosen, the project coordinator assigns one to them.

• Upon completion of registration of teams, the project coordinator will create panels and assign faculty members to each of these panels. The project coordinator will then assign one faculty member from each panel as the coordinator for a panel which we refer to as a panel coordinator.

• The coordinators of each panel are in charge of scheduling reviews for all the teams in their panel. They are also responsible for assigning faculty members to review teams for each review.

• The evaluator performs the job of grading each team assigned to them for every review.

Structure of the solution

1. The application at its core is designed based on a hierarchy. Each user can access and modify the content only within the boundaries set up by their role as described in Figure 1. The 3 Roles are Administrator, Coordinator and Evaluator

2. Administrator

• The administrator is incharge of inserting and maintaing the meta data for students, teams and faculties.
• The administrator can initialize panels with faculties and teams in it.
• The administrator also has a few special permissions like opening/closing the student portal, mail student portal passwords to students and download the marks summary for each student.

3. Coordinator

• The coordinator is incharge of setting up project reviews for all the teams in their panel.
• The coordinator needs to assign a list of faculties for each team for each project review who shall in turn evaluate the team.
• The coordinator must also set up opening and closing time for each project review within which the evaluators must evaluate the teams.

4. Evaluator

• Evaluators' role is to evaluate all the teams assigned to them for every project review.

5. Student Portal

• Students are given a "mini" user status when they submit their project details to the project coordinator. They can set up their profile image, project description as well as view the public comments provided by their reviewers.
• This portal is opened by the administrator.

6. Other Salient Features

• The interface provides E-Mail support. eg:
  • The coordinators can mail the review schedule to all the evaluators in the panel.
  • Passwords are generated and mailed to all students to access the student portal.
• The interface also provides CSV Upload/Download.
• Evaluators can view profile images submitted by students.

Implementation

1. REST-API

• The back-end of the web application is a REST API which interacts with clients using JSON requests and responses.
• The backend of the web application is developed using Django Framework.
• PostgreSQL is the database engine used.
• Django ORM is used to convert the relational data (postgres) to objects.

2. Security

• The web application authenticates the user using a combination of JWT tokens and Django Session Cookies.
• The JWT Access Token is used to authenticate the user to communicate with the REST API and hence its validity is kept short.
• The access token is refreshed periodically using a Refresh Token whose validity is substantially longer.
• Refresh token for the corresponding user is stored at the back end and is accessed using the session cookie.

3. Single Page Application

• Implementation of a Single Page Application using breadcrumb structure with vanilla javascript.
• User Interface is designed with Bootstrap v4.5.

5. Dockerized for cross-platform reliability

• The web application is dockerized for reliability across platforms.
• To reduce the hassle of installing third party libraries and easy deployment, the app is dockerized into containers for the REST API, Nginx, PostgreSQL

Requirements

A more detailed list is available in the source code as requirements.txt

They can be downloaded by running

sudo apt install docker docker-compose nginx gunicorn
pip install -r requirements.txt

The dockerfile is as follows

FROM python:3.6
EXPOSE 8000
ENV PYTHONUNBUFFERED 1
RUN mkdir /code
WORKDIR /code
ADD requrements.txt /code/
RUN pip install -r requirements.txt
ADD . /code/
LABEL creators="the NVRD team"

The docker-compose file is as follows The number of workers have been set to 2, they can be changed accordingly here.

version: "3"

services:
  db:
    image: postgres
    container_name: nvrd-postgres
    volumes:
      - postgres_data:/var/lib/postgresql/data/
    environment:
      - POSTGRES_USER=postgres
      - POSTGRES_PASSWORD=postgres
      - POSTGRES_DB=nvrdb
  web:
    build: .
    command: bash -c "cd Back_end/NVRD && python3 manage.py makemigrations && python3 manage.py migrate && gunicorn --workers=2 NVRD.wsgi:application --bind 0.0.0.0:8000"
    container_name: nvrd-django
    env_file:
      - .env.dev
    volumes:
      - .:/code
    ports:
      - "8000:8000"

  nginx:
    build: ./nginx
    container_name: nvrd-nginx
    ports:
      - 80:80
    depends_on:
      - web

volumes:
  postgres_data:

Run the following commands to start the server in the project directory. Choose 0 when prompted to initialize the database with an admin user.

bash run.sh

run.sh file contains the following code

read -p "enter 1 to initialize database (default NO): " dbinit
sudo docker-compose up --no-start
sudo docker-compose start
echo "NVRD containers are running"
dbinit=$dbinit
if ((dbinit==1)); then
	sudo docker exec nvrd-django bash -c "python Back_end/NVRD/manage.py flush --noinput"
	sudo docker exec nvrd-django bash -c "python Back_end/NVRD/manage.py shell < Back_end/NVRD/eval/values.py"
    echo "nvrdb initialsed"
fi