Course Master - Django project

Installation

Install Python3

Install Python3

sudo apt-get install python3

Check installation with python3 --version

Then, install pip3 if not done yet.

sudo apt-get install python3-pip

Then, install pipenv with pip3 install pipenv

Django installation

Now, you can install Django by doing pipenv install django.

This will create you a virtual environment and install Django into it.

Django-Redis installation

Type pip install redis to install redis.

To run the redis server, type docker run -d --name redis-stack -p 6379:6379 redis/redis-stack:latest

Configuration

In the repo folder, type pipenv shell. This will lead you to the venv of the project.

Then, type python manage.py runserver when you're in the repo root folder.

Usage

Now, you can go into http://localhost:8000/redis to access the project.

Once a user is registered to a course, he's automatically subscribed to courses update for all courses he follows into the notifications page.

Limitations

### Sessions

This app works with sessions. You have to play with this project with 2 browsers to test the pub/sub.

To test pub/sub, please follow these instructions:

Open a browser, register as a Teacher, create a course.

Open another browser, register as a Student, search your course, register to it, and go to notifications.

As the teacher, update, delete or publish a message from the course created.

As a student, you'll see instantly the message.

You can also do this with redis-cli. To do that, in a shell, type:

• docker start redis-stack

• docker exec -it redis-stack bash

• redis-cli

• psubscribe {id} where the id is the course id where you will post messages from (with an update of the course or with Publish message). you can retrieve it with KEYS course:*

• Update the course or post the message from this course id

OR

• docker start redis-stack

• docker exec -it redis-stack bash

• redis-cli

• Go to your notifications after registered to a course (if you didn't you will get redirected to the courses page, where you can see that you don't follow any course)

• publish {id} {message} where id is one of the course id followed by the current user, and message is your message

BUT you can also use only one browser.

While you're a teacher, you're "subscribed" to your own course. So with 2 tabs, it with one, you're on notifications page, and with the other your publish a message, update the course or delete the course, you can also test the pub/sub.

---

Expiration

Expiration of the course: The course expires after 5 minutes. If a student register to the course, it add 1 minute.

---

### Students management

As a teacher, you can't manage students registered to your course.

---

Searching

The search looks for title, description and level fo the course. There's no advanced search for the moment.

Bugs

If you see any bugs, please tell me. I really tried my best by testing every feature every time at every change, but I can miss some tricky things. So please, feel free to mail me at corentin.lemaire@etu.uca.fr to report any bug!

Django model

I'm using django model a little in my code only to create me some "barriers". this standardize my objects from redis to avoid lot's of bugs. This is only kinda sort of structure to format my data.

## Pub/Sub

To implement the Redis pub/sub in Django, I decided to use this mechanism:

• The notifications_view is firstly displayed

• This view calls the notifications function thanks to HTML and JS

• This calls the get_message function, waiting for a new message

• While waiting this message, the HTMl refreshs the page every 10s to avoid the timeout

• When a message is received, notifications function calls back notifications_view and this repeats.

It stops when the user change page. So, the page is constantly "loading" waiting for a message. This allowed me to get real-time message. It's not the most efficient way to do this but I really didn't want to use websockets or JS libraries to do this. It seems the fastest easiest way to do it.

Also, I saw in some docs we could perform some async functions in Django, but seems kinda long and heavy to implement, so I choose only this method.