SkillBuilder

A web application for evaluating students' professional skills in real time.

Description

The purpose of the SkillBuilder application is to allow instructors to assess teams of students in real time using research-based or custom rubrics. Instructors are offered the ability to email their students and view basic statistics and graphs. Additionally, instructors can download the data for analysis.

Getting Started

Dependencies

If you choose to run the application using Docker, then Docker will be the only thing to install from the list below on your system. Otherwise, ensure the following:

• Operating System: Windows 10* or newer*, MacOS11+*, or any modern Linux distribution.
• Runtime Environments:
  • Node.js(v21.6.1 or higher) - Handles frontend logic.
  • Python(v3.12 or higher) - Handles backend server logic.
• Database/Caching:
  • Redis(v5.2.1) - Handles session management and rate limiting.
  • MySQL(v8.0) - Manages the application's data.
• Package Managers:
  • pip for Python backend.
  • npm for Node.js frontend.
• Docker(recommended; skip if not using) - simplifies setup and ensures consistency across operating systems.

Note: Running the application on Windows locally(outside of docker or WSL) is not recommended or stable.

You can install the visual Docker Desktop app or just the CLI here.

Installing

Docker

1. Build the containers.

docker compose build

This step is needed for whenever Docker files are modified.

Linux

Debian/Ubuntu (and its derivatives)

1. Perform any system upgrades.

sudo apt update -y
sudo apt upgrade -y

2. Install Python3:

sudo apt install python3
python3 --version

Ensure that the version is >= 3.12.

Note: Debian uses the last stable release of Python (which is not 3.12), but from testing, it seems to work just fine.

3. Install Redis:

Using the following link to install:

https://redis.io/docs/latest/operate/oss_and_stack/install/install-redis/install-redis-on-linux/

Note: Ubuntu and Debian typically use systemctl as the init system, but if using something different, the docs will not cover those.

4. Install Node.js:

sudo apt install nodejs
node -v # Checking to make sure it is installed.

MacOS

MacOS will require some kind of package manager (this document will use homebrew).

You can find homebrew here: https://brew.sh/

1. Install Python3

You can find the downloads here:

https://www.python.org/downloads/macos/

2. Install Redis

brew install redis

3. Install Node

Either download prebuilt binaries directly, or use a package manager:

https://nodejs.org/en/download/package-manager

Windows

Running this project on bare metal Windows is no longer supported. You will need to get WSL (Windows Subsystem for Linux) or preferably WSL2.

The following shows you how to set it up:

https://learn.microsoft.com/en-us/windows/wsl/install

Once this is installed and set up, open Windows Terminal, Powershell, Command Prompt (or whatever terminal emulator you use) and do:

wsl

If this is working correctly, follow the installation instructions in the Linux section of this README above.

Executing the application

Docker

To start up the built containers run:

docker compose up

The containers are up and running if none exited with an error of 1.

Use ctrl+c to shut down the containers(must be in terminal that started the containers or issues commands to them).

When the frontend finishes compiling, it should serve the web app at the link it notes. I will look like http://localhost:3000. Click on it or open it in your preferred browser.

Not Using Docker

You can also run rubricapp without Docker, but you will need to manually run the setup yourself.

1. Create a virtual environment

python3 -m venv <environment_name>
source <environment_name>/bin/activate

This is where all of the Python dependencies will be stored instead of being stored globally on your system. You will need to do the source command each time to pop into the virtual environment.

2. Start Redis

Enable the Redis service using your appropriate init system (systemctl in this example).

systemctl start redis

Make sure that it is running:

systemctl status redis

3. Launch the backend:

user@/(project root)$ cd BackendFlask
user@/(project root)/BackendFlask$ python3 ./setupEnv.py -irds

The setup flags are as follows:

• -i, --install - install all dependencies
• -r, --reset - reset the database
• -d, --demo - load demo data into the database
• -s, --start - start the backend server

Later iterations of using setupEnv.py only requires the -s flag (unless new depencencies are added or if the database needs to be reset etc).

4. Launch the Frontend Server

user@/(project root)$ cd FrontendReact
user@/(project root)/FrontendReact$ npm install # only do this once
user@/(project root)/FrontendReact$ npm start

This will launch the server on port 3000. Access it by navigating to http://localhost:3000 in your browser and logging in with appropriate credentials.

Help

Docker questions

1. So how will I be running or starting the code most of the time?

docker compose build; docker compose up

2. I am noticing that Docker is not building up the new continers right after modifying docker files; what can I do?

   • Docker does cache things like its builds. Try the following:

docker compose build --no-cache

3. This is all nice, but how do I get into a container to jump into the database or something live?

# The following command gets all the running containers.
docker ps
# OUTPUT:
# CONTAINER ID   IMAGE     COMMAND   CREATED   STATUS    PORTS     NAMES
# Notice the output above was empty. That is because my application is currently down. 
# NOTE: you can only go into a container when it is running.
# Fire up the application then run the command again and look at the [NAMES].
docker exec -t -i [CONTAINER_NAME] /bin/bash
# That previous command will pop you into the terminal of the container named.

4. What if I want to get rid of a container?

# Note the containers must be stopped
docker ps -a # This lists all containers even if they are not running.
docker rm [CONTAINER_ID] # You get the ID from docker ps -a.
# A slightly more efficient command to remove them all is:
docker rm -v $(docker ps -aq) #-q means to only list container ids.

5. The containers are out of control. No matter what I do they will not stop even if I restart my machine. What can I do?
   • This is more prone to happen on machines with lower resources. The following command is tested to run on Linux and WSL2. Try to find the equivalent command online for your operating system.

# This command shuts down AppArmor on Linux.
# Without it working, opening other applications will be impossible until you restart your machine.
sudo aa-remove-unknown && docker stop $(docker ps -aq)
docker rm -v $(docker ps -aq) # Getting rid of the containers so they do not auto-start.

6. Even when I get rid of the container, things like the database are not in a clean new state. Why?
   • Why this is happening depends a little more on your system and Docker settings. It is possible that somewhere things are being cached or saved. I recommend the following command to just clear out the containers and their volumes.

docker-compose down --rmi all --volumes

7. How do I run the tests?
   • Go to Dockerfile.backend and change the flags from -ds to -t for the moment. Likewise change the cmd line in the fontend docker file to get the front end to run through its tests.

It is recommended to make these commands into aliases. Here is how to on Ubuntu. You do not have to use vim; use your favorite text editor.

Troubleshooting outside of Docker

REDIS

If it does not start correctly, there could be a multitude of reasons. I suggest using journalctl to investigate it (systemctl will give out the full command).

But a good starting point is seeing if it is already running:

ps aux | grep redis

This will give the PIDs of all processes with redis in its name. Try killing them with kill <pid1> <pid2> ..., <pidN> and then rerunning systemctl start redis.

Note: if redis is not considered a service, try using redis-server or redis-server.service.

Port Conflicts

The backend runs on port 5000 and the frontend runs on port 3000. You may already have processes running on those ports. If this is the case, you will have conflicts and the server(s) will not run normally.

You can check what is running on those ports with:

lsof -i :5000
lsof -i :3000

If any output appears here, you may either want to kill them with kill, or run those processes on different ports.

Other

If you are testing with adding students/TAs/admins, it may be time consuming to manually do it via the website. There is a script that will automatically insert new users into the database straight from the command line. It is important to note that this script only works if the backend is currently running inside docker.

Run this script with:

./dbinsert.sh

Run this and follow the on-screen instructions.

Testing

This will first explain how to launch the tests as if you do not use Docker. The logic for this is so that you understand why certain docker lines are bing altered.

To launch the backend tests, run the following in the BackEndFlask/ directory:

setupEnv.py -t # The -t flag runs pytests for us.

Note that for this to work you need to have built things up having used -ds or some variation.

To launch the frontend Jest tests, run the following:

npm test

Note that for this to work you need to have have the backend running and serving connections.

Now we swap over to getting tests to run in our Docker containers.

To get the backend tests going, navigate to rubricapp/Dockerfile.backend. Notice that the last line is what is starting up the back end code. It is running setupEnv.py -ds. Change

-ds
# To
-t

Now build fresh without the cache and run Docker compose up. The container will now run through the tests. Do not forget to change this back once you are done testing.

To get the frontend tests going, navigate to rubricapp/Dockerfile.frontend. Like the backend, we are only concerned with the very last line that starts the backend. The line we see is:

CMD ["npm", "start"]

Change it to:

CMD ["sh", "-c", "sleep 35 && npm start"]

Do not forget to change it back. One key difference is that there is a sleep before the command is ran. This is to allow the backend setup since Jest tests make hits to the backend. You can adjust the timer to more or less depending on how fast the backend is ready on your system.

Note that running the Jest tests are not fast. It will look like it is frozen. It will then dump a lot of errors and warnings. Cleaning this up is still in progress.

Because there is a lot of output from Jest tests, I recommend running:

docker compose up > jestTests.txt 2>&1

This will keep your termianl uncluttered and log things in the jestTests.txt file.

Now we will talk about how to create a backend test. Go to BackEndFlask/Functions/test_files/test_genericImport.py. We will be looking at def test_should_fail_with_file_not_found. The first thing to notice is that the function names will start with the word test followed by a brief description of what is being tested in snake case. The function needs to take the object flask_mock_object to be able to connect to the application. Use the same with line there in your function to use the applications context. If ever you want to use the database connection itself, then you can use flask_mock_object.db. Make sure to clean up any data that your function input to the database, so that the next test is working with a clean database.