This project is mainly developed using Linux (Debian or derivative). It is possible that authoring workflow could work on MacOS as well, but we are not testing it.
The requirements are different depending on whether you are only interested in develoment of the Python assignment notebooks (master notebooks) or if you need to modify the server or autograder and build deployment images.
Install virtualenv. The command may differ depending on the system.
# On Debian or Ubuntu linux
apt-get install python-virtualenv # Install virtualenv.
# On MacOS with Homebrew:
brew install python3 # Make sure python3 is installed.
pip3 install virtualenv # Install virtualenv.
After that the setup procedure is common
virtualenv -p python3 ../venv # Create the virtual Python environment in ../venv/
source ../venv/bin/activate # Activate it.
pip install -r requirements.txt # Install Jupyter etc. into ../venv.
Install the prog_edu_assistant_tools, or see
python/prog_edu_assistant_tools/README.md for the details.
cd python/prog_edu_assistant_tools/
source ../../../venv/bin/activate
python setup.py bdist_wheel
pip install --ignore-installed dist/prog_edu_assistant_tools-0.1-py3-none-any.whl
cd ../..
To start the Jupyter notebook run command
jupyter notebook
The local development requires the following tools:
- Virtualenv setup for Python and Jupyter for authoring master notebooks, as described above.
- Go toolchain (https://golang.org), because the server and autograder are implemented in Go.
- Bazel (http://bazel.build).
- Docker (https://docs.docker.com/install) is used for building deployment container images and local testing of autograder backend.
- Google Cloud SDK (https://cloud.google.com/sdk/install) is used for deployment of autograder backend to GCE.