This is the ITPcp plasma group meta-repository that collects scripts to setup development and use of internal and external codes. Our development environment is Visual Studio Code, and we strongly recommend GitHub Copilot there and in the CLI.
On Windows: Prepare your machine with
scripts/setup_windows.bat first.
CODE can then be used via the devcontainer or alternatively via
Debian Linux in WSL.
On Linux: At ITPcp all packages should be installed to get going.
For your own setup, look inside the packages in the
Dockerfiles base, devel and devel-tex inside images or
use the devcontainer.
On Mac: The recommended way is the devcontainer. The next option is
multipass via the most recent LTS version of Ubuntu Linux. The
option on the local system is brew with similar packages as in Linux,
but it is prone to problems with Fortran on Apple Silicon.
If you haven't done so earlier, set up your SSH keys in ~/.ssh via ssh-keygen and
add the content of id_rsa.pub to Gitlab and GitHub for authentication.
Clone the repository to your working copy, at the institute this is
git clone git@gitlab.tugraz.at:plasma/code /proj/plasma/CODE/<username>
Then open the directory in VS Code with
code code
When asked to initialize the devcontainer, do so. If there is no message run F1 and Remote-Containers: Rebuild and reopen in Container. Wait for
Configuring Dev Container to finish in the lower status bar. Then open a
new bash shell. You should see the prompt
Activating /workspaces/code on branch main
(.venv) root@fcb4ad176bde:/workspaces/code#
When working outside the container, run the setup script manually with
source setup.sh
The setup will install external dependencies and create and activate
a Python virtual environment in the hidden .venv directory.
If you work outside a container, also manually put a line
source /path/to/code/activate.sh
in your bashrc.
Tests can currently be performed by using
pytest tests/
This will perform all the tests in tests/ and its subfolders.