nomad-utility-workflows plugin

Utilities for interfacing with NOMAD within workflows, e.g., via a workflow manager, including python API functions for uploading to NOMAD and querying the repository as well as automated generation of NOMAD custom workflow yaml file (NOMAD workflow data models).

This nomad plugin was generated with Cookiecutter along with @nomad's cookiecutter-nomad-plugin template.

Usage

For direct usage and integrating the utility module into other plugins or codes, nomad-utility-workflows is available as a PyPI package:

pip install nomad-utility-workflows>=0.0.8

Linking to your NOMAD account

Create an account on https://nomad-lab.eu/. Store your credentials in a .env file in your working directory, at the root plugin directory for developers, or in some directory that is added to your PYTHONPATH, with the following content

NOMAD_USERNAME="MyLogin"
NOMAD_PASSWORD="MyPassWord"

and insert your username and password.

Caution

Never push your .env file to a repository. This would expose your password.

Development

If you want to develop locally this plugin, clone the project and in the plugin folder, create a virtual environment (you can use Python 3.9, 3.10, or 3.11):

git clone https://github.com/FAIRmat-NFDI/nomad-utility-workflows.git
cd nomad-utility-workflows
python3.11 -m venv .pyenv
. .pyenv/bin/activate

Make sure to have pip upgraded:

pip install --upgrade pip

We recommend installing uv for fast pip installation of the packages:

pip install uv

Install the nomad-lab package:

uv pip install '.[dev]' --index-url https://gitlab.mpcdf.mpg.de/api/v4/projects/2187/packages/pypi/simple

Note! Until we have an official pypi NOMAD release with the plugins functionality make sure to include NOMAD's internal package registry (via --index-url in the above command).

The plugin is still under development. If you would like to contribute, install the package in editable mode (with the added -e flag):

uv pip install -e '.[dev]' --index-url https://gitlab.mpcdf.mpg.de/api/v4/projects/2187/packages/pypi/simple

Run the tests

You can run locally the tests:

python -m pytest -sv tests

where the -s and -v options toggle the output verbosity.

Our CI/CD pipeline produces a more comprehensive test report using the pytest-cov package. You can generate a local coverage report:

uv pip install pytest-cov
python -m pytest --cov=src tests

Run linting and auto-formatting

We use Ruff for linting and formatting the code. Ruff auto-formatting is also a part of the GitHub workflow actions. You can run locally:

ruff check .
ruff format . --check

Debugging

For interactive debugging of the tests, use pytest with the --pdb flag. We recommend using an IDE for debugging, e.g., VSCode. If that is the case, add the following snippet to your .vscode/launch.json:

{
  "configurations": [
      {
        "name": "<descriptive tag>",
        "type": "debugpy",
        "request": "launch",
        "cwd": "${workspaceFolder}",
        "program": "${workspaceFolder}/.pyenv/bin/pytest",
        "justMyCode": true,
        "env": {
            "_PYTEST_RAISE": "1"
        },
        "args": [
            "-sv",
            "--pdb",
            "<path-to-plugin-tests>",
        ]
    }
  ]
}

where <path-to-plugin-tests> must be changed to the local path to the test module to be debugged.

The settings configuration file .vscode/settings.json automatically applies the linting and formatting upon saving the modified file.

Documentation on Github pages

To view the documentation locally, install the related packages using:

uv pip install -r requirements_docs.txt

Run the documentation server:

mkdocs serve

Test Notebooks

To run the test notebooks, create a jupyter kernel using your venv:

python -m ipykernel install --user --name=nomad_utility_workflows

Main contributors

Name	E-mail
Joseph F. Rudzinski	joseph.rudzinski@physik.hu-berlin.de