A cookiecutter template for developing a Python package for mBuild recipes. Based on the MolSSI Cookiecutter.

Continuous integration pipelines on Azure are currently not working. The Azure organization for this project may have been inadvertently flagged due to attempts to cut down on abuse, such as crypto mining, in recent months. See: https://devblogs.microsoft.com/devops/change-in-azure-pipelines-grant-for-public-projects/. A ticket has been submitted to resolve this issue.

• Pre-configured setup.py for installation and packaging which automatically registers with the entry_point group defined in mBuild
• Pre-configured Linux and OSX continuous integration on Azure
• Basic testing with PyTest
• Sample example directory

• Python 3.6 or later
• Cookiecutter

With cookiecutter installed, execute the following command inside the folder you want to create your skeletal mBuild recipe directory.

cookiecutter /path/to/mbuild-cookiecutter/

From here, the user will be prompted for information on how to build the directory such as package name, module name, and authors.

Unit testing with pytest has been integrated within this cookiecutter. This testing framework is used in mBuild. The default tests added in mbuild_{{cookiecutter.directory_name}}/tests/ are to check that the cookiecutter package has been successfully integrated with mBuild recipes via entrypoints.

Additional tests can be added to the project/tests/ folder. Any function starting with test_* will be automatically included into the testing framework.

To run the suite of unit tests, type pytest -v on the command line.

Continuous integration is done with Azure for both Linux and MacOS testing. Azure DevOps is free for open source projects and allows you to test and verify that your mBuild recipe works with various OS and Python versions.

Currently, continuous integration is setup to test Linux and MacOS builds for Python versions 3.7 and 3.8. The build instructions are contained in azure-pipelines.yml and is designed to work out of the box. However, you may have to edit the instructions to include any new dependencies.

To connect your mBuild recipe to mBuild, the first step is to install your recipe locally by running pip install -e . in the project root directory. You can test that your recipe classes are recognized by mBuild by running the following list of commands:

python
import mbuild
mbuild.recipes.name_of_your_plugin()

mBuild recipes using the cookiecutter template can be built using the provided docker image. To run the docker image, simply execute the following command docker run -it rmatsum/mbuild-cookiecutter:latest which will pull the most recent docker image from Docker Hub. Running this image will drop you into the software directory which contains the installed mbuild-cookiecutter package. From there, the instructions in the Usage section can be followed to build the mBuild recipe.