openff-nagl-models 0.2.0

OpenFF NAGL Models

Open NAGL Models

Latest release
Status

This repository contains NAGL models released by the Open Force Field Initiative. They are intended to be used by OpenFF NAGL.

Use

Installing this package exposes an entry point that makes it easy to access models installed in this package.

A convenience function is provided to wrap this for you:

>>> from openff.nagl_models import load_nagl_model_directory_entry_points
>>> load_nagl_model_directory_entry_points()
[PosixPath('/home/.../openff-nagl-models/openff/nagl_models/models/am1bcc')]

You can also list all available models from all entry points:

>>> from openff.nagl_models import list_available_nagl_models
>>> list_available_nagl_models()
[PosixPath('/home/.../openff-nagl-models/openff/nagl_models/models/am1bcc/openff-gnn-am1bcc-0.0.1-alpha.1.pt'), PosixPath('/home/.../openff-nagl-models/openff/nagl_models/models/am1bcc/openff-gnn-am1bcc-0.1.0-rc.1.pt')]

Or validate if a model name is found in the local directory, or an entry point directory:

>>> from openff.nagl_models import validate_nagl_model_path
>>> validate_nagl_model_path("openff-gnn-am1bcc-0.0.1-alpha.1.pt")
PosixPath('/home/.../openff-nagl-models/openff/nagl_models/models/am1bcc/openff-gnn-am1bcc-0.0.1-alpha.1.pt')

Finally, if you want to find all models for a particular type, use get_models_by type. These will come sorted according to semantic versioning, where the latest release is last:

>>> from openff.nagl_models import get_models_by_type
>>> get_models_by_type("am1bcc")
[PosixPath('/home/.../openff-nagl-models/openff/nagl_models/models/am1bcc/openff-gnn-am1bcc-0.0.1-alpha.1.pt'), PosixPath('/home/.../openff-nagl-models/openff/nagl_models/models/am1bcc/openff-gnn-am1bcc-0.1.0-rc.1.pt')]

Versions

• v0.0.1-alpha.1: a pre-production model to use for experimentation. We do not recommend using this model to assign charges in scientific work.
• v0.1.0-rc1: a pre-production model to use to assign AM1-BCC partial charges.
• v0.1.0-rc2: a pre-production model to use to assign AM1-BCC partial charges.

Acknowledgements

Project based on the Computational Molecular Science Python Cookiecutter version 1.0.