Skip to content

The Simmate Toolkit

Danger

Many classes in this module are highly experimental. We strongly recommend using pymatgen and ase for toolkit functionality until Simmate hits v1.0.0. For developers, this means many of the classes are undocumented and untested. This accelerates our testing without spending time on reformating tests and rewriting guides. We include this experimental code on our main branch because higher-level functions (e.g. workflows) rely on some of these features. Higher-level functions are well tested and documented, however, to account for ongoing low-level changes.

The toolkit module is meant to be an extension of pymatgen and ase. It includes low-level classes and functions, such as the Structure class and its associated methods. The toolkit module is written entirely in python and does not use third-party DFT programs. See the simmate.apps module for those.

The most commonly used classes from this toolkit are the Structure and Composition classes. These classes can be imported with...

from simmate.toolkit import Structure, Composition

Outline of submodules

  • base_data_types = defines fundamental classes for materials science, including Structure and Composition classes. For convenience, these classes can be imported directly from simmate.toolkit, as noted above.
  • creators = creates structures, lattices, and periodic sites
  • featurizers = converts properties into numerical descriptors for machine-learning
  • structure_prediction = predicts crystal structures using an evolutionary algorithm
  • symmetry = contains tools and metadata for symmetry analyses, such as spacegroups and wyckoff sites
  • transformations = methods to transform or "mutate" Structures
  • validators = methods to check if a structure meets specific criteria