KIM Quarterly Update (January 2025)

31-Jan-2025

Happy new year! We hope you had an enjoyable winter holiday and wish you a pleasant and productive 2025. We would like to inform you of some recent and upcoming developments in OpenKIM and related projects.

Announcements

  • We are pleased to announce that our paper titled Fundamental microscopic properties as predictors of large-scale quantities of interest: Validation through grain boundary energy trends has been published in Acta Materialia. Using high-throughput OpenKIM property calculations for 304 interatomic models, we quantitatively demonstrate correlation between material properties across length scales.
  • ColabFit, the KIM Initiative project that provides open access to curated and standardized first-principles datasets suitable for fitting IPs, continues to expand. We have recently ingested our 400th dataset!
  • We have updated the call and return signatures of the "get_test_result" simplified query for accessing the OpenKIM Mongo database. The new API is able to access all Test Results in the database, including vacancy tests that were previously only accessible through querying Mongo directly (see the Test Drivers here and here ), and cases where a single Test computes multiple instances of the same property. As we continue work on the Crystal Genome testing framework encompassing all known crystal structures, we intend for "get_test_result" to become the primary query endpoint for accessing OpenKIM Test Results. The query is accessible through a web interface, the kim-query Python package available through pip and Conda, and web requests (curl/wget/etc). See the documentation here.

Ongoing Developments

  • The KIM Developer Platform (KDP) Docker image continues to be updated with new features. The latest release (1.5.5) contains the aforementioned updates to the query interface, as well as improvements to make Test development easier for KIM staff and collaborators. The KIM Binder Sandbox, based on the KDP and accessible through your browser, contains Jupyter tutorials and a terminal for learning about usage and development of KIM content.
  • We are working on a new version of the KIM API. The main development will be improved support for large parameter files, enabling the API to support large machine learning interatomic potentials (MLIP).

How to stay up-to-date