Published April 9, 2026 | Version v0.13.0
Software Open

hyperspy/rosettasciio: v0.13.0

  • 1. UKAEA
  • 2. University of Lille
  • 3. Paul-Drude-Institut
  • 4. University of Warsaw
  • 5. Direct Electron
  • 6. Luminovo.ai
  • 7. Kyoto U.
  • 8. Department of Materials Science & Metallurgy, University of Cambridge
  • 9. VITO
  • 10. Forschungszentrum Jülich
  • 11. Norwegian University of Science and Technology
  • 12. EPFL
  • 13. @microsoft
  • 14. @OxfordRSE
  • 15. @actions
  • 16. Diamond Light Source
  • 17. @questek
  • 18. ABB, Germany

Description

Enhancements

  • Add lazy support compatible with dask distributed scheduler in pantarhei-format reader (#462)

  • Allow saving image with scalebar when the scales have opposite signs. (#475)

  • Lazify import of rsciio and improve documentation. (#477)

  • EDAX TEAM/Genesis: add support for distributed lazy loading. (#481)

  • msa-format enhancements:

    • support reading files with empty fields.
    • support reading files with multiple line titles and comments.
    • support reading files with multiple columns and datatype xy. (#487)
  • EMPAD:

    • Add q_calibration parameter to set the diffraction space calibration in 1/nm.
    • Add remove_nans parameter to remove NaN values in the data by replacing them with zeros. (#495)
  • EMPAD: add support for distributed lazy loading using memmap_distributed. (#498)

Bug Fixes

  • EDAX TEAM/Genesis: fix OverflowError: Python integer out of bounds when loading some spd files. (#481)
  • EMPAD: fix ordering of axes and array shape when reading non-square scan data. (#493)

Improved Documentation

  • bruker-format: add documentation about support of micro-XRF bcf files. (#471)

Maintenance

  • Add codespell to pre-commit hooks to catch common misspellings. (#472)

  • Improve release workflow:

    • Speed up release workflow by avoiding downloading tests data when testing wheels.
    • Add release notes and distribution files to GitHub release. (#482)
  • Update integration tests workflow to use caching shared in CI matrix. (#483)

  • Consistently use numpy.testing.assert_allclose in favour of numpy.allclose, because the failure message is more useful, as it shows the relative and absolute error. Specify tolerance where needed. (#492)

Files

hyperspy/rosettasciio-v0.13.0.zip

Files (95.2 MB)

Name Size Download all
md5:3f6a08f466f71a4c6f9f8486567ad56d
95.2 MB Preview Download

Additional details

Related works