Changelog

Release 2.2.0

• Added support for NumPy 2, in addition to NumPy 1.

Release 2.1.17

• Added a new function to automatically determine a diffraction mask, auto_masking() (#46).

Release 2.1.16

• Fixed an issue where autocenter could return bogus results. A new parameter has been added to address This problem while maintaining backwards-compatibility (#45).

Release 2.1.15

• Add support for scikit-image 0.22+, which dropped for some arguments of skimage.registration.phase_cross_correlation.

Release 2.1.14

• Fixed documentation of powdersim which specified the wrong units for argument q (#44).

• Updated support for scikit-image 0.21, which is deprecating some behaviors of skimage.registration.phase_cross_correlation.

Release 2.1.13

• Fixed some deprecation warnings due to matplotlib 3.6.

Release 2.1.12

• Improved the masking functionality in bragg_peaks_persistence.

Release 2.1.11

• Added new Bragg peak determination functionality to combat new datasets with sub-optimal signal-to-noise ratio. Further added Brillouin zone determination based on Bragg peak locations (#42).

Release 2.1.10

• Fixed an issue where masks needed to be provided in bragg_peaks (#41).

Release 2.1.9

• Updated bounds on pyqtgraph.

Release 2.1.8

• Added explicit testing with python 3.10.

• Cleaned up unused imports.

• Updated some modules which were using deprecated code.

Release 2.1.7

• This release brings no changes, and only fixes the conda-forge package.

Release 2.1.6

• Fixed an issue where gaussian() would trip on a full-width at half-maximum of 0.

• Fixed an issue where the first stage of the dual-tree complex wavelet transform was not shifted properly (#36).

Release 2.1.5

• Releases are now automatically performed using Github Actions

• It is now possible to install all the dependencies required to use diffshow() using the following installation option: pip install scikit-ued[diffshow].

Release 2.1.4

• Increased the reliability of bragg_peaks() to distinguish between noise and Bragg peaks.

Release 2.1.3

• Added the function bragg_peaks() to determine the location of single-crystal diffraction peaks in an image.

• Fixed deprecation warnings regarding NumPy’s dtypes.

Release 2.1.2

• Improved autocenter() for diffraction patterns with large Ewald sphere walkoff.

• diffread() now supports NumPy’s *.npy format.

• Speedup of all routines that use the Fast Fourier transform (autocenter(), align(), ialign(), itrack_peak(), and kinematicsim()) by 50%.

Release 2.1.1

• Added the autocenter() routine, to automatically find the center of diffraction patterns. This works for both single-crystal and polycrystalline patterns.

• Support for Python 3.6 and NumPy<1.17 has been dropped

Release 2.1.0

This release brings major infrastructure improvements, which in turn have allowed to squash some bugs.

• Migration of continuous integration testing to GitHub Actions.

• Migration of test infrastructure to pytest.

• Tests are now included in source distributions.

• Code snippets in documentation are now tested for correctness.

User-facing changes:

• Fixed an issue where the detected shift in skued.align and skued.ialign might be partial (i.e. only shift in one direction).

• Fixed an issue with packaging data on Linux.

• The fast argument to skued.align and skued.ialign has been removed. It was previously-marked as deprecated.

• Added pre-emptive support for Python 3.10+ by removing deprecations.

• Increased the precision of the pseudo-voigt approximation in skued.pseudo_voigt.

• Fixed many issues regarding documentation being out-of-date.

Release 2.0.6

• scikit-ued is being re-licensed from the MIT license to the GPLv3 license.

• The fast argument to skued.align and skued.ialign has been deprecated. Its value has no effect anymore.

• Official support for Python 3.9.

• Removed explicit requirement for the tifffile package.

Release 2.0.5

• Added skued.kinematicsim, a simple function to compute electron diffraction patterns from crystals structures in the kinematic approximation (i.e. thin samples).

• Added the skued.RingArcSelection area.

• Various documentation improvements and fixes.

Release 2.0.4

• Added support for crystals.ElectronicStructure. This requires crystals version 1.1.0 and up.

• Added the function with_irf, which allows to modify fitting functions to include the effects of instrument response.

• Various documentation fixes.

Release 2.0.3

• Added the Selection.mpatch method to draw patches on Matplotlib plots.

• Added the spectrum_cmap Matplotlib colormap, available under the name “spectrum”.

• Fixed an issue where diffracted intensities were not correctly scaled in potential_map.

Release 2.0.2

• Added the Selection class and RectSelection, DiskSelection, RingSelection, and ArbitrarySelection to assemble time-series. This is a generalization of iris-ued’s time-series rects.

• Added real-time pixel value and cursor position to skued.diffshow.

• Added indices_to_text, a plotting utility function to render Miller indices to Mathjax/LaTeX-style text (Matplotlib-compatible).

Release 2.0.1

• skued.diffshow will temporarily switch PyQtGraph’s image axis order to the row-major, which is a saner default.

• Added skued command-line utilities. Images can be shown (with interactive viewer) using skued diffshow [path]. Crystal information can be determined using skued crystinfo [path].

• Fixed an issue where a typo in electron_velocity would raise an exception.

Release 2.0.0

Due to a conflict between scikit-image and scikit-ued conventions, some breaking changes are required. Image conventions will now follow that of scikit-image. Most importantly:

• Changed the convention on image masks to align with the scikit-image convention. Masks will be True for valid pixels, and False on invalid pixels.

We took the opportunity to make other breaking changes:

• Broke off the skued.structure package into its own library, crystals.

• Removed masked_register_translation in favour of the new scikit-image implementation ported from scikit-ued.

• Removed xcorr and mnxc as these were the backbone of masked_register_translation and are no longer needed.

• Added aspherical electron form factor parametrization from Zheng et al. 2009.

• Removed diff_register in favor of an analog of scikit-image’s register_translation and masked_register_translation.

• Removed powder_center due to unpredictable performance.

• Removed calibrate_scattvector, which was deprecated.

• Removed time_shift and time_shifts, which were deprecated.

• Removed shift_image in favor of scipy.ndimage.shift.

• bounded_reflections has been removed in favor of Crystal.bounded_reflections in the crystals library (version >= 0.6.4)

We have also added some features:

• Added the patterson function to calculate Patterson pair-pair distribution functions from polycrystalline diffraction patterns.

• Added the detector_scattvectors function to determine the wavevectors visible on a detector, in transmission, based on experimental geometry.

Release 1.0.1.1

• Added time-series fitting.

Release 1.0.1.0

• Added support for Gatan Digital Micrograph image formats DM3 and DM4

Release 1.0.0.0

• available_dt_filters and available_first_stage_filters have been added to list available baseline-removal filters.