Release 0.9.5

Overview

HEXRD 0.9.5 introduces numerous features and improvements. A few highlights are listed below:

• Refactored instrument, powder, and laue calibrators to use lmfit (including boundary conditions)
• Added pressure and temperature calculations for adjusting lattice parameters
• Added detector and instrument templates
• Improvements to the GE inverse distortion function performance
• Several performance improvements and bug fixes

The full list of changes are provided below.

What's Changed

• Change "hexrd-prerelease" label to "prerelease" by @psavery in #572
• Remove urllib3 version pinning by @psavery in #574
• import angs_to_dvec as well by @saransh13 in #576
• Only include "roi" in the config if it is not None by @psavery in #579
• Inverse distortion by @ZackAttack614 in #584
• Install pybind11 if it is missing by @psavery in #587
• Add yaml dumper that converts numpy to native by @psavery in #588
• Add option to set eta_step size for eta omega maps by @psavery in #573
• Update README.md install instructions by @psavery in #583
• Add property for point group number to material by @psavery in #585
• fixed bug in GrainData.load and added rotation_matrices attribute by @donald-e-boyce in #589
• Add matplotlib as a dependency to HEXRD by @psavery in #590
• Add support for detector "group" key by @bnmajor in #591
• Add pressure and temperature calculations for lattice parameters by @psavery in #593
• Add initial detector templates by @bnmajor in #594
• Remove transform section from these detectors by @psavery in #597
• numba accelerated findDuplicateVectors by @saransh13 in #596
• Remove unused "calibration" directory by @psavery in #598
• Update rawimage.py by @darrencpagan in #601
• ProcessedImageseries with frame_list automatically updates omega metadata by @donald-e-boyce in #603
• Add initial pre-configured instrument by @bnmajor in #602
• Frame cache writer by @donald-e-boyce in #605
• Imagefiles yaml text by @donald-e-boyce in #604
• Remove tbb requirement for non-intel machines by @psavery in #606
• cleans up rawimage _get_length() and get_item() functions by @donald-e-boyce in #610
• Make getitem for raw imageseries threadsafe by @psavery in #611
• Add xtensor support, remove Eigen by @psavery in #607
• Add a test for gvec_to_xy by @psavery in #613
• Move projections into their own module by @psavery in #614
• fixes stiffness matrix typo by @donald-e-boyce in #618
• Add additional detector templates by @bnmajor in #617
• Add hedm_intensity to PlaneData by @psavery in #619
• fix arguments to cylindrical detector. by @saransh13 in #621
• Imageseries docs by @donald-e-boyce in #620
• Do not set values to 0 that are not referenced. by @ZackAttack614 in #624
• Fix hexrd misspelling, resolves #622 by @ZackAttack614 in #627
• Stacking fault wppf by @saransh13 in #626
• Refactor calibrators to work with lmfit by @psavery in #628
• Remove panel buffers from detector templates by @psavery in #629
• Use scipy's convolve for snip1d by @psavery in #630
• Update ca certificates for centos Linux build by @psavery in #632
• Update numpy and qt in Ryan's pinhole script by @psavery in #633
• Invalidate sample layer past critical beta by @psavery in #635
• Fix integer overflow issue in fabio by @psavery in #623
• Downgrade packaging macos to 12 by @psavery in #636

Full Changelog: 0.9.4...0.9.5

Release 0.9.4

• Added support for Python3.11, resulting in noticeable performance improvements
• Added conda packages for osx-arm64 (Apple computers after 2019)
• Implemented build time optimization that results in performance improvements
• Added ability to cancel fit-grains programmatically
• Changed PXRDIP default polar angle to 67.5
• Improved TARDIS distance between plates constraint
• Implemented a few bug fixes, including one with memoization

Release 0.9.3

A variety of new features have been added, including:

• support for a cylindrical detector (used primarily in TARDIS currently)
• new Rygg pinhole distortion method (used primarily for TARDIS and PXRDIP)
• new structureless calibration class
• support for GEL files

For the HEDM workflow:

• large performance improvements to fit-grains (including through the use of updated C transforms, and a major reduction in the amount of memory used)
• namedtuple grains output
• fixed an HKL reset issue in fit-grains
• new options for selecting which HKLs to exclude when running from the CLI

Minor updates for LLNL pinhole camera utilities

Release for FY23Q1 LLNL tutorials

What's Changed

• modified Y translations per JHE's TBSep=23.13mm by @joelvbernier in #482
• added notch to back IP by @joelvbernier in #483

Full Changelog: 0.9.1...0.9.2

Running changes for hexrdgui 0.9.0 compatability

What's Changed

• Create NOTICE by @joelvbernier in #473
• script to calculate pinhole 2theta by @joelvbernier in #474
• added a simple form for pinhole line tth distortion by @joelvbernier in #475
• Add calculating symmetric vectors in cartesian space. by @saransh13 in #476
• Added ability to import HDF5 imageseries with 2-d data by @joelvbernier in #477
• fixed the bad choice of always casting as array by @joelvbernier in #478
• WPPF -- Rietveld fixes by @saransh13 in #479
• add missing colon. by @saransh13 in #481
• Nf chunking by @rachelelim in #469

Full Changelog: 0.9.0...0.9.1

API change and pinhole camera distortion tools

Major release

In addition to the API change in the ff-HEDM config object and associated η-ω maps, the PowderCalibrator class can now accept an xrdutil.phutil.SampleLayerDistortion class to apply the 2θ distortion associated with an offset sample layer in a pinhole camera instrument (e.g., TARDIS, PXRDIP). The corresponding class for actual pinhole lines themselves is not yet implemented (staged, using functions from @ryanrygg). Below is a simulated TARDIS image with the cyan and orange overlays coming from 5μm thick layers at 150μm and 250μm offsets, respectively; the aforementioned distortion caused the nominal 2θ values to appear curved (note the simulation captures the distortion effect on account of being voxelated, while the correction follows @ryanrygg's Eq 10 from the TARDIS paper.

As for the ff-HEDM config API, now the hkls used for constructing the η-ω maps are specified by the explicit list of (h, k, l) indices (any symmetric equivalent) or the list of global G-vector ids from the material (less preferred). This removes the ambiguity associated with the previous convention of using indices into the reduced hkl list. For example,:

find_orientations:
  orientation_maps:
    # A file name must be specified. If it doesn't exist, one will be created
    file: null

    threshold: 250 # threshold on raw images to build maps
    bin_frames: 1 # how many successive frames to bin. default is 1

    # specify which hkls to use
    # "all", or a list of hkl indices.  Note that the default value of "all"
    # means all hkls that are not marked as excluded in the material definition.
    # can be a list of indices or a list of [h, k, l].
    # !!! used to be active_hkls: [0,1,2,3,4,5]
    # !!! this was really [0,1,2,4,6,7] considering the material parameters above
    active_hkls:
      - [0, 1, 2]
      - [1, 0, 4]
      - [1, 1, 0]
      - [1, 1, 3]
      - [0, 2, 4]
      - [1, 1, 6]

What's Changed

• replacing "if" with "elif" in line 604 of materials.py. by @saransh13 in #465
• Guard against unicode when saving HDF5 imageseries by @psavery in #466
• Pinhole distortion by @joelvbernier in #446
• changed hkl specification for indexing to master hklID by @joelvbernier in #454
• Update test.yml by @joelvbernier in #467
• change default source_distance to inf by @joelvbernier in #468
• NF chunking by @rachelelim in #461
• Additional modifications to make tth_distortion for pinhole camera instruments functional by @joelvbernier in #470
• API change to tth_corr_map_sample_layer by @joelvbernier in #471
• Add missing init.py files by @psavery in #472

New Contributors

• @rachelelim made their first contribution in #461

Full Changelog: 0.8.21...0.9.0

Minor bug fix to multi-wedge indexing

There was a check added to the multi-wedge capability in the ff-HEDM indexing workflow that was throwing a TypeError in hexrd 0.8.20. This release resolved that issue. There was also an issue in the CI with fabio relating to the lack of a wheel for the latest.

What's Changed

• Pin fabio to 0.12 on the Mac by @cjh1 in #462

Full Changelog: 0.8.20...0.8.21

Stability and performance upgrades.

What's Changed

• density fix for ruby by @saransh13 in #436
• Plane data updates by @joelvbernier in #437
• Material load by @saransh13 in #438
• some tweaks to the HEDM calibration by @joelvbernier in #439
• restore functionality with array-like tth input by @joelvbernier in #440
• Material speedup by @saransh13 in #431
• Spectrum fitting fixes by @joelvbernier in #442
• add pinhole util module by @joelvbernier in #443
• LP split by @saransh13 in #444
• Unitcell fixes by @saransh13 in #445
• Add support for Python 3.10 by @cjh1 in #447
• Fix use of collections.Sequence by @cjh1 in #448
• Fix use of collections.abc.Iterator by @cjh1 in #449
• Remove duplicate atom sync by @saransh13 in #450
• Bump pinned numpy version to 1.22 by @psavery in #458
• Force cast input to ndarray in GE distortion functions by @joelvbernier in #456
• multiwedge tolerance fix on delta omega by @joelvbernier in #459
• Remove hkl_indices comparison from testing by @psavery in #460
• Update spectrum fitting utils to support wildcards in parameter name matching by @joelvbernier in #455

Full Changelog: 0.8.19...0.8.20

New features and performance enhancements

Some salient updates

1. HEDM indexer can now handle discontinuities in ω range ("broken scans")
2. Robustness added to powder line fitting via lmfit models with constraints
3. Various bug fixes and platform-specific stablility fixes.

What's Changed

• Bump numpy version by @psavery in #397
• Fix multiprocessing context for ProcessPoolExecutor by @joelvbernier in #398
• Check shape of hkl array before write by @saransh13 in #400
• add tqdm package dependency by @joelvbernier in #401
• Fitting fixes jvb by @joelvbernier in #389
• Use logger for InstrumentCalibrator by @psavery in #402
• fix atomic weight of berkelium (Bk) by @saransh13 in #405
• catch None in hdf5 write as TypeError by @joelvbernier in #403
• Powder line extraction by @joelvbernier in #407
• Catch for flat patch data by @joelvbernier in #406
• Update nfutil.py by @darrencpagan in #408
• Update nfutil.py by @darrencpagan in #409
• Downgrade windows CI to windows-2019 by @psavery in #412
• Fix spot export error for multiple grains by @psavery in #429
• Set default dmin to None in load_materials_hdf5() by @psavery in #428
• Update rotations.py by @joelvbernier in #433
• Add a gaussian laplace op to ProcessedImageSeries by @psavery in #430
• attribute for picks and logging by @joelvbernier in #404

Full Changelog: 0.8.18...0.8.19

Bug fixes and a few more features.

What's Changed

• addresses case where cluster has 2 orientations by @joelvbernier in #381
• Simulated peak by @saransh13 in #331
• routines to get absorption length of a material by @saransh13 in #380
• fixing issues with spline background estimation by @saransh13 in #383
• Move generating parameter names into a function by @psavery in #384
• Fix parameter names by @psavery in #385
• fixes to projection code by @joelvbernier in #386
• save two theta width in h5 file. by @saransh13 in #387
• Raw image imageseries by @donald-e-boyce in #388
• Rotations update by @joelvbernier in #391
• Update utils.py by @joelvbernier in #393
• Fix default material tThWidth for HDF5 by @psavery in #395
• Remove duplicate atom fix by @saransh13 in #394
• Fixing write error in material by @saransh13 in #396

Full Changelog: 0.8.17...0.8.18