24.0.2

Bug-fix release in the 24.0.x series.

What's Changed

• FIX: Add per-volume transforms as single transform in chain by @effigies in #226

Full Changelog: 24.0.1...24.0.2

24.0.1

Bug-fix release in the 24.0.x series.

New Contributors

• @shnizzedy made their first contribution in #222

CHANGES

• FIX: Use standard library pathlib by @shnizzedy in #222
• MAINT: Support pre-__or__ types by @effigies in #223
• MAINT: Bump the actions-infrastructure group with 3 updates by @dependabot in #224
• MAINT: Bump codecov/codecov-action from 3 to 4 by @dependabot in #225

Full Changelog: 24.0.0...24.0.1

24.0.0

A new series incorporating several major changes, including bugfixes and taking on several housekeeping/maintenance actions. One relevant change is the outsourcing of the apply() member out of transformation data structures by @jmarabotto. The method apply() is now a standalone method that operates on one transform and images/surfaces/etc. provided as arguments. A later major development is the adoption of a foundation for surface transforms by @feilong and @Shotgunosine.

What's Changed

• FIX: Inefficient iterative reloading of reference and moving images by @oesteban in #186
• FIX: Postpone coordinate mapping on linear array transforms by @oesteban in #187
• FIX: Remove unsafe cast during TransformBase.apply() by @effigies in #189
• FIX: _is_oblique() by @mvdoc in #194
• FIX: Update implementation of ndim property of transforms by @jmarabotto in #197
• FIX: Output displacement fields by @bpinsard in #182
• FIX: Composition of deformation fields by @jbanusco in #188
• FIX: Indexing disallowed in lists introduced by bugfix by @oesteban in #204
• FIX: Do not transpose (see :obj:~scipy.ndimage.map_coordinates) by @oesteban in #207
• FIX: Forgotten test using xfm.apply() by @oesteban in #208
• FIX: Load ITK fields from H5 correctly by @effigies in #211
• FIX: Wrong warning argument name level in warnings.warn by @oesteban in #216
• ENH: Define ndim property on nonlinear transforms by @oesteban in #201
• ENH: Outsource apply() from transform objects by @jmarabotto in #195
• ENH: Restore apply() method, warning of deprecation and calling function by @effigies in #209
• ENH: SurfaceTransform class by @feilong in #203
• ENH: reenable-parallelization-apply-214 (builds on PR #215, solves Issue #214) by @jmarabotto in #217
• ENH: Parallelize serialized 3D+t transforms by @oesteban in #220
• ENH: Implement a memory limitation mechanism in loading data by @oesteban in #221
• ENH: Serialize+parallelize 4D apply() into 3D+t and add 'low memory' loading by @oesteban in #215
• MAINT: Loosen dependencies by @mgxd in #164
• MAINT: Drop Python 3.7 support, test through 3.11 by @effigies in #181
• MAINT: Update CircleCI's infrastructure (machine image and Python version in Docker image) by @oesteban in #206
• MAINT: Fix tests for Python 3.12, numpy 2.0, and pytest-xdist by @effigies in #210
• MAINT: Update ANTs' pinnings by @oesteban in #219

New Contributors

• @mvdoc made their first contribution in #194
• @jmarabotto made their first contribution in #197
• @bpinsard made their first contribution in #182
• @jbanusco made their first contribution in #188
• @feilong made their first contribution in #203

Full Changelog: 23.0.1...24.0.0

23.0.1

What's Changed

• FIX: numpy deprecation errors after 1.22 by @oesteban in #180
• FIX: Load ITK's .mat files with Affine's loaders by @oesteban in #179

Full Changelog: 23.0.0...23.0.1

23.0.0

Release Notes

A new major release preparing for the finalization of the package and migration into
NiBabel, mostly addressing bugfixes and scheduled added new features.

What's Changed

• FIX: Ensure datatype of generated CIFTI2 file in TransformBase unit test by @oesteban in #178
• FIX: Set x-forms on resampled images by @oesteban in #176
• ENH: API change in TransformChain - new composition convention by @oesteban in #165
• ENH: Extend the nonlinear transforms API by @oesteban in #166
• ENH: "Densify" voxel-wise nonlinear mappings with interpolation by @oesteban in #168
• ENH: Read ITK's composite transforms with only affines by @oesteban in #174
• MAINT: Rotate CircleCI secrets and setup up org-level context by @oesteban in #172

Full Changelog: 22.0.0...23.0.0

21.0.1

What's Changed

• Add myself as a contributor by @HippocampusGirl in #159
• FIX: Wrong datatype used for offset when reading ITK's h5 fields. by @oesteban in #161
• FIX: Orientation of displacements field and header when reading ITK's h5 by @oesteban in #162
• MAINT: Conclude migration poldracklab -> nipy by @oesteban in #163
• ENH: Guess open linear transform formats by @oesteban in #160

Full Changelog: 22.0.0...21.0.1

22.0.0

Release Notes

The first stable release of NiTransforms in 2022.
Contains all the new bug-fixes, features, and maintenance executed within the context of the NiBabel EOSS4 grant from the CZI Foundation.

CHANGES

• FIX: Implement AFNI's deoblique operations (#117)
• FIX: Ensure input dtype is kept after resampling (#153)
• FIX: Replace deprecated _read_mat with scipy.io.loadmat (#151)
• FIX: Add FSL-LTA-FSL regression tests (#146)
• FIX: Increase FSL serialization precision (#144)
• FIX: Refactor of LTA implementation (#145)
• FIX: Load arrays of linear transforms from AFNI files (#143)
• FIX: Load arrays of linear transforms from FSL files (#142)
• FIX: Double-check dtypes within tests and increase RMSE tolerance (#141)
• ENH: Base implementation of B-Spline transforms (#138)
• ENH: I/O of FSL displacements fields (#51)
• MAINT: Fix path to test summaries in CircleCI (#148)
• MAINT: Move testdata on to gin.g-node.org & datalad (#140)
• MAINT: scipy-1.8, numpy-1.22 require python 3.8 (#139)

New Contributors

• @HippocampusGirl made their first contribution in #151

Full Changelog: 21.0.1...22.0.0

21.0.0

Release Notes

A first release of NiTransforms. This release accompanies a corresponding JOSS submission.

CHANGES

• FIX: Final edits to JOSS submission (#135)
• FIX: Add mention to potential alternatives in JOSS submission (#132)
• FIX: Misinterpretation of voxel ordering in LTAs (#129)
• FIX: Suggested edits to the JOSS submission (#121)
• FIX: Invalid DOI (#124)
• FIX: Remove the --inv flag from regression mri_vol2vol regression test (#78)
• FIX: Improve handling of optional fields in LTA (#65)
• FIX: LTA conversions (#36)
• ENH: Add more comprehensive comments to notebook (#134)
• ENH: Add an .asaffine() member to TransformChain (#90)
• ENH: Read (and apply) ITK/ANTs' composite HDF5 transforms (#79)
• ENH: Improved testing of LTA handling - ITK-to-LTA, mri_concatenate_lta (#75)
• ENH: Add FS transform regression (#74)
• ENH: Add ITK-LTA conversion test (#66)
• ENH: Support for transforms mappings (e.g., head-motion correction) (#59)
• ENH: command line interface (#55)
• ENH: Facilitate loading of displacements field transforms (#54)
• ENH: First implementation of AFNI displacement fields (#50)
• ENH: Base implementation of transforms chains (composition) (#43)
• ENH: First implementation of loading and applying ITK displacements fields (#42)
• ENH: Refactor of AFNI and FSL I/O with StringStructs (#39)
• ENH: More comprehensive implementation of ITK affines I/O (#35)
• ENH: Added some minimal test-cases to the Affine class (#33)
• ENH: Rewrite load/save utilities for ITK's MatrixOffsetBased transforms in io (#31)
• ENH: Rename resample() with apply() (#30)
• ENH: Write tests pulling up the coverage of base submodule (#28)
• ENH: Add tests and implementation for Displacements fields and refactor linear accordingly (#27)
• ENH: Uber-refactor of code style, method names, etc. (#24)
• ENH: Increase coverage of linear transforms code (#23)
• ENH: FreeSurfer LTA file support (#17)
• ENH: Use obliquity directly from nibabel (#18)
• ENH: Setting up a battery of tests (#9)
• ENH: Revise doctests and get them ready for more thorough testing. (#10)
• DOC: Add Zenodo metadata record (#136)
• DOC: Better document the IPython notebooks (#133)
• DOC: Transfer CoC from NiBabel (#131)
• DOC: Clarify integration plans with NiBabel in the README (#128)
• DOC: Add contributing page to RTD (#130)
• DOC: Add CONTRIBUTING.md file pointing at NiBabel (#127)
• DOC: Add example notebooks to sphinx documentation (#126)
• DOC: Add an Installation section (#122)
• DOC: Display API per module (#120)
• DOC: Add figure to JOSS draft / Add @smoia to author list (#61)
• DOC: Initial JOSS draft (#47)
• MAINT: Add imports of modules in __init__.py to workaround #91 (#92)
• MAINT: Fix missing python3 binary on CircleCI build job step (#85)
• MAINT: Use setuptools_scm to manage versioning (#83)
• MAINT: Split binary test-data out from gh repo (#84)
• MAINT: Add Docker image/circle build (#80)
• MAINT: Drop Python 3.5 (#77)
• MAINT: Better config on setup.py (binary operator starting line) (#60)
• MAINT: add docker build to travis matrix (#29)
• MAINT: testing coverage (#16)
• MAINT: pep8 complaints (#14)
• MAINT: skip unfinished implementation tests (#15)
• MAINT: pep8speaks (#13)

20.0.0rc5

A first beta release of NiTransforms with the purpose of integrating
it within *fMRIPrep*.
This release accompanies the `ISBI2020 preprint
<https://doi.org/10.31219/osf.io/8aq7b>`__.

  * FIX: Accept double precision when reading ANTs' ``.mat`` affines
(#102)
  * FIX: Remove the ``--inv`` flag from regression ``mri_vol2vol``
regression test (#78)
  * FIX: Improve handling of optional fields in LTA (#65)
  * FIX: LTA conversions (#36)
  * ENH: Add an ``.asaffine()`` member to ``TransformChain`` (#90)
  * ENH: Read (and apply) ITK/ANTs' composite HDF5 transforms (#79)
  * ENH: Improved testing of LTA handling - ITK-to-LTA,
``mri_concatenate_lta`` (#75)
  * ENH: Add FS transform regression (#74)
  * ENH: Add ITK-LTA conversion test (#66)
  * ENH: Support for transforms mappings (e.g., head-motion correction)
(#59)
  * ENH: command line interface (#55)
  * ENH: Facilitate loading of displacements field transforms (#54)
  * ENH: First implementation of AFNI displacement fields (#50)
  * ENH: Base implementation of transforms chains (composition) (#43)
  * ENH: First implementation of loading and applying ITK displacements
fields (#42)
  * ENH: Refactor of AFNI and FSL I/O with StringStructs (#39)
  * ENH: More comprehensive implementation of ITK affines I/O (#35)
  * ENH: Added some minimal test-cases to the Affine class (#33)
  * ENH: Rewrite load/save utilities for ITK's MatrixOffsetBased
transforms in ``io`` (#31)
  * ENH: Rename ``resample()`` with ``apply()`` (#30)
  * ENH: Write tests pulling up the coverage of base submodule (#28)
  * ENH: Add tests and implementation for Displacements fields and
refactor linear accordingly (#27)
  * ENH: Uber-refactor of code style, method names, etc. (#24)
  * ENH: Increase coverage of linear transforms code (#23)
  * ENH: FreeSurfer LTA file support (#17)
  * ENH: Use ``obliquity`` directly from nibabel (#18)
  * ENH: Setting up a battery of tests (#9)
  * ENH: Revise doctests and get them ready for more thorough testing.
(#10)
  * DOC: Update sphinx to latest major (4.0) (#116)
  * DOC: Initial setup for sphinx-build (#114)
  * DOC: Add Jupyter notebooks (#93)
  * DOC: Add figure to JOSS draft / Add @smoia to author list (#61)
  * DOC: Initial JOSS draft (#47)
  * MAINT: Move from TravisCI to GHA (#112)
  * MAINT: Revise tests and Docker image dependencies (#104)
  * MAINT: Add imports of modules in ``__init__.py`` to workaround #91
(#92)
  * MAINT: Fix missing python3 binary on CircleCI build job step (#85)
  * MAINT: Use setuptools_scm to manage versioning (#83)
  * MAINT: Split binary test-data out from gh repo (#84)
  * MAINT: Add docker image/circle build (#80)
  * MAINT: Drop Python 3.5 (#77)
  * MAINT: Better config on setup.py (binary operator starting line)
(#60)
  * MAINT: add docker build to travis matrix (#29)
  * MAINT: testing coverage (#16)
  * MAINT: pep8 complaints (#14)
  * MAINT: skip unfinished implementation tests (#15)
  * MAINT: pep8speaks (#13)

20.0.0rc4

20.0.0rc4 (May 14, 2021)

* FIX: Accept double precision when reading ANTs' ``.mat`` affines (#102)
* DOC: Add Jupyter notebooks (#93, #96, #97, #98)
* MAINT: Revise tests and Docker image dependencies (#104)
* MAINT: Add imports of modules in ``__init__.py`` to workaround #91 (#92)