NiBabel is the successor to the much-loved PyNifti package. Here we list the
releases for both packages.
Nibabel releases
Most work on NiBabel so far has been by Matthew Brett (MB), Michael Hanke (MH)
Ben Cipollini (BC), Marc-Alexandre Côté (MC), Chris Markiewicz (CM), Stephan
Gerhard (SG), Eric Larson (EL), Yaroslav Halchenko (YOH) and Chris Cheng (CC).
References like “pr/298” refer to github pull request numbers.
2.3.1 (Tuesday 16 October 2018)
New features
- nib-diff command line tool for comparing image files (pr/617, pr/672,
pr/678) (CC, reviewed by YOH, Pradeep Raamana and CM)
Enhancements
- Speed up reading of numeric arrays in CIFTI2 (pr/655) (Michiel Cottaar,
reviewed by CM)
- Add ndim property to ArrayProxy and DataobjImage (pr/674) (CM,
reviewed by MB)
Bug fixes
- Deterministic deduction of slice ordering in degenerate cases (pr/647)
(YOH, reviewed by CM)
- Allow 0ms TR in MGH files (pr/653) (EL, reviewed by CM)
- Allow for PPC64 little-endian long doubles (pr/658) (MB, reviewed by CM)
- Correct construction of FreeSurfer annotation labels (pr/666) (CM, reviewed
by EL, Paul D. McCarthy)
- Fix logic for persisting filehandles with indexed-gzip (pr/679) (Paul D.
McCarthy, reviewed by CM)
Maintenance
- Fix semantic error in coordinate systems documentation (pr/646) (Ariel
Rokem, reviewed by CM, MB)
- Test on Python 3.7, minor associated fixes (pr/651) (CM, reviewed by Gregory
R. Lee, MB)
2.3 (Tuesday 12 June 2018)
New features
- TRK <=> TCK streamlines conversion CLI tools (pr/606) (MC, reviewed by CM)
- Image slicing for SpatialImages (pr/550) (CM)
Enhancements
- Simplfiy MGHImage and add footer fields (pr/569) (CM, reviewed by MB)
- Force sform/qform codes to be ints, rather than numpy types (pr/575) (Paul
McCarthy, reviewed by MB, CM)
- Auto-fill color table in FreeSurfer annotation file (pr/592) (Paul McCarthy,
reviewed by CM, MB)
- Set default intent code for CIFTI2 images (pr/604) (Mathias Goncalves,
reviewed by CM, Satra Ghosh, MB, Tim Coalson)
- Raise informative error on empty files (pr/611) (Pradeep Raamana, reviewed
by CM, MB)
- Accept degenerate filenames such as .nii (pr/621) (Dimitri
Papadopoulos-Orfanos, reviewed by Yaroslav Halchenko)
- Take advantage of IndexedGzipFile drop_handles flag to release
filehandles by default (pr/614) (Paul McCarthy, reviewed by CM, MB)
Bug fixes
- Preserve first point of LazyTractogram (pr/588) (MC, reviewed by Nil
Goyette, CM, MB)
- Stop adding extraneous metadata padding (pr/593) (Jon Stutters, reviewed by
CM, MB)
- Accept lower-case orientation codes in TRK files (pr/600) (Kesshi Jordan,
MB, reviewed by MB, MC, CM)
- Annotation file reading (pr/592) (Paul McCarthy, reviewed by CM, MB)
- Fix buffer size calculation in ArraySequence (pr/597) (Serge Koudoro,
reviewed by MC, MB, Eleftherios Garyfallidis, CM)
- Resolve UnboundLocalError in Python 3 (pr/607) (Jakub Kaczmarzyk,
reviewed by MB, CM)
- Do not crash on non-ImportError failures in optional imports (pr/618)
(Yaroslav Halchenko, reviewed by CM)
- Return original array from get_fdata for array image, if no cast
required (pr/638, MB, reviewed by CM)
Maintenance
- Use SSH address to use key-based auth (pr/587) (CM, reviewed by MB)
- Fix doctests for numpy 1.14 array printing (pr/591) (MB, reviewed by CM)
- Refactor for pydicom 1.0 API changes (pr/599) (MB, reviewed by CM)
- Increase test coverage, remove unreachable code (pr/602) (CM, reviewed by
Yaroslav Halchenko, MB)
- Move nib-ls and other programs to a new cmdline module (pr/601, pr/615)
(Chris Cheng, reviewed by MB, Yaroslav Halchenko)
- Remove deprecated numpy indexing (EL, reviewed by CM)
- Update documentation to encourage get_fdata over get_data (pr/637,
MB, reviewed by CM)
API changes and deprecations
- Support for keep_file_open = 'auto' as a parameter to Opener() will
be deprecated in 2.4, for removal in 3.0. Accordingly, support for
openers.KEEP_FILE_OPEN_DEFAULT = 'auto' will be dropped on the same
schedule.
- Drop-in support for indexed_gzip < 0.7 has been removed.
2.2.1 (Wednesday 22 November 2017)
Bug fixes
- Set L/R labels in orthoview correctly (pr/564) (CM)
- Defer use of ufunc / memmap test - allows “freezing” (pr/572) (MB, reviewed
by Satra Ghosh)
- Fix doctest failures with pre-release numpy (pr/582) (MB, reviewed by CM)
Maintenance
- Update documentation around NIfTI qform/sform codes (pr/576) (Paul McCarthy,
reviewed by MB, CM) + (pr/580) (Bennet Fauber, reviewed by Paul McCarthy)
- Skip precision test on macOS, newer numpy (pr/583) (MB, reviewed by CM)
- Simplify AppVeyor script, removing conda (pr/584) (MB, reviewed by CM)
2.2 (Friday 13 October 2017)
New features
- CIFTI support (pr/249) (Satra Ghosh, Michiel Cottaar, BC, CM, Demian
Wassermann, MB)
- Support for MRtrix TCK streamlines file format (pr/486) (MC, reviewed by
MB, Arnaud Bore, J-Donald Tournier, Jean-Christophe Houde)
- Added get_fdata() as default method to retrieve scaled floating point
data from ``DataobjImage``s (pr/551) (MB, reviewed by CM, Satra Ghosh)
Enhancements
- Support for alternative header field name variants in .PAR files
(pr/507) (Gregory R. Lee)
- Various enhancements to streamlines API by MC: support for reading TRK
version 1 (pr/512); concatenation of tractograms using +/+= operators
(pr/495); function to concatenate multiple ArraySequence objects (pr/494)
- Support for numpy 1.12 (pr/500, pr/502) (MC, MB)
- Allow dtype specifiers as fileslice input (pr/485) (MB)
- Support “headerless” ArrayProxy specification, enabling memory-efficient
ArrayProxy reshaping (pr/521) (CM)
- Allow unknown NIfTI intent codes, add FSL codes (pr/528) (Paul McCarthy)
- Improve error handling for img.__getitem__ (pr/533) (Ariel Rokem)
- Delegate reorientation to SpatialImage classes (pr/544) (Mark Hymers, CM,
reviewed by MB)
- Enable using indexed_gzip to reduce memory usage when reading from
gzipped NIfTI and MGH files (pr/552) (Paul McCarthy, reviewed by MB, CM)
Bug fixes
- Miscellaneous MINC reader fixes (pr/493) (Robert D. Vincent, reviewed by CM,
MB)
- Fix corner case in wrapstruct.get (pr/516) (Paul McCarthy, reviewed by
CM, MB)
Maintenance
- Fix documentation errors (pr/517, pr/536) (Fernando Perez, Venky Reddy)
- Documentation update (pr/514) (Ivan Gonzalez)
- Update testing to use pre-release builds of dependencies (pr/509) (MB)
- Better warnings when nibabel not on path (pr/503) (MB)
API changes and deprecations
- header argument to ArrayProxy.__init__ is renamed to spec
- Deprecation of header property of ArrayProxy object, for removal in
3.0
- wrapstruct.get now returns entries evaluating False, instead of None
- DataobjImage.get_data to be deprecated April 2018, scheduled for removal
April 2020
2.1 (Monday 22 August 2016)
New features
- New API for managing streamlines and their different file formats. This
adds a new module nibabel.streamlines that will eventually deprecate
the current trackvis reader found in nibabel.trackvis (pr/391) (MC,
reviewed by Jean-Christophe Houde, Bago Amirbekian, Eleftherios
Garyfallidis, Samuel St-Jean, MB);
- A prototype image viewer using matplotlib (pr/404) (EL, based on a
proto-prototype by Paul Ivanov) (Reviewed by Gregory R. Lee, MB);
- Functions for image resampling and smoothing using scipy ndimage (pr/255)
(MB, reviewed by EL, BC);
- Add ability to write FreeSurfer morphology data (pr/414) (CM, BC, reviewed
by BC);
- Read and write support for DICOM tags in NIfTI Extended Header using
pydicom (pr/296) (Eric Kastman).
Enhancements
- Extensions to FreeSurfer module to fix reading and writing of FreeSurfer
geometry data (pr/460) (Alexandre Gramfort, Jaakko Leppäkangas, reviewed
by EL, CM, MB);
- Various improvements to PAR / REC handling by Gregory R. Lee: supporting
multiple TR values (pr/429); output of volume labels (pr/427); fix for
some diffusion files (pr/426); option for more sophisticated sorting of
volumes (pr/409);
- Original trackvis reader will now allow final streamline to have fewer
points than the number declared in the header, with strict=False
argument to read function;
- Helper function to return voxel sizes from an affine matrix (pr/413);
- Fixes to DICOM multiframe reading to avoid assumptions on the position of
the multiframe index (pr/439) (Eric M. Baker);
- More robust handling of “CSA” private information in DICOM files (pr/393)
(Brendan Moloney);
- More explicit error when trying to read image from non-existent file
(pr/455) (Ariel Rokem);
- Extension to nib-ls command to show image statistics (pr/437) and other
header files (pr/348) (Yarik Halchenko).
Bug fixes
- Fixes to rotation order to generate affine matrices of PAR / REC files (MB,
Gregory R Lee).
Maintenance
- Dropped support for Pythons 2.6 and 3.2;
- Comprehensive refactor and generalization of surface / GIFTI file support
with improved API and extended tests (pr/352-355, pr/360, pr/365, pr/403)
(BC, reviewed by CM, MB);
- Refactor of image classes (pr/328, pr/329) (BC, reviewed by CM);
- Better Appveyor testing on new Python versions (pr/446) (Ariel Rokem);
- Fix shebang lines in scripts for correct install into virtualenvs via pip
(pr/434);
- Various fixes for numpy, matplotlib, and PIL / Pillow compatibility (CM,
Ariel Rokem, MB);
- Improved test framework for warnings (pr/345) (BC, reviewed by CM, MB);
- New decorator to specify start and end versions for deprecation warnings
(MB, reviewed by CM);
- Write qform affine matrix to NIfTI images output by parrec2nii (pr/478)
(Jasper J.F. van den Bosch, reviewed by Gregory R. Lee, MB).
API changes and deprecations
- Minor API breakage in original (rather than new) trackvis reader. We are now
raising a DataError if there are too few streamlines in the file,
instead of a HeaderError. We are raising a DataError if the track
is truncated when strict=True (the default), rather than a TypeError
when trying to create the points array.
- Change sform code that parrec2nii script writes to NIfTI images; change
from 2 (“aligned”) to 1 (“scanner”);
- Deprecation of get_header, get_affine method of image objects for
removal in version 4.0;
- Removed broken from_filespec method from image objects, and deprecated
from_filespec method of ECAT image objects for removal in 4.0;
- Deprecation of class_map instance in imageclasses module in favor of
new image class attributes, for removal in 4.0;
- Deprecation of ext_map instance in imageclasses module in favor of
new image loading API, for removal in 4.0;
- Deprecation of Header class in favor of SpatialHeader, for removal
in 4.0;
- Deprecation of BinOpener class in favor of more generic Opener
class, for removal in 4.0;
- Deprecation of GiftiMetadata methods get_metadata and get_rgba;
GiftiDataArray methods get_metadata, get_labeltable,
set_labeltable; GiftiImage methods get_meta, set_meta. All
these deprecated in favor of corresponding properties, for removal in 4.0;
- Deprecation of giftiio read and write functions in favor of
nibabel load and save functions, for removal in 4.0;
- Deprecation of gifti.data_tag function, for removal in 4.0;
- Deprecation of write-access to GiftiDataArray.num_dim, and new error
when trying to set invalid values for num_dim. We will remove
write-access in 4.0;
- Deprecation of GiftiDataArray.from_array in favor of GiftiDataArray
constructor, for removal in 4.0;
- Deprecation of GiftiDataArray to_xml_open, to_xml_close methods in
favor of to_xml method, for removal in 4.0;
- Deprecation of parse_gifti_fast.Outputter class in favor of
GiftiImageParser, for removal in 4.0;
- Deprecation of parse_gifti_fast.parse_gifti_file function in favor of
GiftiImageParser.parse method, for removal in 4.0;
- Deprecation of loadsave functions guessed_image_type and
which_analyze_type, in favor of new API where each image class tests the
file for compatibility during load, for removal in 4.0.
2.0.2 (Monday 23 November 2015)
- Fix for integer overflow on large images (pr/325) (MB);
- Fix for Freesurfer nifti files with unusual dimensions (pr/332) (Chris
Markiewicz);
- Fix typos on benchmarks and tests (pr/336, pr/340, pr/347) (Chris
Markiewicz);
- Fix Windows install script (pr/339) (MB);
- Support for Python 3.5 (pr/363) (MB) and numpy 1.10 (pr/358) (Chris
Markiewicz);
- Update pydicom imports to permit version 1.0 (pr/379) (Chris Markiewicz);
- Workaround for Python 3.5.0 gzip regression (pr/383) (Ben Cipollini).
- tripwire.TripWire object now raises subclass of AttributeError when trying
to get an attribute, rather than a direct subclass of Exception. This
prevents Python 3.5 triggering the tripwire when doing inspection prior to
running doctests.
- Minor API change for tripwire.TripWire object; code that checked for
AttributeError will now also catch TripWireError.
2.0.1 (Saturday 27 June 2015)
Contributions from Ben Cipollini, Chris Markiewicz, Alexandre Gramfort,
Clemens Bauer, github user freec84.
- Bugfix release with minor new features;
- Added axis parameter to concat_images (pr/298) (Ben Cipollini);
- Fix for unsigned integer data types in ECAT images (pr/302) (MB, test data
and issue report from Github user freec84);
- Added new ECAT and Freesurfer data files to automated testing;
- Fix for Freesurfer labels error on early numpies (pr/307) (Alexandre
Gramfort);
- Fixes for PAR / REC header parsing (pr/312) (MB, issue reporting and test
data by Clemens C. C. Bauer);
- Workaround for reading Freesurfer ico7 surface files (pr/315) (Chris
Markiewicz);
- Changed to github pages for doc hosting;
- Changed docs to point to neuroimaging@python.org mailing list.
2.0.0 (Tuesday 9 December 2014)
This release had large contributions from Eric Larson, Brendan Moloney,
Nolan Nichols, Basile Pinsard, Chris Johnson and Nikolaas N. Oosterhof.
- New feature, bugfix release with minor API breakage;
- Minor API breakage: default write of NIfTI / Analyze image data offset
value. The data offset is the number of bytes from the beginning of file
to skip before reading the image data. Nibabel behavior changed from
keeping the value as read from file, to setting the offset to zero on
read, and setting the offset when writing the header. The value of the
offset will now be the minimum value necessary to make room for the header
and any extensions when writing the file. You can override the default
offset by setting value explicitly to some value other than zero. To read
the original data offset as read from the header, use the offset
property of the image dataobj attribute;
- Minor API breakage: data scaling in NIfTI / Analyze now set to NaN when
reading images. Data scaling refers to the data intercept and slope
values in the NIfTI / Analyze header. To read the original data scaling
you need to look at the slope and inter properties of the image
dataobj attribute. You can set scaling explicitly by setting the
slope and intercept values in the header to values other than NaN;
- New API for managing image caching; images have an in_memory property
that is true if the image data has been loaded into cache, or is already
an array in memory; get_data has new keyword argument caching to
specify whether the cache should be filled by get_data;
- Images now have properties dataobj, affine, header. We will
slowly phase out the get_affine and get_header image methods;
- The image dataobj can be sliced using an efficient algorithm to avoid
reading unnecessary data from disk. This makes it possible to do very
efficient reads of single volumes from a time series;
- NIfTI2 read / write support;
- Read support for MINC2;
- Much extended read support for PAR / REC, largely due to work from Eric
Larson and Gregory R. Lee on new code, advice and code review. Thanks also
to Jeff Stevenson and Bennett Landman for helpful discussion;
- parrec2nii script outputs images in LAS voxel orientation, which
appears to be necessary for compatibility with FSL dtifit /
fslview diffusion analysis pipeline;
- Preliminary support for Philips multiframe DICOM images (thanks to Nolan
Nichols, Ly Nguyen and Brendan Moloney);
- New function to save Freesurfer annotation files (by Github user ohinds);
- Method to return MGH format vox2ras_tkr affine (Eric Larson);
- A new API for reading unscaled data from NIfTI and other images, using
img.dataobj.get_unscaled(). Deprecate previous way of doing this,
which was to read data with the read_img_data function;
- Fix for bug when replacing NaN values with zero when writing floating
point data as integers. If the input floating point data range did not
include zero, then NaN would not get written to a value corresponding to
zero in the output;
- Improvements and bug fixes to image orientation calculation and DICOM
wrappers by Brendan Moloney;
- Bug fixes writing GIfTI files. We were using a base64 encoding that didn’t
match the spec, and the wrong field name for the endian code. Thanks to
Basile Pinsard and Russ Poldrack for diagnosis and fixes;
- Bug fix in freesurfer.read_annot with orig_ids=False when annot
contains vertices with no label (Alexandre Gramfort);
- More tutorials in the documentation, including introductory tutorial on
DICOM, and on coordinate systems;
- Lots of code refactoring, including moving to common code-base for Python
2 and Python 3;
- New mechanism to add images for tests via git submodules.
1.3.0 (Tuesday 11 September 2012)
Special thanks to Chris Johnson, Brendan Moloney and JB Poline.
- New feature and bugfix release
- Add ability to write Freesurfer triangle files (Chris Johnson)
- Relax threshold for detecting rank deficient affines in orientation
detection (JB Poline)
- Fix for DICOM slice normal numerical error (issue #137) (Brendan Moloney)
- Fix for Python 3 error when writing zero bytes for offset padding
1.2.2 (Wednesday 27 June 2012)
- Bugfix release
- Fix longdouble tests for Debian PPC (thanks to Yaroslav Halchecko for
finding and diagnosing these errors)
- Generalize longdouble tests in the hope of making them more robust
- Disable saving of float128 nifti type unless platform has real IEEE
binary128 longdouble type.
1.2.1 (Wednesday 13 June 2012)
Particular thanks to Yaroslav Halchecko for fixes and cleanups in this
release.
- Bugfix release
- Make compatible with pydicom 0.9.7
- Refactor, rename nifti diagnostic script to nib-nifti-dx
- Fix a bug causing an error when analyzing affines for orientation, when the
affine contained all 0 columns
- Add missing dicomfs script to installation list and rename to
nib-dicomfs
1.2.0 (Sunday 6 May 2012)
This release had large contributions from Krish Subramaniam, Alexandre
Gramfort, Cindee Madison, Félix C. Morency and Christian Haselgrove.
- New feature and bugfix release
- Freesurfer format support by Krish Subramaniam and Alexandre Gramfort.
- ECAT read write support by Cindee Madison and Félix C. Morency.
- A DICOM fuse filesystem by Christian Haselgrove.
- Much work on making data scaling on read and write more robust to rounding
error and overflow (MB).
- Import of nipy functions for working with affine transformation matrices.
- Added methods for working with nifti sform and qform fields by Bago
Amirbekian and MB, with useful discussion by Brendan Moloney.
- Fixes to read / write of RGB analyze images by Bago Amirbekian.
- Extensions to concat_images by Yannick Schwartz.
- A new nib-ls script to display information about neuroimaging files, and
various other useful fixes by Yaroslav Halchenko.
1.1.0 (Thursday 28 April 2011)
Special thanks to Chris Burns, Jarrod Millman and Yaroslav Halchenko.
- New feature release
- Python 3.2 support
- Substantially enhanced gifti reading support (SG)
- Refactoring of trackvis read / write to allow reading and writing of voxel
points and mm points in tracks. Deprecate use of negative voxel sizes;
set voxel_order field in trackvis header. Thanks to Chris Filo
Gorgolewski for pointing out the problem and Ruopeng Wang in the trackvis
forum for clarifying the coordinate system of trackvis files.
- Added routine to give approximate array orientation in form such as ‘RAS’
or ‘LPS’
- Fix numpy dtype hash errors for numpy 1.2.1
- Other bug fixes as for 1.0.2
1.0.2 (Thursday 14 April 2011)
- Bugfix release
- Make inference of data type more robust to changes in numpy dtype hashing
- Fix incorrect thresholds in quaternion calculation (thanks to Yarik H for
pointing this one out)
- Make parrec2nii pass over errors more gracefully
- More explicit checks for missing or None field in trackvis and other
classes - thanks to Marc-Alexandre Cote
- Make logging and error level work as expected - thanks to Yarik H
- Loading an image does not change qform or sform - thanks to Yarik H
- Allow 0 for nifti scaling as for spec - thanks to Yarik H
- nifti1.save now correctly saves single or pair images
1.0.1 (Wednesday 23 Feb 2011)
- Bugfix release
- Fix bugs in tests for data package paths
- Fix leaks of open filehandles when loading images (thanks to Gael
Varoquaux for the report)
- Skip rw tests for SPM images when scipy not installed
- Fix various windows-specific file issues for tests
- Fix incorrect reading of byte-swapped trackvis files
- Workaround for odd numpy dtype comparisons leading to header errors for
some loaded images (thanks to Cindee Madison for the report)
1.0.0 (Thursday, 13, Oct 2010)
- This is the first public release of the NiBabel package.
- NiBabel is a complete rewrite of the PyNifti package in pure python. It was
designed to make the code simpler and easier to work with. Like PyNifti,
NiBabel has fairly comprehensive NIfTI read and write support.
- Extended support for SPM Analyze images, including orientation affines from
matlab .mat files.
- Basic support for simple MINC 1.0 files (MB). Please let us know if you
have MINC files that we don’t support well.
- Support for reading and writing PAR/REC images (MH)
- parrec2nii script to convert PAR/REC images to NIfTI format (MH)
- Very preliminary, limited and highly experimental DICOM reading support (MB,
Ian Nimmo Smith).
- Some functions (nibabel.funcs) for basic image shape changes, including
the ability to transform to the image with data closest to the cononical
image orientation (first axis left-to-right, second back-to-front, third
down-to-up) (MB, Jonathan Taylor)
- Gifti format read and write support (preliminary) (Stephen Gerhard)
- Added utilities to use nipy-style data packages, by rip then edit of nipy
data package code (MB)
- Some improvements to release support (Jarrod Millman, MB, Fernando Perez)
- Huge downward step in the quality and coverage by the docs, caused by MB,
mostly fixed by a lot of good work by MH.
- NiBabel will not work with Python < 2.5, and we haven’t even tested it with
Python 3. We will get to it soon...