1.0.0rc1

This is a prerelease for the version that will be used by the HBCD project for their first data release.
Starting with this release, QSIPrep is now being deployed to pennlinc/qsiprep
on DockerHub instead of pennbbl/qsiprep.

We plan to not introduce any backwards-incompatible changes for a while.

What's Changed

🛠 Breaking Changes

• Replace pennbbl with pennlinc in package by @tsalo in #869
• Move common metadata to top of JSON by @tsalo in #875

🐛 Bug Fixes

• [FIX] synthstrip extra arg by @mattcieslak in #881
• Pin nireports version and allow GIFs by @tsalo in #883

Other Changes

• Restructure documentation and document how to run tests locally by @tsalo in #879

Full Changelog: 0.24.0...1.0.0rc1

0.24.0

What's Changed

🛠 Breaking Changes

• Rename --anat-space-definition to --subject-anatomical-reference by @tsalo in #855
• Remove underscore versions of all parameters by @tsalo in #861
• Change T1wACPC to ACPC and T1wNative to anat by @tsalo in #859
• Change _cnr to stat-cnr_dwimap by @tsalo in #867
• Rename confounds file to desc-confounds_timeseries by @tsalo in #874

🎉 Exciting New Features

• Infer the appropriate MNIInfant cohort based on the participant's age by @tsalo in #856
• Make grouping account for acquisition entity by @tsalo in #862

🐛 Bug Fixes

• [FIX] Change to Linear for upsampling by @mattcieslak in #820
• Add MNIInfant to template enum by @tsalo in #824
• Add ability to ignore phase data by @tsalo in #849
• Use correct base name with --distortion-group-merge average by @tsalo in #863

Other Changes

• [CI] Swap synthstrip for bet when sloppy mode is chosen by @mattcieslak in #817
• [FIX] Constrain threading for intramodal template by @mattcieslak in #825
• [FIX] use correct arguments for merging distortion groups by @mattcieslak in #831
• rf(py314): Replace deprecated pkgutil.find_loader by @effigies in #828
• [ENH] Enable session selection by @mattcieslak in #832
• [FIX] Add denoising reportlets back to html by @mattcieslak in #841
• [ENH] Add derivatives metadata by @mattcieslak in #844
• Use "Apptainer" instead of "Singularity" by @arokem in #846
• Document the transforms produced by QSIPrep by @tsalo in #866
• Add warning for pennbbl Docker image by @tsalo in #870
• Add QSIPrep paper to boilerplate by @tsalo in #873
• Remove aggr_ses_reports field by @tsalo in #872

New Contributors

• @effigies made their first contribution in #828

Full Changelog: 0.23.0...0.24.0

0.22.0

A huge number of infrastructure upgrades to improve testing and test coverage. An update to PyAFQ! Also a number of bug fixes, including for the hsvs recon workflows

What's Changed

🐛 Bug Fixes

• Fix FUGUE bug by @tsalo in #769

Other Changes

• [ENH] Update to NiPreps-style config file by @mattcieslak in #745
• Fix API documentation by @tsalo in #748
• Run integration tests with pytest by @tsalo in #763
• MAINT: Upgrade pyAFQ to 1.3.2 by @arokem in #764
• [ENH] Update dependencies in docker image by @mattcieslak in #768
• [CI] add optional bundles by @mattcieslak in #771
• [FIX] Fix the mif2fib workflow by @mattcieslak in #778
• Save SIFT2 mu parameter by @araikes in #774
• Drop functions to read sidecar JSONs by @tsalo in #760
• [FIX] Use new bundle names by @mattcieslak in #783
• [FIX] Make hsvs workflow work by @mattcieslak in #784
• [ENH] Compress recon derivatives by @mattcieslak in #787

New Contributors

• @araikes made their first contribution in #774

Full Changelog: 0.21.4...0.22.0

0.21.4

What's Changed

The effort to update to current NiPreps standards has begun, plus another bugfix in the TOPUP workflow.

Other Changes

• Remove copy of Niworkflows in favor of dependency by @tsalo in #709
• [FIX] stop bizarre argsort behavior in best b=0 by @mattcieslak in #744

Full Changelog: 0.21.3...0.21.4

0.21.3

What's Changed

This update addresses an important coregistration bug (#740), makes CPU Eddy much more usable and fixes a bug in the PyAFQ workflow.

Other Changes

• [FIX] PyAFQ was not being written to derivatives by @mattcieslak in #738
• [FIX] use b0 ref from eddy-processed data by @mattcieslak in #739
• [MAINT] remove pypi by @mattcieslak in #742
• [ENH] add parallel argument for topup by @mattcieslak in #741
• [ENH] Allow multithreading in eddy by @mattcieslak in #743

Full Changelog: 0.21.1...0.21.3

0.21.1

What's Changed

There are important software updates in this release, along with a lot of infrastructure improvements.

Important Changes

• FSL version 6.0.7.8 is now in qsiprep. This contains 2 serious bugfixes. One has to do with susceptibility-by-volume correction and the other resulted in incorrect CNR values being calculated.
• DSI Studio has been updated to fix a bug in Neighboring DWI Correlation.
• You can use "mporder" in your eddy config json file and the slice timings will automatically be created and passed to eddy (even if you're concatenating runs)
• Recon workflows can now include a "bundle mapping" and "scalar mapping", where scalars created in individual workflows can be mapped to a template or summarized inside autotrack bundles. This does not do tract profiles or surface mapping - yet. See qsiprep/data/pipelines/hbcd_scalar_maps.json for an example.
• The recon derivatives are written approximately according to BEP-016

🎉 Exciting New Features

• [ENH] Add workflow to resample recon scalars to template space by @mattcieslak in #688
• Support complex-valued dwidenoising by @tsalo in #679
• [ENH] Add ReconScalars node to recon workflows by @mattcieslak in #683
• Pin current version of pyAFQ (1.3.1) by @arokem in #729
• [ENH] Add TORTOISE Estimator recon workflows by @mattcieslak in #674
• [ENH] Make Eddy's slice2vol much easier to use by @mattcieslak in #710

Other Changes

• Apply stylistic changes to workflows/base.py by @tsalo in #678
• Pin Nilearn version by @tsalo in #687
• Add a series of infrastructure files from ASLPrep/XCP-D by @tsalo in #684
• Run isort and remove unused imports by @tsalo in #690
• Apply stylistic changes to workflows/utils.py by @tsalo in #680
• Remove unused classes flagged by vulture by @tsalo in #693
• [RF] Use Black to reformat package by @mattcieslak in #701

Bugfixes

• [FIX] Prevent undecodable terminal output in calc_connectivity by @mattcieslak in #711
• [FIX] save recon anat files by @mattcieslak in #702
• [FIX] revert citeproc for old pandoc by @mattcieslak in #736
• [FIX] recon workflows when --anat-modality T2w was used by @mattcieslak in #734
• [FIX] remove old nodes from csdsi_3dshore.json by @kjamison in #733
• [ENH] Update TORTOISE for lower cuda memory use by @mattcieslak in #730

New Contributors

• @kjamison made their first contribution in #733

Full Changelog: 0.20.0...0.21.1

0.20.0

Please note there is no corresponding release on PyPI for this version

New features

• [ENH] allow topup+drbuddi for hbcd by @mattcieslak in #667
  • Adds a 2-stage distortion correction option --pepolar-method TOPUP+DRBUDDI, which will run TOPUP -> Eddy -> DRBUDDI
• [ENH] Use UKB processed data as input for recon workflows by @mattcieslak in #651
  • This adds the --recon-input-pipeline, which lets you run recon workflows on UKB data
• [ENH] Update to python 3.10 by @mattcieslak in #670
  • Hopefully this will address the hang-after-crashing problem in the recent releases

Bugfixes/Docs

• DOC: Add SMeisler and JHLegarreta to contributors list by @jhlegarreta in #642
• Fixes typos on FreeSurfer requirements for ss3t hsvs recon by @pcamach2 in #414
• Fix RTD build by @tsalo in #652
• ENH: conform bvals to shells separated by b0_threshold by @cookpa in #660
• [FIX] remove unneeded "method" from tracking by @mattcieslak in #641
• FIX: allow finding of lesion rois by @psadil in #659
• MISC: Remove outdated dsi_studio tracking parameters by @cookpa in #668
• [DOC] Add documentation for dsi_studio_autotrack reconstruction workflow by @valeriejill in #669
• [ENH] Update BIDS validator to 1.8.4 by @mattcieslak in #671

New Contributors

• @psadil made their first contribution in #659

Full Changelog: 0.19.1...0.20.0

0.19.1

What's Changed

Updates in performance stability in synthseg/synthstrip, bugfixes and a few documentation updates. Sincere thanks to @jhlegarreta, @mayayab and @pierre-nedelec for contributions.

• [FIX] Forces synthseg to use 1 thread by @mattcieslak in #614
• ENH: Rework the PR template and first time contributor message by @jhlegarreta in #622
• DOC: Remove the destination paths in container documentation by @jhlegarreta in #621
• [FIX] Make recon workflows work on multi-session data by @mattcieslak in #639
• Fix #630 — update in BIDS dataset_description by @pierre-nedelec in #631
• DOC: Document how to obtain help by @jhlegarreta in #637
• DOC: Clarify the output space convention by @jhlegarreta in #638
• DOC: add --recon-only flag in reconstruction.rst by @pierre-nedelec in #628

New Contributors

• @jhlegarreta made their first contribution in #622
• @pierre-nedelec made their first contribution in #631

Full Changelog: 0.19.0...0.19.1

0.19.0

A new version of DSI Studio that brings more stable multithreading to the AutoTrack workflow. Memory management is improved in synthseg and synthstrip

What's Changed

• [FIX] steinhardt computation by @mattcieslak in #603
• [CI] Add test for tensor-based head motion correction by @mattcieslak in #605
• [ENH] Stabilize autotrack performance by @mattcieslak in #604
• typo by @arovai in #606
• [ENH] limit the synths to 1 thread by @mattcieslak in #608

New Contributors

• @arovai made their first contribution in #606

Full Changelog: 0.18.1...0.19.0

0.18.0

What's Changed

• [FIX] Twine upload issue by @mattcieslak in #579
• [DOC] Add citations for anatomical workflow and dsi studio by @mattcieslak in #584
• [FIX] allow non-gzipped T1ws for recon workflows by @mattcieslak in #586

Full Changelog: 0.18.0alpha0...0.18.0