Smriprep

=========================
Bug-fix release in the 0.7.x series.

All releases since 0.4.0 have incorrectly labelled T1w images normalized to a
template space as SkullStripped in the corresponding JSON sidecar files.
This affects 0.4.x through 0.4.2, 0.5.x through 0.5.3, 0.6.x through 0.6.2, and
0.7.0. Prior to 0.4.0, the images were actually skull-stripped, and the metadata
labels were not incorrect.

For backwards compatibility reasons, any future releases of these series will
have SkullStripped set to False. In 0.8 and above, the images will be skull-stripped
and the metadata set back to True.

* CI: CircleCI housekeeping (234, 235)

==========================
Minor release in preparation for *fMRIPrep* 20.2.x LTS series.
Includes minor features and bug-fixes over the previous 0.6 series.

* FIX: Pin *NiWorkflows* 1.3.1 including bugfix for INU-correction failures (nipreps/niworkflows567)
* FIX: Generate anatomical conversions with full spec from ``--output-spaces`` (219)
* FIX: Ordering of ``probseg`` maps with anatomical *fast-track* (214)
* FIX: Progress partial volume maps instead of posteriors (FSL FAST) (213)
* ENH: Retain session info when multi-session data are not averaged (225)
* ENH: Update derivatives description, ``.bidsignore`` for derivatives (220)
* ENH: Add ``--no-tty`` option to Docker wrapper (216)
* ENH: Add function to handle stale ``IsRunning`` files (207)
* MAINT: Upgrade ANTs to 2.3.4 in ``Dockerfile`` (365673b)
* MAINT: Make workflows keyword-only (PEP 3102) (208)

====================
Bug-fix release addressing minor problems related to FreeSurfer handling.

* FIX: Adapt to the new FS canary interface (backwards compatible) (205)
* FIX: Use ``t1w2fsnative_xfm`` to resample segmentations (201)

====================
Hotfix release to address an issue recently encountered in fMRIPrep 20.1.0rc3.

* FIX: ``MultiLabel`` interpolations should not use ``float=True`` (196)

====================
Minor release in preparation for fMRIPrep 20.1.x series.
Features the new implementation of derivatives writers in NiWorkflows,
and additional flexibility to use previously computed results (in particular,
skull-stripped brains, and the new *fast-track* that allows skipping the
anatomical workflow in full, if all the expected derivatives are provided).
Most of the the bug-fixes correspond to amendments over these newly added
features.

* FIX: Convert LTA to ITK with nitransforms (188)
* FIX: Dismiss ``session`` entity on most of anatomical derivatives (193)
* FIX: Revise tissue probability maps connections and order (190)
* FIX: Make TPMs label ordering in ``io_spec.json`` consistent with workflow (179)
* FIX: Correct the ``dseg`` labeling from FSL FAST earlier (177)
* FIX: Ensure ``bias_corrected`` is single file, not list (174)
* ENH: Use new ``DerivativesDataSink`` from NiWorkflows 1.2.0 (183)
* ENH: Use FreeSurfer's canary to exit fast and with a clear message when the license is missing (182)
* ENH: Execute FSL FAST only with ``--fs-no-reconall`` (180)
* ENH: Enable anatomical fast track reusing existing derivatives (107)
* ENH: Add option to skip brain extraction (167)
* MAINT: Remove unused nwf interfaces (187)
* MAINT: Pin troublesome sphinx (175)
* MAINT: Update dependencies to be inline with fMRIPrep-20.1.x (173)

0.5.x Series
============

--------------------
Bug-fix release in the 0.5.x series.

This release fixes a bug where pre-run FreeSurfer that was not in alignment with the
T1w template generated by fMRIPrep could result in misaligned segmentation and mask
derivatives.

The bug is most likely to occur with pre-run FreeSurfer where multiple T1w images are found.
It is easily evident in the first figure in the anatomical section of the reports, and will
show heavily misaligned brain mask.

* FIX: Use t1w2fsnative_xfm to resample segmentations (201) effigies