Niworkflows

======================
A bugfix release ensuring compatibility with Python 3.7.

* FIX: Add ``unlink()`` with ``missing_ok`` for Python 3.7 (695)

======================
A bugfix release addressing several issues and improving the carpet plot.

* FIX: Set ``pixdim[4]`` to match ``RepetitionTime`` (679)
* FIX: Unlink destination files before saving derivative (683)
* FIX: Smarter decision to remap input segmentation to carpetplot (691)
* FIX: Error handling the segmentation in NIfTI extraction (689)
* ENH: Add a legend to carpet plots with more than one segment (690)
* ENH: Miscellaneous improvements to carpetplot (688)
* ENH: Refactor carpetplot reasigning responsibilities (687)
* ENH: Initiate a module of nipype interfaces for mathematical morphology (651)
* ENH: Reorder within-segment rows of carpetplot with hierarchical clustering (686)
* DOC: Fix doc builds in ``maint/1.3.x`` (692)

=========================
First release in the 1.5.x series.
This release upgrades the Dockerfile to use FSL6, and includes some new interfaces.

* ENH: Additional nibabel interfaces (680)
* ENH: Add utility function to generate sample BIDS directory (677)
* ENH: Update to FSL6 (675)
* FIX: Scipy docs path (681)
* TEST: Drop excessively long interface equivalence tests (674)

===========================
Bug-fix release in the 1.4.x series

Backports and minor maintenance.

* FIX: TSV2JSON should convert empty TSV files to empty JSON files (747)
* FIX: Stop relying on deprecated indexing behavior (730)
* MNT: Update docs and GitHub actions config from master in LTS branch (748)

======================
Bug-fix release in the 1.4.x series.

This release includes improvements to the reliability of the BOLD masking workflow,
as well as an fMRIPrep bug-fix when reusing multi-echo anatomicals.

* FIX: Account for potential lists of lists in multi-echo cases (719)
* FIX: Improve reliability of BOLD masking workflow (712)

======================
Bug-fix release in the 1.4.x series.

This change enables a bug-fix in sdcflows where images with slightly
different affines were unable to be concatenated for use with TOPUP.

* ENH: Add ``affine_tolerance`` flag to ``MergeSeries`` (706)