Release 1.4.2

[schuenke]

This is meant as the final PR before the release of version 1.4.2

The changelog is quite messy, so if nobody wants to create a more useful one, I would simply use the autogenerated changelog from GitHub, which I cleaned from temporary changes. Currently it would look like this:

---

What's Changed

• bug fix to event_lib.py by @tblazey in bug fix to event_lib.py #111
• sequence.py fix by @tblazey in sequence.py fix #110
• latest make_sinc_pulse dont return gradient by default by @yupbank in latest make_sinc_pulse dont return gradient by default #113
• Replace np.float with np.float64 by @btasdelen in Replace np.float with np.float64 #115
• Update gre_walkthrough for v1.4.0 by @btasdelen in Update gre_walkthrough for v1.4.0 #121
• Bug fixes that produced incorrect test reports by @FrankZijlstra in Bug fixes that produced incorrect test reports #125
• Hacky fix for waveforms_and_times rf append floating point issue. by @btasdelen in Hacky fix for waveforms_and_times rf append floating point issue. #114
• Bug fixes in Sequence and make_trapezoid by @FrankZijlstra in Bug fixes in Sequence and make_trapezoid #128
• More pythonic handling of shape_pieces in waveforms_and_times by @FrankZijlstra in More pythonic handling of shape_pieces in waveforms_and_times #130
• Several bug fixes in Sequence, add_gradients, make_trapezoid, and rotate by @FrankZijlstra in Several bug fixes in Sequence, add_gradients, make_trapezoid, and rotate #132
• Added PNS calculations by @FrankZijlstra in Added PNS calculations #131
• change sigpy dependency to GitHub repo by @schuenke in change sigpy dependency to GitHub repo #135
• Performance improvements by @FrankZijlstra in Performance improvements #133
• Add init,py to utils to help setuptools discover the package by @fzimmermann89 in Add __init__,py to utils to help setuptools discover the package #137
• Reworked piecewise polynomial implementation in Sequence to use scipy's PPoly by @FrankZijlstra in Reworked piecewise polynomial implementation in Sequence to use scipy's PPoly #140
• Enh: New options in make_arbitrary_rf.py for adiabatic pulses by @wtclarke in Enh: New options in make_arbitrary_rf.py for adiabatic pulses #143
• Enh: Enable GitHub CI by @wtclarke in Enh: Enable GitHub CI #142
• Bump sigpy version to 0.1.26 by @Andrew-Dupuis in Bump sigpy version to 0.1.26 #148
• BF: make_trapezoid with amplitude + duration as inputs. by @wtclarke in BF: make_trapezoid with amplitude + duration as inputs. #146
• BF: Minor bug fixes in sequence.py and make_gauss_pulse.py by @wtclarke in BF: Minor bug fixes in sequence.py and make_gauss_pulse.py #147
• BF: Refine argument options for make_block_pulse by @wtclarke in BF: Refine argument options for make_block_pulse #151
• Fix BlockDurationRaster being ignored in read_seq.py by @gabuzi in Fix BlockDurationRaster being ignored in read_seq.py #157
• Update confusing documentation for calc_duration.py by @gabuzi in Update confusing documentation for calc_duration.py #158
• Fixed set_block to give errors when multiple events of the same type are specified by @FrankZijlstra in Fixed set_block to give errors when multiple events of the same type are specified #141
• Various bug fixes, features, and cleanups by @FrankZijlstra in Various bug fixes, features, and cleanups #159
• Fixed read_seq for reading v1.3.1 sequence files by @FrankZijlstra in Fixed read_seq for reading v1.3.1 sequence files #162
• Feature: evaluate_labels function in Sequence by @FrankZijlstra in Feature: evaluate_labels function in Sequence #164
• A few small bugfixes by @FrankZijlstra in A few small bugfixes #165
• Bug fixes in make_extended_trapezoid and add_gradients by @FrankZijlstra in Bug fixes in make_extended_trapezoid and add_gradients #167
• Added gradient spectrum calculations by @FrankZijlstra in Added gradient spectrum calculations #170
• Fix make_extended_trapezoid area by @schuenke in Fix make_extended_trapezoid area #171
• Added functionality to remove duplicate events by @FrankZijlstra in Added functionality to remove duplicate events #178
• Add continous deployment (CD) workflow by @schuenke in Add continous deployment (CD) workflow #173
• remove_duplicates() without in_place = True causes signature_value to be not updated after write. by @btasdelen in remove_duplicates() without in_place = True causes signature_value to be not updated after write. #179
• feat: add TRID label (GE interpreter). by @mcencini in feat: add TRID label (GE interpreter). #180
• Changed get_gradients to actually use extrapolation in the piecewise polynomial by @FrankZijlstra in Changed get_gradients to actually use extrapolation in the piecewise polynomial #181

New Contributors

• @tblazey made their first contribution in bug fix to event_lib.py #111
• @yupbank made their first contribution in latest make_sinc_pulse dont return gradient by default #113
• @FrankZijlstra made their first contribution in Bug fixes that produced incorrect test reports #125
• @fzimmermann89 made their first contribution in Add __init__,py to utils to help setuptools discover the package #137
• @wtclarke made their first contribution in Enh: New options in make_arbitrary_rf.py for adiabatic pulses #143
• @Andrew-Dupuis made their first contribution in Bump sigpy version to 0.1.26 #148
• @gabuzi made their first contribution in Fix BlockDurationRaster being ignored in read_seq.py #157
• @mcencini made their first contribution in feat: add TRID label (GE interpreter). #180

Full Changelog: 1.4.0...1.4.1

[schuenke]

I would vote to remove the citations section from the Readme. Currently only citations until 2022 are included, but the section is already quite prominent. With all citations from 2022, 2023, 2024 this will get worse and keeping the list up to date is unnecessary work IMO.

[btasdelen]

@schuenke I agree that citations can be removed. I would argue that contributors can also be removed, as they are already shown by GitHub automatically.

[FrankZijlstra]

The auto-generated changelog looks fine to me.
I agree on removing citations, it would be hard to keep up-to-date anyway.
Removing contributors also seems fine.

One thing to consider about the version number and how we manage it going forward: the package version number (major, minor, revision) is written to the generated .seq files. I'm not 100% certain if any interpreter currently uses this to change any behaviour, but it could in theory have an effect. So if we start deviating from the Matlab Pulseq version numbers, we may need to keep a separate version number for the .seq file format.

[FrankZijlstra]

Apart from merging the last commits to the new release 1.4.1, are there any objections to going forward with this? Things left to do, etc?

[schuenke]

Apart from merging the last commits to the new release 1.4.1, are there any objections to going forward with this? Things left to do, etc?

I don't think so. As suggested, I removed the contributors list from the README file. Further, I replaced the references in the beginning of the file with a link to the references section and added Philips to the list of supported hardware.

Let's get your latest bug fix merged and then we should be good to go.

UPDATE: while writing this, I realized that some tests seem to fail... lets check this...

[FrankZijlstra]

Failing tests were fixed by #180

[schuenke]

I think we are now finally good to go now. Do we wanna stick with 1.4.1 now or directly jump to 1.4.2 to mirror the current Pulseq version? I think this is the last thing to decide 😁

[aTrotier]

Great work!
This PR fixes a lot of incompatibility issues with sigpy and numpy 2.0 :)