Fix width logic in SliceProjectionOp

[fzimmermann89]

There was some logic to autodetect width of a profile, i.e. in what range it is nonzero.
But this check was completly broken with a too large step size (only 2 steps over the volume), resulting in wrong effective slice width for width larger than 4 voxels.
See failing test in first commit:

    width = 11
    input_shape = SpatialDimension(100, 100, 100)
    slice_profile = SliceSmoothedRectangular(width, 0)
    operator = SliceProjectionOp(input_shape=input_shape, slice_profile=slice_profile, slice_shift=-0.5)
    slices = torch.ones(1, 1, 100, 100)
    (volume,) = operator.adjoint(slices)
    assert (volume[:, 50, 50] > 0).sum() == width

FAILED tests/operators/test_slice_projection_op.py::test_slice_projection_op_width - assert tensor(4) == 11

[github-actions]

Coverage Report

File	Stmts	Miss	Cover	Missing
src/mrpro
_version.py	6	2	67%	7–8
src/mrpro/algorithms/csm
inati.py	25	1	96%	43
src/mrpro/algorithms/dcf
dcf_voronoi.py	55	4	93%	15, 55–56, 89
src/mrpro/algorithms/optimizers
adam.py	30	6	80%	108, 125–129
cg.py	52	1	98%	139
pdhg.py	81	3	96%	180–181, 187
pgd.py	53	4	92%	108, 153–156
src/mrpro/algorithms/reconstruction
DirectReconstruction.py	30	2	93%	64, 80
Reconstruction.py	56	8	86%	57–59, 123–126, 128, 136
RegularizedIterativeSENSEReconstruction.py	52	2	96%	125, 153
TotalVariationRegularizedReconstruction.py	54	4	93%	106, 112, 129, 132
src/mrpro/data
AcqInfo.py	165	7	96%	49, 56, 134–135, 137, 243, 367
Dataclass.py	311	26	92%	58, 319, 335, 401, 465–467, 480, 575, 595–596, 598, 613–614, 616, 663–664, 669–670, 855–856, 881, 888, 893–894, 896
DcfData.py	33	1	97%	62
EncodingLimits.py	97	3	97%	37, 127, 130
IData.py	178	12	93%	138, 155–156, 217–225, 258, 304, 322
IHeader.py	174	8	95%	49, 98–101, 284, 288, 292, 296
KData.py	230	24	90%	122–123, 138, 145, 156–167, 178, 186, 197, 236, 258–260, 307–308, 380, 546, 548, 620
KHeader.py	176	13	93%	115–121, 148, 196, 203–204, 231–238
KNoise.py	22	1	95%	44
KTrajectory.py	95	3	97%	163, 165, 185
Rotation.py	737	42	94%	104, 202, 339, 437, 481, 499, 586, 588, 597, 634, 636, 699, 776, 781, 784, 799, 816, 821, 897, 1085, 1090, 1093, 1117, 1121, 1145, 1265, 1267, 1275–1276, 1340, 1422, 1626, 1633–1635, 1694, 1790, 1942, 1977, 1981, 2156, 2177
SpatialDimension.py	150	19	87%	34, 103, 146, 158, 278–280, 293–295, 329, 347, 360, 373, 386, 399, 408–409, 437
src/mrpro/data/traj_calculators
KTrajectoryCalculator.py	26	2	92%	84, 95
KTrajectoryCartesian.py	31	4	87%	129–132, 136
KTrajectoryIsmrmrd.py	19	1	95%	57
KTrajectorySpiral2D.py	57	13	77%	63–66, 69, 71, 73, 75, 77, 105, 107, 134–136
src/mrpro/operators
AveragingOp.py	38	2	95%	73, 115
B0InformedFourierOp.py	75	2	97%	151–154
CartesianSamplingOp.py	112	4	96%	152, 191, 266, 387
ConjugateGradientOp.py	89	7	92%	62, 64, 100, 106, 228, 230, 233
ConstraintsOp.py	85	4	95%	78, 80, 250, 255
EndomorphOperator.py	28	2	93%	71, 77
FiniteDifferenceOp.py	29	2	93%	40, 126
FourierOp.py	105	9	91%	85–87, 192–193, 212, 257, 321, 326
Functional.py	70	2	97%	116, 118
GridSamplingOp.py	165	15	91%	72–73, 82–83, 90–91, 94, 96, 98, 282, 290–291, 306, 312–313
LinearOperator.py	202	6	97%	217, 255, 296, 305, 313, 330
LinearOperatorMatrix.py	174	19	89%	99, 136, 169, 178, 183, 192–195, 208–211, 219–220, 225–226, 238, 347, 377, 404
MultiIdentityOp.py	16	2	88%	58, 63
NonUniformFastFourierOp.py	198	10	95%	72, 99, 221, 223, 261, 263, 340, 397, 471, 476
Operator.py	88	3	97%	82, 115, 125
PatchOp.py	49	3	94%	93, 129, 144
ProximableFunctionalSeparableSum.py	44	3	93%	118, 213, 224
SliceProjectionOp.py	178	10	94%	45, 62, 64, 70, 154, 180, 216, 253, 290, 330
WaveletOp.py	119	3	97%	184, 228, 254
ZeroPadOp.py	18	1	94%	30
src/mrpro/operators/functionals
SSIM.py	73	7	90%	60–80, 82, 86, 114, 147
src/mrpro/operators/models
EPG.py	256	19	93%	31–32, 283, 288, 304–306, 326–327, 332, 356, 361, 386, 391, 545, 599, 738, 755, 782
src/mrpro/phantoms
EllipsePhantom.py	43	2	95%	66, 131
b0map.py	26	21	19%	37–63
brainweb.py	297	40	87%	276, 290–294, 349–359, 398, 454–457, 479–480, 485–486, 488–489, 493, 501, 508–509, 550, 621, 624–625, 644, 653–656, 667, 669, 700–701, 715, 723
fastmri.py	106	10	91%	50–51, 59, 65, 162, 169–171, 174–175, 189
m4raw.py	74	4	95%	58–59, 74, 76
mdcnn.py	71	7	90%	58, 62–63, 70, 82, 88, 135
src/mrpro/utils
RandomGenerator.py	155	15	90%	23–24, 36, 38, 188, 212, 428, 446, 528, 799, 829–832, 895, 898
ema.py	42	4	90%	49, 74, 78, 82
filters.py	61	1	98%	46
indexing.py	177	1	99%	321
pad_or_crop.py	39	6	85%	40, 43, 46, 49, 66, 73
reshape.py	153	10	93%	176, 370, 482–484, 505, 508–509, 515, 530
slice_profiles.py	49	6	88%	21, 37, 119–122, 155
split_idx.py	10	2	80%	43, 47
summarize.py	57	6	89%	40–41, 70–73, 81
unit_conversion.py	72	15	79%	34, 44, 51, 53, 62, 69, 71, 78, 80, 89, 100, 121, 123, 144, 146
TOTAL	8113	501	94%

Tests	Skipped	Failures	Errors	Time
3301	0 💤	0 ❌	0 🔥	2m 30s ⏱️

[github-actions]

📚 Documentation

📁 Download as zip
🔍 View online

[schuenke]

Looks reasonable. Maybe we can document (at least in this PR) why you choose 4 * max_shape + 1 ?!

[fzimmermann89]

It think half-voxelwise checking would be enough. That we do double the steps is because I am not 100% sure about it. And it is cheap... If you have a better suggestion, we can change it. Patrick Schuenke ***@***.***> schrieb am Mi., 6. Mai 2026, 07:34:

…

***@***.**** approved this pull request. Looks reasonable. Maybe we can document (at least in this PR) why you choose 4 * max_shape + 1 ?! — Reply to this email directly, view it on GitHub <#990 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABH3DV2ZMMNXPTBVI4ZIRR34ZLFFXAVCNFSM6AAAAACYN23ZXCVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHM2DEMZTGQ2TMOJRG4> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[schuenke]

It's fine for me. Just wanted to have it documented in case we are wondering in 3 years where these numbers come from...