Codestin Search App

KrisThielemans · 2026-08-10T12:14:34Z

Add support for blocks on cylindrical scanners for reading GATE ROOT list-mode files.

Currently no tests for it yet. Help appreciated. @NikEfth @VietAnhDao

try to future-proof changes in the hierarchy

add "blocks" keywords to .hroot

codacy-production · 2026-08-10T12:15:23Z

Up to standards ✅

🟢 Issues 0 issues

Results:
0 new issues

View in Codacy

🟢 Metrics 4 complexity · 14 duplication

Metric Results

Complexity 4

Duplication 14

View in Codacy

_{NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer}
_{TIP This summary will be updated as you push new changes.}

- add RunTests::check_if_equal for LORInAxialAndNoArcCorrSinogramCoordinates (and DetectoinPositionPair) - add text output to stream for LORInAxialAndNoArcCorrSinogramCoordinates - use the latter in Python

- add basic checks on crystal/blocks spacing when constructing the geometry (call error() when they fail). - add check on ring-distance (call warning() when it fails) See UCL#1753 - add block-info for GE Signa PET/MR The checks will be performed when calling Scanner::set_up(). WARNING: set_up() now needs to be called explicitly. (We used to call it in the constructor, but can no longer do that as it might then throw in unexpected cases).

no longer set tof_mashing=0 and remove some checks that were throwing errors if TOF data was detected. adapt test_time_of_flight.cxx to test the Blocks case. This needed coping with LOR end-points not necessarily corresponding to the detectors.

If the scanner was recognised, but the axial_distance_between_crystals_in_cm was not explicitly given, it was initialised from the transaxial distance. Fixes UCL#1752

KrisThielemans · 2026-09-08T06:08:47Z

All ctests work on my machine. However, the headers of the test files on Zenodo need updating, so CI will currently still fail.

We could fix the headers during CI, but I suppose it's better to upload a new version on Zenodo anyway.

IMPORTANT: to get all the tests to work, I had to enable TOF for blocks-on-cylindrical (and fix a few bugs elsewhere).

NikEfth · 2026-09-08T07:42:17Z

I can work on this after Thursday

KrisThielemans · 2026-09-08T16:25:16Z

CI failures:

zenodo download failed (but CI would have failed anyway)
test_OSMAPOSL_parallelproj fails. I've seen that on my machine as well, although I don't understand why. (It seems to be reconstructing arc-corrected data).

KrisThielemans · 2026-09-09T18:08:55Z

  else
    set_scanner_geometry(scanner_geometry_v);
-
-  set_up();


Probably best to re-instate this, but I had some trouble during testing.

KrisThielemans · 2026-09-09T18:31:47Z

test_OSMAPOSL_parallelproj fails. I've seen that on my machine as well, although I don't understand why. (It seems to be reconstructing arc-corrected data).

reconstructed image with parallelproj seems to use very different dimensions. Still no idea why.

KrisThielemans · 2026-09-09T19:17:08Z

I had to call Scanner:set_up() in more cases. Tests should work now (aside from the Zenodo problem)

valgrind detected that ProjectorByBinPairUsingParallelproj::_already_set_up was used before initialisatoin in set_defaults(). It shouldn't matter as it was afterwards set to false anyway. Nevertheless, best to fix it.

KrisThielemans · 2026-09-09T21:44:14Z

test_OSMAPOSL_parallelproj still fails on GHA, but in a different way (reconstructed images are apparently zero), and not on my laptop. grrr

future-proof and simplify code in a few places

Also run on non-arccorrected data, as that's what we do these days.

* Ask scanner if has been set_up before constructing ProjDataInfo * Make sure set_up() is called in every test.

KrisThielemans · 2026-09-12T20:14:07Z

I was a bit surprised that test_consistency_with_GATE is working fine with the old (i.e. cylindrical) .hroot files, but of course it should, as otherwise we would have broken something.

I still think it's better to test with "blocks" .hroot. (I suppose we could test with both, but I'm not sure that's really a good idea).

KrisThielemans · 2026-09-13T07:54:12Z

Appveyor build for VS 2019 fails while compiling test_IO_ITKMulticomponent.cxx with

https://ci.appveyor.com/project/KrisThielemans/stir/builds/54710682/job/jucuo2u66rcx3u2d#L2838

C:\projects\stir\src\include\stir/LORCoordinates.inl(297): warning C4244: '=': conversion from 'double' to 'coordT', possible loss of data
        with
        [
            coordT=size_t
        ]
C:\projects\stir\src\include\stir/modulo.h(140): error C2668: 'stir::modulo': ambiguous call to overloaded function
C:\projects\stir\src\include\stir/modulo.h(87): note: could be 'int stir::modulo(const int,const int)'
C:\projects\stir\src\include\stir/modulo.h(70): note: or       'float stir::modulo(const float,const float)'
C:\projects\stir\src\include\stir/modulo.h(52): note: or       'double stir::modulo(const double,const double)'
C:\projects\stir\src\include\stir/modulo.h(140): note: while trying to match the argument list '(const FloatOrDouble, FloatOrDouble)'
        with
        [
            FloatOrDouble=size_t
        ]
C:\projects\stir\src\include\stir/LORCoordinates.inl(134): note: see reference to function template instantiation 'FloatOrDouble stir::to_0_2pi<coordT>(const FloatOrDouble)' being compiled
        with
        [
            FloatOrDouble=size_t,
            coordT=size_t
        ]
C:\projects\stir\src\include\stir/LORCoordinates.inl(130): note: while compiling class template member function 'stir::LORInCylinderCoordinates<coordT>::LORInCylinderCoordinates(const stir::LORInAxialAndNoArcCorrSinogramCoordinates<coordT> &)'
        with
        [
            coordT=size_t
        ]
C:\projects\stir\src\include\stir/LORCoordinates.inl(492): note: see reference to function template instantiation 'stir::LORInCylinderCoordinates<coordT>::LORInCylinderCoordinates(const stir::LORInAxialAndNoArcCorrSinogramCoordinates<coordT> &)' being compiled
        with
        [
            coordT=size_t
        ]
C:\projects\stir\src\include\stir/LORCoordinates.inl(367): warning C4244: 'argument': conversion from 'float' to 'coordT', possible loss of data
        with
        [
            coordT=size_t
        ]
C:\projects\stir\src\include\stir/LORCoordinates.inl(407): note: see reference to function template instantiation 'stir::Succeeded stir::find_LOR_intersections_with_cylinder<coordT,coordT>(stir::LORInCylinderCoordinates<coordT> &,const stir::LORAs2Points<coordT> &,const double)' being compiled
        with
        [
            coordT=size_t
        ]
C:\projects\stir\src\include\stir/LORCoordinates.inl(406): note: while compiling class template member function 'stir::Succeeded stir::LORAs2Points<coordT>::change_representation(stir::LORInCylinderCoordinates<coordT> &,const double) const'
        with
        [
            coordT=size_t
        ]
C:\projects\stir\src\include\stir/LORCoordinates.inl(492): note: see reference to class template instantiation 'stir::LORAs2Points<coordT>' being compiled
        with
        [
            coordT=size_t
        ]
C:\projects\stir\src\include\stir/LORCoordinates.inl(369): warning C4146: unary minus operator applied to unsigned type, result still unsigned
C:\projects\stir\src\include\stir/LORCoordinates.inl(370): warning C4146: unary minus operator applied to unsigned type, result still unsigned

All other compilers including VS 2022 are fine.

It's trying to instantiate LORInCylinderCoordinates<size_t> which should indeed fail. (Other warnings are related to LORInAxialAndNoArcCorrSinogramCoordinates<size_t>` ).

It looks like this is caused by

STIR/src/test/IO/test_IO_ITKMulticomponent.cxx

Line 62 in 3da6880

check_if_equal<size_t>(voxels_coords->size(), 64);

and the RunTests addition in this PR

template <class T>
bool
RunTests::check_if_equal(const LORInAxialAndNoArcCorrSinogramCoordinates<T>& a,
                         const LORInAxialAndNoArcCorrSinogramCoordinates<T>& b,
                         const std::string& str)

Looks like a compiler bug to me.

KrisThielemans · 2026-09-13T13:01:53Z

Aside from problems with downloading files from Zenodo, this is all fine now.

I'll revert the change that removed the call to set_up() from set_params() to avoid backwards compatibility problems, and make sure it is only enabled in 7.0.

@VietAnhDao @emikhay this PR that should help you with GATE simulations for blocks and hence NeuroLF. As long as you edit the .hroot file.

Summary of all changes is in the release_6.5.htm. Commit is 0eedd70.

KrisThielemans added 2 commits August 10, 2026 13:11

add ProjDataInfoPETScannerWithDiscreteDetectors

6994040

try to future-proof changes in the hierarchy

[ROOT] support BlocksOnCylindrical

7bda3cc

add "blocks" keywords to .hroot

KrisThielemans self-assigned this Aug 10, 2026

KrisThielemans added the enhancement label Aug 10, 2026

NikEfth self-assigned this Aug 10, 2026

KrisThielemans added 6 commits September 7, 2026 18:06

Merge remote-tracking branch 'origin/master' into ROOTBlocksSupport

7b4567c

add helpers for LORInAxialAndNoArcCorrSinogramCoordinates

d5e2933

- add RunTests::check_if_equal for LORInAxialAndNoArcCorrSinogramCoordinates (and DetectoinPositionPair) - add text output to stream for LORInAxialAndNoArcCorrSinogramCoordinates - use the latter in Python

Fix bug in Interfile parsing for Blocks info

bfd00e6

If the scanner was recognised, but the axial_distance_between_crystals_in_cm was not explicitly given, it was initialised from the transaxial distance. Fixes UCL#1752

fix ROOT header for test of consistency with GATE

38accbc

KrisThielemans added this to the v6.5 milestone Sep 8, 2026

KrisThielemans linked an issue Sep 8, 2026 that may be closed by this pull request

Interfile parsing of BlocksOnCylindrical uses wrong default for axial crystal spacing #1752

Open

KrisThielemans added this to TOF improvements and fixes Sep 8, 2026

github-project-automation Bot moved this to In progress in TOF improvements and fixes Sep 8, 2026

KrisThielemans requested review from NikEfth and danieldeidda September 8, 2026 06:09

KrisThielemans commented Sep 9, 2026

View reviewed changes

KrisThielemans added 2 commits September 9, 2026 13:59

call Scanner::set_up()

f43c906

fix uninitialised _already_set_up warning

134646f

valgrind detected that ProjectorByBinPairUsingParallelproj::_already_set_up was used before initialisatoin in set_defaults(). It shouldn't matter as it was afterwards set to false anyway. Nevertheless, best to fix it.

KrisThielemans force-pushed the ROOTBlocksSupport branch from 8d9852f to 134646f Compare September 9, 2026 21:17

KrisThielemans added 2 commits September 12, 2026 09:10

Use ProjDataInfoPETScannerWithDiscreteDetectors in more places

f2e8f21

future-proof and simplify code in a few places

added release notes and updated .hroot example file

0eedd70

KrisThielemans and others added 2 commits September 12, 2026 09:10

Scanner::set_up() in ReconstructionTests

3f5c43b

Also run on non-arccorrected data, as that's what we do these days.

more Scanner::set_up()

1334f4f

* Ask scanner if has been set_up before constructing ProjDataInfo * Make sure set_up() is called in every test.

more Scanner::set_up() calls

aaeec0b

KrisThielemans force-pushed the ROOTBlocksSupport branch from c0ae735 to 5bb772b Compare September 13, 2026 07:57

This was referenced Sep 13, 2026

ProjDataInfoGeneric et al need TOF changes #1307

Open

Enable TOF support for blocks geometry in SAFIR_genericScanner #1506

Open

Core dump while trying to access Scanner object created for an unknown scanner #1520

Open

KrisThielemans linked an issue Sep 13, 2026 that may be closed by this pull request

ProjDataInfoGeneric et al need TOF changes #1307

Open

This was referenced Sep 13, 2026

Make ProjDataInfoCylindricalNoArcCorr, ProjDataInfoGeneric/Blocks etc hierarchy sensible #1757

Open

Improvements for blocks/generic PET support #993

Open

rewrite code to avoid VS2019 compiler bug

52f74f9

KrisThielemans force-pushed the ROOTBlocksSupport branch from 5bb772b to 52f74f9 Compare September 13, 2026 08:30

KrisThielemans linked an issue Sep 13, 2026 that may be closed by this pull request

Core dump while trying to access Scanner object created for an unknown scanner #1520

Open

KrisThielemans changed the title ~~ROOT blocks-on-cylindrical support~~ Better generic/blocks-on-cylindrical support: ROOT support and enabled TOF Sep 13, 2026

re-enable call to Scanner::set_up() in set_params() until v7

777fe2d

Navigation Menu

Uh oh!

Better generic/blocks-on-cylindrical support: ROOT support and enabled TOF - #1746

Better generic/blocks-on-cylindrical support: ROOT support and enabled TOF#1746
KrisThielemans wants to merge 17 commits into
UCL:masterfrom
KrisThielemans:ROOTBlocksSupport

KrisThielemans commented Aug 10, 2026

Uh oh!

codacy-production Bot commented Aug 10, 2026 •

edited

Loading

Uh oh!

KrisThielemans commented Sep 8, 2026 •

edited

Loading

Uh oh!

NikEfth commented Sep 8, 2026

Uh oh!

KrisThielemans commented Sep 8, 2026

Uh oh!

KrisThielemans Sep 9, 2026

Uh oh!

KrisThielemans commented Sep 9, 2026 •

edited

Loading

Uh oh!

KrisThielemans commented Sep 9, 2026

Uh oh!

KrisThielemans commented Sep 9, 2026

Uh oh!

KrisThielemans commented Sep 12, 2026

Uh oh!

KrisThielemans commented Sep 13, 2026

Uh oh!

KrisThielemans commented Sep 13, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

KrisThielemans commented Aug 10, 2026

Uh oh!

codacy-production Bot commented Aug 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Up to standards ✅

Uh oh!

KrisThielemans commented Sep 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

NikEfth commented Sep 8, 2026

Uh oh!

KrisThielemans commented Sep 8, 2026

Uh oh!

KrisThielemans Sep 9, 2026

Choose a reason for hiding this comment

Uh oh!

KrisThielemans commented Sep 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

KrisThielemans commented Sep 9, 2026

Uh oh!

KrisThielemans commented Sep 9, 2026

Uh oh!

KrisThielemans commented Sep 12, 2026

Uh oh!

KrisThielemans commented Sep 13, 2026

Uh oh!

KrisThielemans commented Sep 13, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

codacy-production Bot commented Aug 10, 2026 •

edited

Loading

KrisThielemans commented Sep 8, 2026 •

edited

Loading

KrisThielemans commented Sep 9, 2026 •

edited

Loading