[PWGCF] Femto Sigma0 implementation: builder, table producer, ppSigma0 task by yakparo · Pull Request #18723 · AliceO2Group/O2Physics

yakparo · 2026-11-13T06:16:53Z

No description provided.

Please consider the following formatting changes to AliceO2Group#18677

github-actions · 2026-11-13T06:17:10Z

O2 linter results: ❌ 0 errors, ⚠️ 2 warnings, 🔕 0 disabled

alibuild · 2026-11-13T06:30:33Z

Error while checking build/O2Physics/staging for be49428 at 2026-11-13 07:30:

No log files found

Full log here.

alibuild · 2026-11-13T06:50:04Z

Error while checking build/O2Physics/o2 for be49428 at 2026-11-13 07:50:

## sw/BUILD/O2-latest/log
2026-11-13T07:49:35.209+0100 [61325:140554152803968] [executioncontext.cpp:739] [ERROR] Failed to download output blob e5d511c38df4775d4f73fdda31ac5aa87d7b178a825fada5e12f7c1742efbce2: 13: Blob absent from sink after replication: Object not found
ninja: build stopped: subcommand failed.

Full log here.

alibuild · 2026-11-13T06:51:26Z

Error while checking build/O2Physics/code-check for be49428 at 2026-11-13 07:51:

## sw/BUILD/O2Physics-code-check-latest/log
--
========== List of issues found ==========
++ echo 'To reproduce the results below locally, run: '\''aliBuild build O2Physics-code-check -e ALIBUILD_BASE_HASH=upstream/master'\''.'
To reproduce the results below locally, run: 'aliBuild build O2Physics-code-check -e ALIBUILD_BASE_HASH=upstream/master'.
++ echo 'To apply all automatic fixes, add the '\''-e O2PHYSICS_CHECKER_FIX=1'\'' option.'
To apply all automatic fixes, add the '-e O2PHYSICS_CHECKER_FIX=1' option.
++ echo 'See the documentation for more details at: <https://aliceo2group.github.io/analysis-framework/docs/tools/#running-with-o2physics-code-check>.'
See the documentation for more details at: <https://aliceo2group.github.io/analysis-framework/docs/tools/#running-with-o2physics-code-check>.
++ echo 'Found 3 errors and 8 warnings.'
Found 3 errors and 8 warnings.
++ [[ 3 -gt 0 ]]
++ cat /sw/BUILD/83febe54fa5b052ceb3375f772bf76bcd29c5d71/O2Physics-code-check/errors.txt
PWGCF/Femto/Core/sigma0Builder.h:267:84: error: no header providing "o2::framework::HistType" is directly included [misc-include-cleaner,-warnings-as-errors]
PWGCF/Femto/Core/sigma0HistManager.h:486:20: error: a template argument list is expected after a name prefixed by the template keyword [clang-diagnostic-missing-template-arg-list-after-template-kw]
PWGCF/Femto/TableProducer/femtoSigma0Producer.cxx:43:44: error: no header providing "o2::framework::OutputObjHandlingPolicy" is directly included [misc-include-cleaner,-warnings-as-errors]
++ [[ 8 -gt 0 ]]
++ cat /sw/BUILD/83febe54fa5b052ceb3375f772bf76bcd29c5d71/O2Physics-code-check/warnings.txt
PWGCF/Femto/Core/sigma0Builder.h:180:3: warning: function 'checkFilters' should be marked [[nodiscard]] [modernize-use-nodiscard]
PWGCF/Femto/Core/sigma0Builder.h:264:38: warning: do not implicitly decay an array into a pointer; consider using gsl::array_view or an explicit cast instead [cppcoreguidelines-pro-bounds-array-to-pointer-decay]
PWGCF/Femto/Core/tripletHistManager.h:697:26: warning: statement should be inside braces [readability-braces-around-statements]
PWGCF/Femto/Core/tripletHistManager.h:699:26: warning: statement should be inside braces [readability-braces-around-statements]
PWGCF/Femto/Core/tripletHistManager.h:701:26: warning: statement should be inside braces [readability-braces-around-statements]
PWGCF/Femto/Core/tripletHistManager.h:816:26: warning: statement should be inside braces [readability-braces-around-statements]
PWGCF/Femto/Core/tripletHistManager.h:818:26: warning: statement should be inside braces [readability-braces-around-statements]
PWGCF/Femto/Core/tripletHistManager.h:820:26: warning: statement should be inside braces [readability-braces-around-statements]
++ [[ 3 -gt 0 ]]
++ exit 1
--

Full log here.

ariedel-cern

Hi @yakparo ,
many thanks for preparing this PR. There are some issues with the data model of the sigma0 and the builder.
Here are the 2 main points:

It does not provide useful information to store the index of the daughter lambda and its decay products for the sigma0. I would suggest doing it the way we do it for cascades, where we only store the indices of the daughter tracks. That is all the information necessary for pair cleaning/CPR/CTR.
Please check how TwoTrackResonance/TwoPhotonResonance builder are implemented. Apply selections on the daughters before they even enter the builder. Applying filters and bitmask by hand is way slower than using the Filter and Partitioning features of the O2 framework

ariedel-cern · 2026-11-13T13:21:04Z

+DECLARE_SOA_INDEX_COLUMN_FULL(Photon, photon, int32_t, FPhotons, "_photon");    ///< row in FPhotons
+// id columns for Lambda daughter tracks
+DECLARE_SOA_INDEX_COLUMN_FULL(PosDau, posDau, int32_t, FTracks, "_PosDau"); //! index column for positive daughter track
+DECLARE_SOA_INDEX_COLUMN_FULL(NegDau, negDau, int32_t, FTracks, "_NegDau"); //! index column for negative daughter track


I am not sure all these indices are necessary. For example, for Cascade we only store the indices to the the daughter tracks (the secondary lambda we dont care about much and the when we we do pair cleaning or CPR, we always check on the track level). So I would remove here the index of the lambda

ariedel-cern · 2026-11-13T13:21:47Z

          PUBLIC_LINK_LIBRARIES O2::Framework O2Physics::AnalysisCore
          COMPONENT_NAME Analysis)
+
+o2physics_add_dpl_workflow(femto-sigma0-producer


Please rename the task to femto-producer-sigma0 to use the same convention as for the other producer tasks

ariedel-cern · 2026-11-13T13:22:32Z

+  sigma0builder::ConfSigma0Tables tables;
+  sigma0builder::Sigma0Builder sigma0Builder;
+
+  HistogramRegistry registry{"Sigma0", {}, OutputObjHandlingPolicy::AnalysisObject};


Use a more descriptive name like FemtoProducerSigma0

ariedel-cern · 2026-11-13T13:25:11Z

+  // {
+  //   sigma0Builder.fillSigma0s(products, lambdas, photons, tracks);
+  // }
+  void process(Lambdas const& lambdas, Photons const& photons, o2::aod::FTracks const& tracks, o2::aod::FCols const& /**/, o2::aod::FColMasks const& /**/, o2::aod::FTrackMasks const& /**/)


This does not seem correct. Like this you would pare every lambda with every photon in every collision.

You have to subscribe to the collision iterate first to group the other tables by the collision index. then you can pare all particle in one collision to search for sigma0 candidates.

ariedel-cern · 2026-11-13T13:26:11Z

+#include "PWGCF/Femto/Core/pairCleaner.h"
+#include "PWGCF/Femto/Core/particleCleaner.h"
+#include "PWGCF/Femto/Core/partitions.h"
+// #include "PWGCF/Femto/Core/photonHistManager.h"


Please done leave code commented

ariedel-cern · 2026-11-13T13:44:12Z

+using FSigma0s = FSigma0s_001;
+
+DECLARE_SOA_TABLE_STAGED_VERSIONED(FSigma0Masks_001, "FSIGMA0MASK", 1, //! sigma0 masks
+                                   femtosigma0s::Mask);


Here you should store the bitmask of the daugher lambda and photon

ariedel-cern · 2026-11-13T13:47:37Z

+    // Preselect photons once and group their row indices by collision.
+    std::unordered_map<int64_t, std::vector<int64_t>> photonRowsPerCollision;
+    for (const auto& photon : photons) {
+      if (selectPhoton(photon)) {


Don't do it like that. Specify a selection bit for the lambda and photon, and then partition them before feeding them to the builder. Please check how the twotrackresonancebuilder or the twophotonresonance builder implements the reconstruction of resonances.

ariedel-cern · 2026-11-13T13:48:05Z

+      return;
+    }
+
+    mPhotonMask = filter.photonMask.value;


Filters for the daughters should be applied outside of the builder with a Filter

ariedel-cern · 2026-11-13T13:49:12Z

+  }
+
+  template <typename T>
+  bool selectLambda(T const& lambda)


Remove, This should all happen, filtering and partioning should happen outside of the builder

ariedel-cern · 2026-11-13T13:50:15Z

+  template <typename TProducts, typename TLambda, typename TPhoton, typename TPosDau, typename TNegDau>
+  bool buildSigma0(TProducts& products, TLambda const& lambda, TPhoton const& photon, TPosDau const& posDau, TNegDau const& negDau)
+  {
+    const float lambdaPt = std::abs(lambda.signedPt());


Please use ROOT::Math::PtEtaPhiMVector and then cache pt/eta/phi as class variables. Please check how twotrackresonance producer is implemented

yakparo and others added 6 commits November 10, 2026 08:02

Created sigma0 builder and table producer and wrote a ppSigma0 task

4318160

Please consider the following formatting changes

88cef53

Merge pull request #9 from alibuild/alibot-cleanup-18677

d1e353d

Please consider the following formatting changes to AliceO2Group#18677

Merge branch 'AliceO2Group:master' into sigma0builder

fed8ccb

Adjusted for lastest Femto update

94cd061

Formatting

be49428

github-actions Bot added pwgcf datamodel labels Nov 13, 2026

github-actions Bot changed the title ~~Femto Sigma0 implementation: builder, table producer, ppSigma0 task~~ [PWGCF] Femto Sigma0 implementation: builder, table producer, ppSigma0 task Nov 13, 2026

yakparo marked this pull request as ready for review November 13, 2026 06:29

yakparo requested review from alibuild, ariedel-cern, dimihayl, dsarkaralice, lauraser, victor-gonzalez and wrzesaCERN as code owners November 13, 2026 06:29

ariedel-cern requested changes Nov 13, 2026

View reviewed changes

Conversation

yakparo commented Nov 13, 2026

Uh oh!

github-actions Bot commented Nov 13, 2026

Uh oh!

alibuild commented Nov 13, 2026

Uh oh!

alibuild commented Nov 13, 2026

Uh oh!

alibuild commented Nov 13, 2026

Uh oh!

ariedel-cern left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Milestone

Development

Uh oh!

3 participants