[PWGJE] pTHat handling, remove HepMCXSections dependency, add R-d… by joonsukbae · Pull Request #15284 · AliceO2Group/O2Physics

joonsukbae · 2026-03-05T22:23:14Z

…ependent matching

trackEfficiency: Remove HepMCXSections/JMcCollisionPIs table subscriptions from all MC processes. Use ptHard() from JMcCollisions with weight-based fallback. Add applyRCTSelections configurable. Remove getPtHatFromHepMCXSection configurable.
jetSpectraCharged: Pass pTHat as parameter to fill functions. Fix pTHatMaxMCD→pTHatMaxMCP bug. Change MCD weighted subscriptions to JetCollisionsMCD. Add applyRCTSelections configurable.
Jet matching (all 6 headers + JetMatchingUtilities.h): Add per-R matching distance override (jetRadiiForMatchingDistance + maxMatchingDistancePerJetR) on top of existing maxMatchingDistance fallback. Backward compatible — empty per-R vectors (default) give identical results.

…ependent matching - trackEfficiency: Remove HepMCXSections and JMcCollisionPIs table subscriptions from all MC processes; use ptHard() from JMcCollisions with weight-based fallback. Add applyRCTSelections configurable. - jetSpectraCharged: Pass pTHat as parameter to fill functions instead of computing internally; fix pTHatMaxMCD->pTHatMaxMCP bug in fillGeoMatchedAreaSubHistograms; change MCD weighted process subscriptions to JetCollisionsMCD for mcCollision() access. - All jet matching tasks: Replace single maxMatchingDistance with per-R configurables (jetRadiiForMatchingDistance + maxMatchingDistancePerJetR) for R-dependent geometric matching distance. Applied to jetMatchingMC, jetMatchingMCSub, jetMatchingSub, jetMatchingDuplicates, jetSubstructureMatching, and jetSubstructureMatchingSub. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

nzardosh

Thanks for the PR Joonsuk. Please make the changes and then locally test that you get the same results.
Also in the future if you would like to make changes to the core tasks, please discuss them in advance first
Thanks

nzardosh · 2026-03-06T09:28:28Z

PWGJE/Core/JetMatchingUtilities.h

    }
-    std::tie(baseToTagMatchingGeoIndex, tagToBaseMatchingGeoIndex) = MatchJetsGeometrically(jetsBasePhi, jetsBaseEta, jetsTagPhi, jetsTagEta, maxMatchingDistance); // change max distnace to a function call
+    float effectiveMatchingDistance = maxMatchingDistance;
+    for (std::size_t i = 0; i < jetRadiiForMatchingDistance.size() && i < maxMatchingDistancePerJetR.size(); i++) {


here i would only check one of i < jetRadiiForMatchingDistance.size() or i < maxMatchingDistancePerJetR.size() since we constrain them to be the same

nzardosh · 2026-03-06T09:32:47Z

PWGJE/Core/JetMatchingUtilities.h

      jetsTagGlobalIndex.emplace_back(jetTag.globalIndex());
    }
-    std::tie(baseToTagMatchingGeoIndex, tagToBaseMatchingGeoIndex) = MatchJetsGeometrically(jetsBasePhi, jetsBaseEta, jetsTagPhi, jetsTagEta, maxMatchingDistance); // change max distnace to a function call
+    float effectiveMatchingDistance = maxMatchingDistance;


here take the first one in the list so we can get rid of maxMatchingDistance as variable. So do
float effectiveMatchingDistance = maxMatchingDistance[0];

it will break if an empty list is given at runtime but i think thats good. or where you do the LOGP fatal you can also check that the lists are not empty

nzardosh · 2026-03-06T09:33:01Z

PWGJE/Core/JetMatchingUtilities.h

 }

 template <typename T, typename U>
-void MatchGeo(T const& jetsBasePerCollision, U const& jetsTagPerCollision, std::vector<std::vector<int>>& baseToTagMatchingGeo, std::vector<std::vector<int>>& tagToBaseMatchingGeo, float maxMatchingDistance)


remove the variable float maxMatchingDistance,

nzardosh · 2026-03-06T09:33:19Z

PWGJE/TableProducer/Matching/Duplicates/jetMatchingDuplicates.h

@@ -35,6 +35,8 @@ struct JetMatchingDuplicates {
  o2::framework::Configurable<bool> doMatchingPt{"doMatchingPt", true, "Enable pt matching"};
  o2::framework::Configurable<bool> doMatchingHf{"doMatchingHf", false, "Enable HF matching"};
  o2::framework::Configurable<float> maxMatchingDistance{"maxMatchingDistance", 0.24f, "Max matching distance"};


remove this variable everywhere

nzardosh · 2026-03-06T09:33:54Z

PWGJE/TableProducer/Matching/Duplicates/jetMatchingDuplicates.h

  o2::framework::Configurable<bool> doMatchingHf{"doMatchingHf", false, "Enable HF matching"};
  o2::framework::Configurable<float> maxMatchingDistance{"maxMatchingDistance", 0.24f, "Max matching distance"};
+  o2::framework::Configurable<std::vector<double>> jetRadiiForMatchingDistance{"jetRadiiForMatchingDistance", {}, "Jet R values for per-R matching distance override, e.g. {0.2, 0.4, 0.6}"};
+  o2::framework::Configurable<std::vector<double>> maxMatchingDistancePerJetR{"maxMatchingDistancePerJetR", {}, "Max matching distance for each R in jetRadiiForMatchingDistance, e.g. {0.12, 0.24, 0.36}"};


have you checked up in previous analysis notes that these are the correct default values? Please also find the appropriate ones for R=0.3 and R=0.5 and add them too

nzardosh · 2026-03-06T09:36:11Z

PWGJE/Core/JetMatchingUtilities.h

    }
-    std::tie(baseToTagMatchingGeoIndex, tagToBaseMatchingGeoIndex) = MatchJetsGeometrically(jetsBasePhi, jetsBaseEta, jetsTagPhi, jetsTagEta, maxMatchingDistance); // change max distnace to a function call
+    float effectiveMatchingDistance = maxMatchingDistance;
+    for (std::size_t i = 0; i < jetRadiiForMatchingDistance.size() && i < maxMatchingDistancePerJetR.size(); i++) {


this whole addition could be moved to line 314, after these :
for (auto jetR : jetsR) {
std::vector jetsBasePhi;
std::vector jetsBaseEta;
std::vector jetsBaseGlobalIndex;
std::vector baseToTagMatchingGeoIndex;
std::vector tagToBaseMatchingGeoIndex;

Because that way then you evaluate it once per jetR and you dont have to do the evaluation each time per jet

nzardosh · 2026-03-06T09:38:36Z

PWGJE/Core/JetMatchingUtilities.h


 template <typename T, typename U>
-void MatchGeo(T const& jetsBasePerCollision, U const& jetsTagPerCollision, std::vector<std::vector<int>>& baseToTagMatchingGeo, std::vector<std::vector<int>>& tagToBaseMatchingGeo, float maxMatchingDistance)
+void MatchGeo(T const& jetsBasePerCollision, U const& jetsTagPerCollision, std::vector<std::vector<int>>& baseToTagMatchingGeo, std::vector<std::vector<int>>& tagToBaseMatchingGeo, float maxMatchingDistance, std::vector<double> const& jetRadiiForMatchingDistance = {}, std::vector<double> const& maxMatchingDistancePerJetR = {})


you dont need to initialise with empty vectors. Its good to make sure these vectors are always given to the function and no default empty value is taken

- Remove maxMatchingDistance from all headers and JetMatchingUtilities.h - Non-empty defaults: {0.3, 0.4, 0.5} -> {0.18, 0.24, 0.30} - init() fatal on empty/mismatched vectors - Runtime fatal if jet R not in configured list - Per-R lookup once per R, not per jet - No default empty vectors in function signatures Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

joonsukbae · 2026-03-07T00:08:58Z

Thanks for the PR Joonsuk. Please make the changes and then locally test that you get the same results. Also in the future if you would like to make changes to the core tasks, please discuss them in advance first Thanks

Dear Nima,
Thanks for the review! I've addressed the comments:

Removed maxMatchingDistance from all 6 matching headers and JetMatchingUtilities.h
Per-R matching distance with non-empty defaults: {0.3, 0.4, 0.5} → {0.18, 0.24, 0.30}
Fail-early strategy: init() fatal on empty/mismatched vectors, runtime fatal if jet R not configured, no default empty
vectors in function signatures
Per-R lookup moved to once per R, not per jet

Tested locally: ran full pipeline (jet finding + matching + trackEfficiency + jetSpectraCharged) on LHC25a2. All 125
histograms identical between pre-PR and post-PR code.

Understood — I'll discuss changes to core tasks in advance before making edits. Thanks for the guidance.

joonsukbae requested review from alibuild, fjonasALICE, maoyx, mfasDa, mhemmer-cern, nzardosh and raymondEhlers as code owners March 5, 2026 22:23

github-actions bot added the pwgje label Mar 5, 2026

joonsukbae force-pushed the fix/pthat-matching-cleanup branch from 3e8d5b5 to 7ebc7cc Compare March 5, 2026 22:30

joonsukbae force-pushed the fix/pthat-matching-cleanup branch from 7ebc7cc to 35b52a7 Compare March 5, 2026 22:34

nzardosh reviewed Mar 6, 2026

View reviewed changes

nzardosh marked this pull request as draft March 6, 2026 09:39

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[PWGJE] pTHat handling, remove HepMCXSections dependency, add R-d…#15284

[PWGJE] pTHat handling, remove HepMCXSections dependency, add R-d…#15284
joonsukbae wants to merge 2 commits intoAliceO2Group:masterfrom
joonsukbae:fix/pthat-matching-cleanup

joonsukbae commented Mar 5, 2026 •

edited

Loading

Uh oh!

nzardosh left a comment

Uh oh!

nzardosh Mar 6, 2026

Uh oh!

nzardosh Mar 6, 2026

Uh oh!

nzardosh Mar 6, 2026

Uh oh!

nzardosh Mar 6, 2026

Uh oh!

nzardosh Mar 6, 2026

Uh oh!

nzardosh Mar 6, 2026

Uh oh!

nzardosh Mar 6, 2026

Uh oh!

nzardosh Mar 6, 2026

Uh oh!

joonsukbae commented Mar 7, 2026

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

2 participants

Conversation

joonsukbae commented Mar 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

nzardosh 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!

joonsukbae commented Mar 7, 2026

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

2 participants

joonsukbae commented Mar 5, 2026 •

edited

Loading