Add configurable Q2 minimum cuts and spline consistency metadata by Pabce · Pull Request #492 · GENIE-MC/Generator

Pabce · 2026-05-21T16:35:06Z

Summary

Adds a configurable phase-space Q2 minimum cut, initially used for EM/electron interactions, with optional weak-interaction support through XML configuration or --q2-min.

Main Changes

Add KPhaseSpaceCuts as the central service for configurable Q2 cuts.
Add CommonPhaseSpaceCuts.xml configuration XML as the single source of truth for Q2 cuts, and potential future kinematic cuts in other variables.
Add --q2-min to gevgen and gmkspl.
Apply the configured cut consistently across phase-space and affected physics modules.
Record Q2 cut metadata in spline XML output.
Make gevgen reject spline files whose recorded Q2 cut is missing or inconsistent when an effective cut is active.

Validation

Added gtestKPhaseSpace coverage for EM defaults, weak defaults, command-line overrides, probe-aware spline metadata, and XML round-tripping.

IMPLEMENTATION_REPORT.md

Add EM-Q2-min parameter to CommonParam.xml [Kinematics] section, allowing users to configure the minimum Q^2 threshold for EM scattering events without recompiling. Default value is 0.02 GeV^2 (unchanged). CRITICAL FIX: DISKinematicsGenerator was sampling (x,y) and converting to Q^2 without checking if Q^2 >= Q2min, allowing events to leak below the threshold. Added explicit Q^2 check in event generation loop. Changes: - Add EM-Q2-min parameter to 5 config files (main + 4 GEM tunes) - Add KPhaseSpace::GetQ2MinEM() to load config value at runtime - Update KineUtils electromagnetic functions to accept Q2min parameter - Fix 8 physics modules that used hardcoded kMinQ2Limit constant: * DISKinematicsGenerator: Add Q^2 threshold check (critical bug fix) * RESKinematicsGenerator: Use config value in max xsec calculation * MECGenerator, MECUtils, SuSAv2MECPXSec, EmpiricalMECPXSec2015 * QELEventGeneratorSuSA, SuSAv2QELPXSec Users must regenerate cross-section splines after changing EM-Q2-min. 17 files modified. See IMPLEMENTATION_REPORT.md for details. Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

The explicit Q² threshold check in DISKinematicsGenerator.cxx has been removed after testing confirmed that the implicit enforcement mechanism works correctly. The Q²min threshold is properly enforced through the existing kinematic validation chain: ValidKinematics() → IsAllowed() → Q2Lim_W() → InelQ2Lim_W() This approach is cleaner and more consistent with how other kinematic constraints are handled in GENIE. Updated IMPLEMENTATION_REPORT.md to reflect this change: - Removed the "Critical Issue" section about DISKinematicsGenerator - Updated file count from 17 to 16 modified files - Added note about the implicit enforcement mechanism Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

Add the ability to override the EM-Q2-min parameter (minimum Q^2 threshold for electromagnetic scattering events) from the command line, without requiring modification of CommonParam.xml. Changes to both gevgen (gEvGen.cxx) and gmkspl (gMakeSplines.cxx): - Parse new --em-q2-min <value> argument (value in GeV^2) - Override the Registry value after BuildTune() loads XML configs - Add documentation in help text and header comments - Log the override value in job configuration output This allows users to easily test different Q^2 thresholds without editing XML configuration files or regenerating builds. Usage examples: gevgen ... --em-q2-min 0.5 gmkspl ... --em-q2-min 0.5

Pabce · 2026-05-21T16:37:16Z

The "enable pi0 decays by default" was a local change that got into the branch by accident - but it is now reverted.

Pabce · 2026-05-22T08:02:35Z

Example spline generated with gmkspl -p 11 -t 1000180400 -e 8.0 -n 3 --event-generator-list EM -o gxspl_eAr40_Q2min010.xml --tune G18_10a_02_11a --q2-min 0.10.

The relevant Q2 cut information is placed as:

<?xml version="1.0" encoding="ISO-8859-1"?>

<!-- generated by genie::XSecSplineList::SaveSplineList() -->

<genie_xsec_spline_list version="3.00" uselog="1">

  <genie_tune name="G18_10a_02_11a">

    <kinematics>
      <q2 min="0.1" unit="GeV^2" source="command-line --q2-min"/>
    </kinematics>
[...]

gxspl_eAr40_Q2min010.xml

jtenavidal · 2026-05-27T13:14:09Z

+  // Q2 minimum override
+  gOptQ2MinSet = false;
+  if(parser.OptionExists("q2-min")) {
+    LOG("gmkspl", pINFO) << "Reading Q2 minimum cut override";


I would add a more descriptive message like
LOG("gmkspl", pINFO) << "Overiding minimum Q2 phase-space minimum by "<< gOptQ2Min ;

Also agreed, added more descriptive wording + value.

jtenavidal · 2026-05-27T13:16:10Z

              cross-section calculation for nuclear targets.
+           --q2-min
+              Override the minimum Q^2 phase-space cut (in GeV^2). For weak
+              interactions, this explicitly opts the run into the Q2 cut.


As this is a new feature, I would specify where this is by default setup for EM interactions.
Quote the CommonPhaseSpaceCuts.xml file

Agreed. Updated the help text here and in gEvGen.cxx to say that the default EM cut is configured in CommonPhaseSpaceCuts.xml as EM-Q2-min.

jtenavidal · 2026-05-27T15:08:38Z


 ClassImp(KPhaseSpace)

+namespace {


Is this needed?

The namespace here is just making ApplyQ2MinCut local to this .cxx file, like a static helper function. If this is not to GENIE-code I can remove it with no change to functionality.

jtenavidal · 2026-05-27T15:09:33Z

    Q2l = kinematics::DarkQ2Lim_W(Ev,M,ml,W);
  } else {
-     Q2l = is_em ? kinematics::electromagnetic::InelQ2Lim_W(Ev,ml,M,W) : kinematics::InelQ2Lim_W(Ev,M,ml,W);
+     Q2l = is_em ? kinematics::electromagnetic::InelQ2Lim_W(Ev,ml,M,W,0.) : kinematics::InelQ2Lim_W(Ev,M,ml,W);


Is this enforcing 0 for everything but em?
Doesn't it need to still be configurable?

No, this does not enforce 0 (it is actually setting 0 for EM in this cut, not for non-EM). The purpose is to pass a default floor for EM so that the actual EM floor comes only from KPhaseSpaceCuts - and not from genie::utils::kinematics::electromagnetic::kMinQ2Limit . Weak interactions still use controls::kMinQ2Limit unless --q2-min or Weak-Q2-min opts them into the configurable cut. I added a comment to make it clearer.

It is worth discussing if we want to get rid of the legacy Q2 floor-setting from KPhaseSpace.cxx altogether (i.e., also for weak interactions).

jtenavidal · 2026-05-27T15:10:26Z

+  Range1D_t  Q2Lim_W_SPP_iso (void) const;  ///< Q2 limits @ fixed W for resonance single pion production on isoscalar nucleon

  static double GetTMaxDFR();
+  static double GetQ2MinEM();  ///< Deprecated; use KPhaseSpaceCuts instead.


Shouldn't we remove the function?

Yes, function was removed and also corresponding unused KPhaseSpaceCuts::EMQ2MinCut().

jtenavidal · 2026-05-27T15:12:38Z

+  bool   fHasQ2MinOverride;
+  double fQ2MinOverride;
+
+  struct Cleaner {


I think this is also not the standard way to do this in GENIE. To discuss later.

Discussed, this is a common pattern in GENIE singleton classes and so is left as-is.

Pabce · 2026-06-05T14:01:03Z

Example before-after splines for neutrino and electron --q2-min limits, showcasing identity except for new kinematics metadata.

after_em_q2_0p02.xml
after_em_q2_0p25.xml
after_numu_q2_1em4.xml
before_em_q2_0p02.xml
before_em_q2_0p25.xml
before_numu_q2_1em4.xml

Pabce and others added 7 commits January 13, 2026 23:25

Final candidate implementation for command line q2 limit

2f4ed24

Enable pi0 decays by default

472ff9b

blup

dd2fe53

Record Q2 cut metadata in spline XML

83ea8a9

Pabce force-pushed the command-line-q2 branch from 36d703e to a41866d Compare May 21, 2026 16:41

Cleanup of irrelevant local files

28a9ba2

Pabce force-pushed the command-line-q2 branch from a41866d to 28a9ba2 Compare May 21, 2026 16:48

jtenavidal requested review from jtenavidal and nusense May 27, 2026 12:33

jtenavidal reviewed May 27, 2026

View reviewed changes

Clarify Q2 cut config and remove obsolete EM helper

e5a6e4b

jtenavidal mentioned this pull request Jun 10, 2026

Make the EM Q2 threshold (kMinQ2Limit) configurable via CommonParam #502

Open

Uh oh!

Conversation

Pabce commented May 21, 2026

Summary

Main Changes

Validation

Uh oh!

Pabce commented May 21, 2026

Uh oh!

Pabce commented May 22, 2026

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!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Pabce commented Jun 5, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants