change clusterizer and CASeeder to reject zsize==1 clusters #1024

bogui56 · 2024-12-16T16:57:39Z

Set Clusterizer to split common mode noise patterns into zsize = 1 clusters and add machinery to reject them in PHCAseeding

PR Summary: Reject zsize==1 clusters in TPC seeding and cap seed propagation

Motivation / Context

TPC clustering can produce common-mode noise patterns that split into clusters with zsize==1 (single z-bin width), reducing seed quality. This PR filters such clusters during track seeding reconstruction and introduces a cap on the number of seeds propagated through the Kalman filter.

Key Changes

PHCASeeding rejection: Add reject_zsize1_clusters(true) flag in Tracking_Reco_TpcSeed_run2pp() to filter out single-z-bin clusters during track seeding
Seed count limit: Introduce set_max_seeds(5000) on PHSimpleKFProp to cap the maximum number of seeds propagated during Kalman filter expansion
Modifications limited to common/Trkr_Reco.C (+4/-2 lines)

Potential Risk Areas

Reconstruction behavior changes: Rejecting zsize==1 clusters alters track seed demographics and may affect downstream tracking efficiency
Low pT performance: Discussion comments note a small efficiency loss at low transverse momentum from the zsize==1 rejection; impact is currently considered acceptable
Inconsistent seeding configuration: The reject_zsize1_clusters flag is only explicitly added to Tracking_Reco_TpcSeed_run2pp(); multiple other seeding functions in the file (for cosmics, zero-field, iterative seeding) may not have this applied, potentially creating inconsistency across reconstruction paths
Seed cap implications: The 5000 seed limit may suppress rare high-multiplicity events if seed propagation would naturally exceed this threshold

Possible Future Improvements

Apply the reject_zsize1_clusters flag consistently across all TPC seeding functions in the macro suite
Centralize seeding initialization as suggested in PR discussion to ensure uniform module/flag configuration across reconstruction macros
Monitor low-pT tracking efficiency in QA to quantify performance trade-off
Document typical seed counts to assess whether 5000 is an appropriate cap

Note: AI may make mistakes in reconstructing algorithm details or impact assessment. Cross-reference with actual tracking performance QA and discuss with physics/detector experts before deployment.

sphenix-jenkins-ci · 2024-12-16T16:58:38Z

For repository maintainers, please start the CI check manually (feedback)

This is an automatic message to assist manually starting CI check for this pull request, commit 921e604743be291ba2820bf23d3cbe1cab4efd2b. macros pull request require a manual start for CI checks, in particular selecting which coresoftware and calibrations versions to check against this macros pull request.

sPHENIX software maintainers: please make your input here and start the Build:

Note:

if needed, fill in the pull request ID for the coresoftware pull request, e.g. origin/pr/1697/merge for PR#1697 in sha_coresoftware. Default is to check with the master branch.
click Build button at the end of the long web page to start the test

Automatically generated by sPHENIX Jenkins continuous integration

sphenix-jenkins-ci · 2025-02-13T15:23:40Z

For repository maintainers, please start the CI check manually (feedback)

This is an automatic message to assist manually starting CI check for this pull request, commit 6b09d9cd04f65b5dab93a658be9e9200042978b6. macros pull request require a manual start for CI checks, in particular selecting which coresoftware and calibrations versions to check against this macros pull request.

sPHENIX software maintainers: please make your input here and start the Build:

Note:

if needed, fill in the pull request ID for the coresoftware pull request, e.g. origin/pr/1697/merge for PR#1697 in sha_coresoftware. Default is to check with the master branch.
click Build button at the end of the long web page to start the test

Automatically generated by sPHENIX Jenkins continuous integration

osbornjd · 2025-02-13T17:42:03Z

Should we add this flag to our reco macros too? Better yet, is it time for us to have a centralized function call for our seeding reconstruction?

sPHENIX-GitHub-l · 2025-02-13T17:45:57Z

First part of the quesion: yes! Second part, what do you mean by centralized function call? On 13. Feb 2025, at 18:42, Joseph (Joe) Osborn ***@***.***> wrote: Should we add this flag to our reco macros too? Better yet, is it time for us to have a centralized function call for our seeding reconstruction? — Reply to this email directly, view it on GitHub<#1024 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AEALMH7MT3JS4H3HVHVHTNL2PTKQLAVCNFSM6AAAAABTWSQFV6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDMNJXGMYTSNBRGE>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***> [osbornjd]osbornjd left a comment (sPHENIX-Collaboration/macros#1024)<#1024 (comment)> Should we add this flag to our reco macros too? Better yet, is it time for us to have a centralized function call for our seeding reconstruction? — Reply to this email directly, view it on GitHub<#1024 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AEALMH7MT3JS4H3HVHVHTNL2PTKQLAVCNFSM6AAAAABTWSQFV6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDMNJXGMYTSNBRGE>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

osbornjd · 2025-02-13T17:51:30Z

For example we are not currently calling the function Tracking_Reco_TrackSeed() in our reconstruction macros, e.g. here. The seeding may be at a point where it has converged enough to have it a central place, which would also make it such that we would have a consistent set of modules/flags that the macros by default call

bogui56 · 2025-02-13T21:03:53Z

Ah, that way, well for me that would be ok. Modulo the changed Michael and David will still come up with...

osbornjd · 2025-02-14T16:31:40Z

It does cause a small loss in efficiency at low pT
https://nbviewer.org/github/sPHENIX-Collaboration/QA-gallery/blob/jenkins-sPHENIX-test-tracking-low-occupancy-qa-6871-test-tracking_Event400_Sum28/QA-tracking.ipynb
but that is probably not an issue right now

sphenix-jenkins-ci · 2025-02-19T16:35:34Z

For repository maintainers, please start the CI check manually (feedback)

This is an automatic message to assist manually starting CI check for this pull request, commit aba70f9b802fd3c5a8e69889b684b1a671840af0. macros pull request require a manual start for CI checks, in particular selecting which coresoftware and calibrations versions to check against this macros pull request.

sPHENIX software maintainers: please make your input here and start the Build:

Note:

if needed, fill in the pull request ID for the coresoftware pull request, e.g. origin/pr/1697/merge for PR#1697 in sha_coresoftware. Default is to check with the master branch.
click Build button at the end of the long web page to start the test

Automatically generated by sPHENIX Jenkins continuous integration

coderabbitai · 2026-01-19T13:10:33Z

📝 Walkthrough

Walkthrough

Modifies common/Trkr_Reco.C to add cluster rejection filtering to PHCASeeding seeders and enforce a maximum seed limit on PHSimpleKFProp propagation, adding two new methods for these configurations.

Changes

Cohort / File(s)	Summary
Seeding Configuration `common/Trkr_Reco.C`	Adds `reject_zsize1_clusters(bool)` method to PHCASeeding with four instances enabling z-size-1 cluster rejection; adds `set_max_seeds(int)` method to PHSimpleKFProp with max seed limit of 5000 applied in two propagation call sites

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

sphenix-jenkins-ci · 2026-01-19T13:11:36Z

For repository maintainers, please start the CI check manually (feedback)

This is an automatic message to assist manually starting CI check for this pull request, commit dac2f9f58c568ee68d5422baa19c3c5c2fee714b. macros pull request require a manual start for CI checks, in particular selecting which coresoftware and calibrations versions to check against this macros pull request.

sPHENIX software maintainers: please make your input here and start the Build:

Note:

if needed, fill in the pull request ID for the coresoftware pull request, e.g. origin/pr/1697/merge for PR#1697 in sha_coresoftware. Default is to check with the master branch.
click Build button at the end of the long web page to start the test

Automatically generated by sPHENIX Jenkins continuous integration

change clusterizer and CASeeder to reject zsize==1 clusters

921e604

osbornjd marked this pull request as draft January 21, 2025 17:25

remove changes to TpcClusterizer settings from PR

6b09d9c

bogui56 marked this pull request as ready for review February 13, 2025 15:24

add max seed = 5000 cut

aba70f9

using theirs

dac2f9f

osbornjd merged commit dc8a5a6 into sPHENIX-Collaboration:master Jan 20, 2026
18 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

change clusterizer and CASeeder to reject zsize==1 clusters #1024

change clusterizer and CASeeder to reject zsize==1 clusters #1024

Uh oh!

bogui56 commented Dec 16, 2024 •

edited by coderabbitai bot

Loading

Uh oh!

sphenix-jenkins-ci bot commented Dec 16, 2024

Uh oh!

sphenix-jenkins-ci bot commented Feb 13, 2025

Uh oh!

osbornjd commented Feb 13, 2025

Uh oh!

sPHENIX-GitHub-l commented Feb 13, 2025 via email

Uh oh!

osbornjd commented Feb 13, 2025

Uh oh!

bogui56 commented Feb 13, 2025

Uh oh!

osbornjd commented Feb 14, 2025

Uh oh!

sphenix-jenkins-ci bot commented Feb 19, 2025

Uh oh!

coderabbitai bot commented Jan 19, 2026 •

edited

Loading

Walkthrough

Changes

Uh oh!

sphenix-jenkins-ci bot commented Jan 19, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

change clusterizer and CASeeder to reject zsize==1 clusters #1024

change clusterizer and CASeeder to reject zsize==1 clusters #1024

Uh oh!

Conversation

bogui56 commented Dec 16, 2024 • edited by coderabbitai bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR Summary: Reject zsize==1 clusters in TPC seeding and cap seed propagation

Motivation / Context

Key Changes

Potential Risk Areas

Possible Future Improvements

Uh oh!

sphenix-jenkins-ci bot commented Dec 16, 2024

For repository maintainers, please start the CI check manually (feedback)

Uh oh!

sphenix-jenkins-ci bot commented Feb 13, 2025

For repository maintainers, please start the CI check manually (feedback)

Uh oh!

osbornjd commented Feb 13, 2025

Uh oh!

sPHENIX-GitHub-l commented Feb 13, 2025 via email

Uh oh!

osbornjd commented Feb 13, 2025

Uh oh!

bogui56 commented Feb 13, 2025

Uh oh!

osbornjd commented Feb 14, 2025

Uh oh!

sphenix-jenkins-ci bot commented Feb 19, 2025

For repository maintainers, please start the CI check manually (feedback)

Uh oh!

coderabbitai bot commented Jan 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Walkthrough

Changes

Uh oh!

sphenix-jenkins-ci bot commented Jan 19, 2026

For repository maintainers, please start the CI check manually (feedback)

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

bogui56 commented Dec 16, 2024 •

edited by coderabbitai bot

Loading

coderabbitai bot commented Jan 19, 2026 •

edited

Loading