refactor: improve full_chain_odd.py example

[andiwand]

• align odd full chain with itk full chain

see #1513

[timadye]

Running with ttbar_pu200 = True, I got:

 PYTHIA Error in BeamRemnants::setKinematics: kinematics construction failed

Do we need to worry about that?

By the way, I noticed that this one has addVertexFitting. Do we want to include that in full_chain_itk.py, or is it still broken for the ITk?

[andiwand]

Running with ttbar_pu200 = True, I got:

 PYTHIA Error in BeamRemnants::setKinematics: kinematics construction failed

Do we need to worry about that?

did you see that in the ITk example too? I wonder why this would only show up in the ODD example

[timadye]

did you see that in the ITk example too? I wonder why this would only show up in the ODD example

No, I didn't see it with the ITk example

[andiwand]

By the way, I noticed that this one has addVertexFitting. Do we want to include that in full_chain_itk.py, or is it still broken for the ITk?

I think it should work with a similar setup like here after the fixes I added for the vertexing. since we do not have ambiguity resolution in Acts the results are quite funky sometimes but it should produce some OK results

[timadye]

I think it should work with a similar setup like here after the fixes I added for the vertexing. since we do not have ambiguity resolution in Acts the results are quite funky sometimes but it should produce some OK results

Could you add that, so we have the example for the workshop? I can do it if you like, but you know more about this?

[codecov]

Codecov Report

Merging #1538 (1fb5e67) into main (bfc3676) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main    #1538   +/-   ##
=======================================
  Coverage   48.47%   48.47%           
=======================================
  Files         381      381           
  Lines       20699    20699           
  Branches     9503     9503           
=======================================
  Hits        10034    10034           
  Misses       4112     4112           
  Partials     6553     6553           

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[andiwand]

@timadye I added the vertexing to the ITk example but didn't try it out yet because I don't know how to run it

I also need to look at the pythia error in the ODD example

[timadye]

@timadye I added the vertexing to the ITk example but didn't try it out yet because I don't know how to run it

I can test it!

[andiwand]

@timadye I added the vertexing to the ITk example but didn't try it out yet because I don't know how to run it

I can test it!

thank you!

Running with ttbar_pu200 = True, I got:

 PYTHIA Error in BeamRemnants::setKinematics: kinematics construction failed

Do we need to worry about that?

I did't see this error in the first few events. when did it happen for you? @timadye

But I see a bunch of these

13:45:16    TrackFinding   ERROR     Step failed with EigenStepperError:3: Step size adjustment exceeds maximum trials
13:45:16    TrackFinding   ERROR     Propapation failed: EigenStepperError:3 Step size adjustment exceeds maximum trials with the initial parameters 72670 : 
 3.85578
 24.0004
 2.70553
0.682366
-1.54399
-11.8624

which is likely related to #1385 I think

[timadye]

I did't see this error in the first few events. when did it happen for you? @timadye

It comes immediately after the Sequencer initialisation, before the first of the EigenStepperError errors:

13:02:22    Sequencer      INFO      Starting event loop with 1 threads
13:02:22    Sequencer      INFO        0 services
13:02:22    Sequencer      INFO        0 context decorators
13:02:22    Sequencer      INFO        1 readers
13:02:22    Sequencer      INFO        10 algorithms
13:02:22    Sequencer      INFO        12 writers
 PYTHIA Error in BeamRemnants::setKinematics: kinematics construction failed
13:02:27    TrackFinding   ERROR     Step failed with EigenStepperError:3: Step size adjustment exceeds maximum trials

I compiled with the latest key4hep installation (2022-08-22). I just tested again with your latest branch. The only changes are to set ttbar_pu200 = True. (I also reduced the number of events and tried with numThreads=1, but that doesn't affect the error.)

[andiwand]

very interesting I cannot reproduce it. it might be due to a different pythia version? I use pythia8307

also just freshly built acts from this branch

I will double check if my acts install was clean

[andiwand]

using this setup https://codimd.web.cern.ch/A108z_6tRiWJaIa5yAabdg# I can reproduce the error and there are also a few warnings

 PYTHIA Warning in MultipartonInteractions::init: maximum increased by factor 4.149
...
 PYTHIA Warning in Pythia::check: not quite matched particle energy/momentum/mass  
 PYTHIA Error in BeamRemnants::setKinematics: kinematics construction failed  
 PYTHIA Warning in SimpleSpaceShower::pT2nextQCD: weight above unity  
 PYTHIA Warning in StringFragmentation::fragmentToJunction: bad convergence junction rest frame  
 PYTHIA Warning in SimpleSpaceShower::pT2nextQCD: small daughter PDF

cc @paulgessinger

[timadye]

I suggested some changes for full_chain_itk.py, but I guess some of these could also apply to full_chain_odd.py.

[timadye]

Since you move outputDir to itk_output subdirectory (a nice convenience), can use also specify it for the Sequencer?

s = acts.examples.Sequencer(events=100, numThreads=-1, outputDir=str(outputDir))

That puts the timing.tsv also in itk_output.

[timadye]

@timadye I added the vertexing to the ITk example but didn't try it out yet because I don't know how to run it

with my suggested changes, it seems to run OK for dimuon events or ttbar_pu200. I don't know what sort of performance is to be expected with CKF output. For ttbar_pu200, I'm getting <nRecoVtx>=44 with <nVtxReconstructable>=148.

[andiwand]

, outputDir=str(outputDir)

good idea! will add it

with my suggested changes, it seems to run OK for dimuon events or ttbar_pu200. I don't know what sort of performance is to be expected with CKF output. For ttbar_pu200, I'm getting <nRecoVtx>=44 with <nVtxReconstructable>=148.

I am quite surprised that you can reconstruct 44 vertices. @paulgessinger ODD performed worse no? or was that just with AMVF?

[timadye]

I don't know if I need to approve this again, but it does no harm I think.

The Pythia error message doesn't seem to stop the job working, so I think we can go ahead with this PR. If required to marge, maybe resolve that thread and when merged, unresolve it to continue the discussion?

[andiwand]

as far as I know you need a re-approve after additional changes are made.

The Pythia error message doesn't seem to stop the job working, so I think we can go ahead with this PR. If required to marge, maybe resolve that thread and when merged, unresolve it to continue the discussion?

sounds good will do 👍

[acts-project-service]

The backport to develop/v19.x failed:

The process '/usr/bin/git' failed with exit code 1

To backport manually, run these commands in your terminal:

# Fetch latest updates from GitHub
git fetch
# Create a new working tree
git worktree add .worktrees/backport-develop/v19.x develop/v19.x
# Navigate to the new working tree
cd .worktrees/backport-develop/v19.x
# Create a new branch
git switch --create backport-1538-to-develop/v19.x
# Cherry-pick the merged commit of this pull request and resolve the conflicts
git cherry-pick -x --mainline 1 e5d19700e5daa9d60922a2402fe61308eefb7639
# Push it to GitHub
git push --set-upstream origin backport-1538-to-develop/v19.x
# Go back to the original working tree
cd ../..
# Delete the working tree
git worktree remove .worktrees/backport-develop/v19.x

Then, create a pull request where the base branch is develop/v19.x and the compare/head branch is backport-1538-to-develop/v19.x.