Build tms with updated geometry for PDR

[AsaNehm]

With the PDR for TMS coming up, the design of TMS has been adapted. The number of thin and thick steel layers has been changed, with a third option (double layers) being added as well. In addition the horizontal modules will be studied and need to be implemented as well.
Also the number of number of scintillator bars per module changed from 48 to 32 with 6 instead of 4 modules now

[AsaNehm]

Allows for the generation of two different types of geometries: pure stereo and full hybrid (every third layer an horizontal layer). To switch between them change in line 409 the variable to True for hybrid and False for stereo

[jdkio]

Running small tests of 10 files of 1e15 pot each using ProcessND.py
https://landscape.fnal.gov/lens/view/job/[email protected] - stereo
https://landscape.fnal.gov/lens/view/job/[email protected] - hybrid
If this works, I'd like to go up to 100-1000 files to be really sure, and also plot interaction positions as a way of seeing the geometry.

Ideally, we'd make some pictures of the geometry to verify everything is adjusted the way we want. We've had geometry issues before, like where it was rotated 90 degrees. I've asked about plotting geometry efficiently on slack

It also sounds like we need to verify whether the reconstruction parameters may need adjustments. These test files could be used to validate that

[jdkio]

Here are some jobs that worked. Only ran genie and g4 stages. Any failures were due to issues copying flux files
[email protected] - stereo
[email protected] - hybrid

Running tmsreco over 50 files from the output above. This version has a fix to the bar sanity check and I verfied it runs over at least one stereo output file above. But it seems to fail on most hybrid files, see below. It's tagged kleykamp_2024-11-06_candidate and branch kleykamp_2024-11-06_tar_candidate
[email protected] - stereo
[email protected] - hybrid

The hybrid file errors are very sparse. It just says,

Sliced event 111 into 2 slices
not in TMS?? hit 77(backwards) : 218.005, -1473.48, 11315
not in TMS?? hit 78(backwards) : 218.005, -1490.74, 11250
terminate called without an active exception
Aborted (core dumped)

Note that "not in TMS??" messages may or may not be related. There are additional "not in TMS??" messages that don't crash. The message comes from TMS_Kalman. If there's no simple fix, we should make an issue to track this in dune-tms

The program doesn't crash at the same location if I turn off the kalman filter using the config file. But it still crashes with a seg fault:
ConvertToTMSTree.exe /pnfs/dune/scratch/users/kleykamp/nd_production/2024-11-06_test_geom_v104_hybrid_small_edep_only/edep/FHC/00m/00/neutrino.29_1730927987.edep.root

#6  0x00007f9f8d333eeb in TMS_TrackFinder::TrackMatching3D (this=this
entry=0x20a3b60 <TMS_TrackFinder::GetFinder()::Instance>) at /cvmfs/larsoft.opensciencegrid.org/products/gcc/v9_3_0/Linux64bit+3.10-2.17/i
nclude/c++/9.3.0/bits/std_abs.h:72
#7  0x00007f9f8d3405e3 in TMS_TrackFinder::FindTracks (this=this
entry=0x20a3b60 <TMS_TrackFinder::GetFinder()::Instance>, event=...) at TMS_Reco.cpp:523
#8  0x000000000040b326 in ConvertToTMSTree (filename=..., output_filename=...) at ../src/TMS_Reco.h:181
#9  0x0000000000408a86 in main (argc=<optimized out>, argv=<optimized out>) at ConvertToTMSTree.cpp:268

[ast0815]

Ideally, we'd make some pictures of the geometry to verify everything is adjusted the way we want. We've had geometry issues before, like where it was rotated 90 degrees. I've asked about plotting geometry efficiently on slack

I tried and failed to implement that. The next best solution I could come up with was the 3D browser thing at https://dune.github.io/dunendggd/

It will always have the latest build geometry, so you can inspect the PR build there. Though it only builds the "default geometries" at the moment. So when you did not change those, you have to load your own geometry files from your PC into it. Which is no problem at all, just click on the three dots next to the file selector.

[jdkio]

Thanks! I couldn't find the loading buttons. I had to adjust the center bar until the three dots became visible

[ast0815]

Let me know when this is actually ready for review. So far I assume that this is a WIP. Should maybe be marked as such.

[ast0815]

Looks good overall, but please see my comment about commented-out code.

Also, you have not recorded the changes in the CHANGELOG.md. Please do that.

[ast0815]

Also, I just saw, that there is an inconsistency in the layer order when switching from the thin to the thicker layers:

As you can see, there is goes XUVUX rather than XUVX (or whatever they are called now)

[AsaNehm]

Also, I just saw, that there is an inconsistency in the layer order when switching from the thin to the thicker layers:

As you can see, there is goes XUVUX rather than XUVX (or whatever they are called now)

Thanks. I fixed this

[AsaNehm]

I updated the Changelog. Please let me know, if I need to change this further

[ast0815]

If you don't actually see the alternative geometry ever being used again, one could argue that it would be better just to remove those commented lines. Keeping track of things for posterity is handled by git, and it would be better not to clutter up the actual code. But fine, it looks good to me now. Just the changelog still is not in the format it should be in. You can just accept my suggested change to that one if you are happy with it.

[ast0815]

Thanks for incorporating my requests. Would you like to wait for @muether to review this as well, or should I merge?

[AsaNehm]

I'll let you know after the TMS studies meeting today. Most likely waiting for @muether is not necessary