Partial fix to HGCal hexagons position, and SimCluster proxy

[tcuisset]

Hi,
I've made a temporary fix for the HGCal hexagon positions, fixing partially issue #128
Now the center of the hexagons are in the correct position (before it seems they were rotated around the origin). It is still missing the rotation around the center of the hexagon (though that is less critical).

I've also added a proxy for SimCluster (very similar to CaloParticle), and a small interface convenience for trackster layer display mode.

[alja]

@tcuisset

Thank you for the research on mispositioned boxsets!

I looked at the REveBoxSet::AddHex interface and saw that it is not compatible with the old TEveBoxSet::AddHex interface. The REveBoxSet::AddHex expects the rotation angle in radians instead of degrees. I made a simple root macro that tests the orientation of the hexagon here: http://xrd-cache-1.t2.ucsd.edu/alja/mail/boxset_hex.C.

Can you check if you check the digit position is correct if the rotation angle is set angle to TMath::Pi() * 0.5? Here is an example to test a single digit:
root.exe boxset_hex.C

Let me know if this solves the problem.
If not, please send me your data file and instructions for reproducing the problem.

[tcuisset]

I tried changing the angle parameter to pi/2, and it does not solve the problem. The radians-degrees issue might be a separate issue. See the screen capture below where I changed the angle of different collections (the most correct one is angle=0)

The issue is that the rotation angle seems to induce a rotation around the z-axis instead of a rotation around the hexagon center. I've put in a drawing what seems to be the issue:

Here is the link to the input file and geometry file in case you want to test.

[alja]

@tcuisset
Thank Theo!
I will run the test using the same input and geometry files.
Which CMSSW release are you using?

[tcuisset]

I'm using CMSSW_15_1_0_pre4_ROOT636 for fireworks & geometry
(and the samples were made with CMSSW_16_0_0_pre2 though they seem compatible)

[alja]

@tcuisset

The bug was in REveBoxSet, where rotation should be in local instead of the parent frame:
alja/root@3b6275c

Fixing the bug in root may also be an opportunity to switch to degrees instead of radians, as in TEve. What units in rotation should REveBoxset::AddHaxagon have, in your opinion?

I will update the FireworksWeb service to include the ROOT correction next week.
The bugfix will also go to the ROOT main branch.
I believe new CMSSW builds with the root master branch are available every second week.

How do you set up a developer environment for FireworksWeb?

[tcuisset]

Hi @alja
Thanks for finding the fix !
For degrees vs radians I don’t have a particular opinion (as long as it’s documented). I guess you can take whatever convention is most used in ROOT/REve for rotations.

To setup a developper environment I just install a CMSSW with a ROOT636 build. I guess to test the fix I can switch to the IB with ROOT master. Then I just clone the FireworksWeb repo, scram build and run (then SSH port forward).