search

fcs-proj / FastCaloSim

An experiment-independent library for fast calorimeter simulation
Apache License 2.0
2 stars 0 forks source link

Merge barrels of the ATLAS geometry #36

Closed tong-qiu closed 1 month ago

tong-qiu commented 1 month ago

This PR merges barrel layers which are continuous at z=0 of the ATLAS geometry. This is needed to avoid particles getting stuck in the shared boundary between the positive part and the negative part during transportation. The upstream change to pygeosimplify was done in https://github.com/jbeirer/pygeosimplify/pull/164.

Closes https://github.com/jbeirer/FastCaloSim/issues/15

jbeirer commented 1 month ago

Hi @tong-qiu,

the CI is not running, presumably because Github actions are deactivated to run on forks by default.

I now changed the setting, but it seems like changing the setting does not trigger the CI immediately. Would you be so kind and re-open the pull request?

Many thanks!

jbeirer commented 1 month ago

Good to see that simulation output remains unaffected :)

jbeirer commented 1 month ago

Hi @tong-qiu,

I just noticed that the strange extrapolations at z=0 in https://github.com/jbeirer/FastCaloSim/blob/main/test/refs/AtlasTransport/AtlasTransportTests_AtlasTransport/photon/particle_extrapolation.png are still there after the fix.

Do you understand why or could you please investigate what is happening?

Thanks!

tong-qiu commented 1 month ago

Hi @jbeirer,

I have looked into the issue.

I believe the problem is with L214 of FastCaloSimCaloExtrapolation.cxx. When eta=0, if it is not possible to scale z. The unexpected points on the plots are from the endcap layers.

Cheers, Tong Qiu

jbeirer commented 1 month ago

Hi @tong-qiu,

thanks, I will have a closer look at what is happening. It is good to know that this does not come from the geo definition, in which case that might also have an impact on our current simulation.

Cheers,

Joshua