Skip to content

tauRecTools: fix FPE for taus without clusters

Bertrand Martin Dit Latour requested to merge martindl/athena:R22_tauRecToolsFPE into master

Hello,

This MR fixes an issue with taus having at least one cluster within 0.2 of the seed jet axis corrected w.r.t. PV0, but with no cluster within 0.2 of the seed jet axis corrected w.r.t the tau vertex (TV). Before, we would reject taus without clusters within 0.2 w.r.t. PV0, while we ultimately care about clusters within 0.2 w.r.t. TV. In case the tau had no cluster within 0.2 w.r.t. TV, ptIntermediateAxis was null, which was causing PFEs as we take log10(ptIntermediateAxis) in the NN decay mode classifier. With this MR, we instead discard taus that have no cluster within 0.2 w.r.t. TV. The treatment is the trigger is unchanged, as there is no vertex correction.

The reconstruction output is unchanged for "non-pathological taus". And for the edge cases where we had FPEs, these no longer show up.

Cheers, Bertrand

Merge request reports