GLO VTX: add missing protection against accessing MC labels of background events - #2684
GLO VTX: add missing protection against accessing MC labels of background events#2684ehellbar wants to merge 1 commit into
Conversation
|
like this, we will fill the data vertex plots, but not the MC plots. |
|
For the record, this fixes O2-7132 |
| mBeamSpot->Fill(x, y); | ||
|
|
||
| if (mUseMC && mcLbl[i].isSet()) { // make sure the label was set | ||
| if (mUseMC && mcLbl[i].isSet() && mcLbl[i].getSourceID() == 0) { // make sure the label was set and to use only the underlying event |
There was a problem hiding this comment.
I have some doubts about this: In an O2DPG simulation when using embedding, we will have both SourceID == 0 (the background) and SourceID == 1 (the signal) ... and both might contribute to the vertexing task.
We could however use mcLbl.isNoise() or mcLbl.isValid() or something. ... or directly exclude QED by sourceID != 99.
There was a problem hiding this comment.
we do the same check also twice earlier in the function, when we loop over the labels. So for simulations with embedding, this never worked properly, then, always skipping signal and only filling background?
There was a problem hiding this comment.
or, is background == simulated events and signal == embedded signal in this case? Then we might not want the embedded signal in those histogram?
There was a problem hiding this comment.
This particular task extracts the MCHeader multiplicity as a very crude alias of the occupancy, relevant only for the PbPb collision. Assuming that the underlying PbPb event is always source 0, a few tracks coming from the overlaid signal with source>0 should be irrelevant.
There was a problem hiding this comment.
If it's only used for PbPb then it's probably fine
5 participants
No description provided.