Feature t0 vertex #133
Feature t0 vertex #133
Conversation
This allows us to use the globally unique event IDs. Didn't notice this before because we were either using the spill sim (where this change was already made) or using zero-based event IDs in the non-spill (bomb) case. Should probably be backported to `main`.
Feature t0 vertex
| #print("Class: ", event.ClassName()) | ||
| #print("Event number:", event.EventId) | ||
|
|
||
| globalVertexID = (event.RunId * 1E6) + event.EventId |
| vertex["y_vert"] = primaryVertex.GetPosition().Y() | ||
| vertex["z_vert"] = primaryVertex.GetPosition().Z() | ||
| vertices_list.append(vertex) | ||
| vertices[iVtx]["eventID"] = event.EventId |
|
|
||
| # Output array datatypes | ||
| segments_dtype = np.dtype([("vertexID", "u4"), ("segment_id", "u4"), ("z_end", "f4"), | ||
| ("trackID", "u4"), ("tran_diff", "f4"), | ||
| segments_dtype = np.dtype([("eventID","u4"),("vertexID", "u8"), ("segment_id", "u4"), |
There was a problem hiding this comment.
Change eventID to a u8 so that we can store globalEventId?
| @@ -26,27 +27,19 @@ | |||
| ("y", "f4"), ("x", "f4"), ("z", "f4"), | |||
| ("n_photons","f4")], align=True) | |||
|
|
|||
| trajectories_dtype = np.dtype([("vertexID", "u4"), ("trackID", "u4"), | |||
| ("parentID", "i4"), | |||
| trajectories_dtype = np.dtype([("eventID","u4"), ("vertexID", "u8"), | |||
| ("Elep", "f4"), ("lep_mom", "f4"), ("lep_ang", "f4"), ("lep_pdg", "i4"), | ||
| ("q0", "f4"), ("q3", "f4"), ("Q2", "f4"), | ||
| ("x", "f4"), ("y", "f4")], align=True) | ||
| vertices_dtype = np.dtype([("eventID","u4"), ("vertexID","u8"), |
| vertex["z_vert"] = primaryVertex.GetPosition().Z() | ||
| vertices_list.append(vertex) | ||
| vertices[iVtx]["eventID"] = event.EventId | ||
| vertices[iVtx]["vertexID"] = iVtx |
There was a problem hiding this comment.
This is non-unique, which is fine if we're using eventID as the EVENT_SEPARATOR. But if for some reason (memory constraints??) we ever want to use vertexID as the EVENT_SEPARATOR, we'll need it to be unique. Change to something hacky like globalEventId*100 + iVtx?
| trajectories[iTraj]["vertexID"] = event.EventId | ||
| trajectories[iTraj]["trackID"] = trajectory.GetTrackId() | ||
| trajectories[iTraj]["parentID"] = trajectory.GetParentId() | ||
| trajectories[iTraj]["eventID"] = event.EventId |
| trajectories[iTraj]["trackID"] = trajectory.GetTrackId() | ||
| trajectories[iTraj]["parentID"] = trajectory.GetParentId() | ||
| trajectories[iTraj]["eventID"] = event.EventId | ||
| trajectories[iTraj]["vertexID"] = globalVertexID |
There was a problem hiding this comment.
Eventually change to an actual unique vertex ID? Seems to require some changes in edep-sim so that it stores the vertex associated with each trajectory. For now I guess trajectories['eventID'] and trajectories['vertexID'] will have to refer to the same thing (for the non-spill case)
| segment = np.empty(len(hitSegments), dtype=segments_dtype) | ||
| for iHit, hitSegment in enumerate(hitSegments): | ||
| segment[iHit]["vertexID"] = event.EventId | ||
| segment[iHit]["eventID"] = event.EventId | ||
| segment[iHit]["vertexID"] = globalVertexID |
There was a problem hiding this comment.
See comments above re trajectories
|
Successfully tested on a 2x2 (beam) file. Review comments are for future consideration. |
No description provided.