Hello,
Here are the new Jira items on the topic of improving LCIO event output:
"Add ReconstructedParticle output."
https://jira.slac.stanford.edu/browse/HPSJAVA-21
"Make the BilliorVertex into a lcio vertex object"
https://jira.slac.stanford.edu/browse/HPSJAVA-20
"Add ParticleID output."
https://jira.slac.stanford.edu/browse/HPSJAVA-22
"Persist extra track information to LCGenericObjects."
https://jira.slac.stanford.edu/browse/HPSJAVA-23
When each task is completed, please resolve as "Fixed" and select "1.6" under "Fix Version".
They're all assigned to Matt but the actual person who closes out the bug should re-assign to themselves (if they did the work).
Thanks.
________________________________________
From: [log in to unmask] [[log in to unmask]] On Behalf Of McCormick, Jeremy I. [[log in to unmask]]
Sent: Thursday, March 07, 2013 1:18 PM
To: hps-software
Subject: improving the LCIO event output
Hi,
I spoke with Omar a few weeks back about improving the LCIO event output, and I'm glad to see that we will focus on it.
The details of what we sketched out for adding to the event at that time were:
-tracking
-track residuals
-additional track errors
-other track info
-Reconstructed Particles
-Track associations
-Cluster associations
-Vertex
-ParticleID
The first set of information for tracking is mostly supplementary, and I'm not sure what should or should not be carried along in the event.
I think, in general, to make associations between LCIO objects and arbitrary data, it is best to use LCGenericObjects and LCRelations.
This association then looks like this (e.g. for a Track):
Track <=> LCRelation(Track, LCGenericObject) <=> LCGenericObject
Each generic object can contain a variable length list of floats and doubles with whatever extra information we want to persist.
Let me know if more information is needed on creating these collections. There are some examples I can point to.
For the particle output, we basically need to add ReconstructedParticles and all the associations that it has, including to tracks and clusters.
The following LCIO interfaces (that we have maybe not used much before) will be of use here:
http://www.lcsim.org/software/lcsim/2.8-SNAPSHOT/apidocs/org/lcsim/event/ReconstructedParticle.html
http://www.lcsim.org/software/lcsim/2.8-SNAPSHOT/apidocs/org/lcsim/event/Vertex.html
http://www.lcsim.org/software/lcsim/2.8-SNAPSHOT/apidocs/org/lcsim/event/ParticleID.html
Each of these should have a base implementation that can be instantiated and then added to the LCIO events in the normal way.
These are the application base classes I could find:
http://www.lcsim.org/software/lcsim/2.8-SNAPSHOT/apidocs/org/lcsim/event/base/BaseReconstructedParticle.html
http://www.lcsim.org/software/lcsim/2.8-SNAPSHOT/apidocs/org/lcsim/event/base/BaseVertex.html
http://www.lcsim.org/software/lcsim/2.8-SNAPSHOT/apidocs/org/lcsim/event/base/BaseParticleID.html
For these improvements, we should probably make a few issues in the Jira:
1) add ReconstructedParticle output to the LCIO event (this can be a parent issue which is possible to do in Jira)
a) add Vertices
b) add particle ID
2) persist extra track information (more specifics needed here???)
This will be nice for getting the changes into the next set of release notes.
Then, on another note, a whole separate set of issues revolves around:
-stripping unused or redundant collections from the final events
-filtering out uninteresting events from the output stream
-making separate event streams containing certain types of events (2-track, etc.)
These topics can probably be addressed later once the overall integrity and completeness of the events has been improved as outlined above.
I think that should be it. If there are any issues or roadblocks, please send back to the list so everyone can be kept in the loop.
--Jeremy
########################################################################
Use REPLY-ALL to reply to list
To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1
|
