 |
 |
Hello Yuri, This came up at yesterday's software meeting; it will be addressed. Norman -----Original Message----- From: Yuri Gershtein [mailto:[log in to unmask]] Sent: Friday, August 17, 2012 9:25 AM To: Uemura, Sho Cc: Joey Reichert; Graham, Mathew Thomas; McCormick, Jeremy I.; Graf, Norman A.; hps-software Subject: Re: ze problem Hi Sho, thanks, that makes sense. I guess may be at some point it would be good to have a "tour" of the drivers to understand what can be varied so that questions like "what can we do to the detector configuration to be sensitive to this" and "what cuts in the reconstruction reject our tracks" can be answered. -y On Aug 17, 2012, at 12:06 PM, Sho Uemura wrote: > I don't know much about the tracking strategies or about the detectors, but I think the relevant difference between TestRun-v2 and EcalTest is the beam gap. > > TestRun-v2 uses the actual SVT position (roughly) we had in May, which was for photon running - so the tracker halves are spread wide and your acceptance for far-forward decays is terrible. It looks like EcalTest has the nominal (electron running) beam gap. So maybe you don't want to use TestRun-v2. I can look around and see if there's a more suitable detector for you to use. > > Make a plot driver, stick it into the steering file, and put AidaSaveDriver in the steering file. You might look at ecal_fadc_bkgd.lcsim as an example of how to do that. But yes, it's fairly simple. > > Pushing this thread into hps-software, by the way - we're trying to use this list more. > > On Fri, 17 Aug 2012, Joey Reichert wrote: > >> Hi Sho, >> >> I am seeing MatchedTracks now, I'm surprised at how few of them >> though. I'm seeing about 30 tracks in files with 500 events with this >> setup whereas the HPS-EcalTest setup has been giving us closer to 300 tracks per 500 events. >> Since we're generating events with four electrons in the final state, >> that's about 1% of electrons reconstructed, which seems rather low. I >> tried loosening up all of the cutoffs in the tracking strategy too, >> but that didn't seem to help...Should there be that many fewer >> reconstructed particles with the TestRun-v2 detector compared with the EcalTest? >> >> In either case, the next thing we'd want to do is learn how to create >> plots with this type of analysis. Is it going to be as simple as >> creating a plot driver and loading it into the HPSTrackingDefaults.lcsim? >> >> Thanks again, >> Joey >> >> On Thu, Aug 16, 2012 at 7:42 PM, Sho Uemura <[log in to unmask]> wrote: >> >>> Try changing the tracking strategy to HPS-TestRun-v2-357.xml. I was >>> just now testing stuff with HPS-Test-4pt0.xml and not getting >>> MatchedTracks. I don't honestly know what strategy you're supposed to be using though. >>> >>> >>> On Thu, 16 Aug 2012, Joey Reichert wrote: >>> >>> Hi Matt, >>>> >>>> I'm still slightly unclear about this. I have a slcio file made via >>>> SLIC, which ran over a stdhep file. Is there some intermediate >>>> slcio file involved in SimDist that I'm forgetting about? >>>> >>>> I understand that my slcio file does not have the extended classes >>>> like SeedTrack, HelicalTrackHits, etc. to start, but the >>>> HPSTrackingDefaults.lcsim opens up the slcio file and finds these >>>> specific types of hits, tracks, etc. So all of these become >>>> accessible, including the TrackerHits, which I'm seeing plenty of. >>>> In principle, these TrackerHits should be converted to >>>> MatchedTracks by one of the drivers called in the xml file (I think >>>> the TrackerReconDriver?), which we could then do analysis on and do >>>> things like create plots as long as we do it all within one job >>>> (which means the *_trackRecon.slcio files would not show any >>>> extended class information). But since I'm not seeing >>>> MatchedTracks, I can't go beyond that point yet. >>>> >>>> So it sounds like it's an issue on my end, since you're saying >>>> MatchedTracks should be created with HPSTrackingDefaults.lcsim. Do >>>> you guys have a sample slcio file created by SLIC that works with >>>> the HPSTrackingDefaults.lcsim that I could use to check that >>>> everything is consistent? >>>> >>>> Also: we're interested in looking at the record of what happens in >>>> the detector (using HPS-EcalTest), which we believe GEANT should be >>>> able to tell us. Is that a better question for Norman, perhaps? >>>> >>>> Thanks again, >>>> Joey >>>> >>>> On Wed, Aug 15, 2012 at 4:09 PM, Graham, Mathew Thomas < >>>> [log in to unmask]> wrote: >>>> >>>> >>>>> Let me just try to summarize: >>>>> >>>>> ---> you have a slcio file (selected.slcio) made by running on >>>>> ---> another >>>>> slcio file (orig. slcio) and selecting some events (say, requiring >>>>> at least one track) >>>>> ---> now you want to read selected.slcio and do some analysis on >>>>> ---> the >>>>> tracks?. >>>>> >>>>> ======== This likely doesn't work out the box because lcio only >>>>> knows about "Tracks", so the "SeedTrack" info is dropped, >>>>> similarly for the hits >>>>> ---> you should be able to run selected.slcio with >>>>> HPSTrackingDefaults.lcsim and get HelicalTrackHits and SeedTracks (i.e. >>>>> "MatchedTracks") though?as long as there are TrackerHits in your >>>>> collection >>>>> >>>>> >>>>> >>>>> >>>>> btw, you HPS-TestRun-v2 is in the JLAB from (beam is in z, >>>>> non-bend is y) >>>>> >>>>> On Aug 15, 2012, at 12:53 PM, Joey Reichert <[log in to unmask] >>>>> <mailto: >>>>> [log in to unmask]>> wrote: >>>>> >>>>> Hi All, >>>>> >>>>> Is what I said in the last email about seeing MatchedTracks >>>>> correct? And are you guys seeing MatchedTracks with SLIC lcio >>>>> files? If so, I'd imagine it comes down to either a problem with >>>>> my (SLIC) lcio file, or my build of one of the hps-java drivers. >>>>> In either case, I can investigate, but I just want to check to >>>>> make sure my understanding of the problem is correct before going >>>>> too far. >>>>> >>>>> Thanks for all the help, >>>>> Joey >>>>> >>>>> On Fri, Aug 10, 2012 at 3:20 PM, Joey Reichert >>>>> <[log in to unmask] <mailto:[log in to unmask]>**> wrote: >>>>> Hi Guys, >>>>> >>>>> This explains why things were not working in JAS, but I should >>>>> still be seeing MatchedTracks after calling >>>>> HPSTrackingDefaults.lcsim, right? With the debug option on >>>>> TrackerReconDriver set to on, I'm seeing the HelicalTrackHits, but still no MatchedTracks. >>>>> >>>>> Also, what reference frame does HPS-TestRun-v2 assume? I'm >>>>> guessing the JLab frame, and I've run lcio files in either frame >>>>> through the HPSTrackingDefaults.lcsim without seeing >>>>> MatchedTracks, but it'll be one less thing to worry about. >>>>> >>>>> Thanks again, >>>>> Joey >>>>> >>>>> >>>>> On Tue, Aug 7, 2012 at 10:00 PM, Graham, Mathew Thomas < >>>>> [log in to unmask]<**mailto:[log in to unmask]**edu<mgr >>>>> [log in to unmask]>>> >>>>> wrote: >>>>> >>>>> Ah?this is likely related to the issue Elwin is having too. >>>>> >>>>> >>>>> Thanks! >>>>> >>>>> On Aug 7, 2012, at 6:39 PM, Joey Reichert <[log in to unmask]<mailto: >>>>> [log in to unmask]><mailto:**[log in to unmask]<[log in to unmask] >>>>> tgers.edu> >>>>> <mailto: >>>>> [log in to unmask]>>> wrote: >>>>> >>>>> Hi Jeremy, >>>>> >>>>> Ah, that makes sense. No wonder it wasn't working...Are there any >>>>> downsides to running everything in one job? That's essentially >>>>> what JAS does, correct? >>>>> >>>>> Thanks for all the help! >>>>> Joey >>>>> >>>>> On Tue, Aug 7, 2012 at 7:53 PM, McCormick, Jeremy I. < >>>>> [log in to unmask]<**mailto:[log in to unmask]**edu<jer >>>>> [log in to unmask]> >>>>>> <mailto: >>>>> [log in to unmask]<**mailto:[log in to unmask]**edu<jer >>>>> [log in to unmask]>>>> >>>>> wrote: >>>>> Hi, Matt. >>>>> >>>>> Joey and I were able to figure out the issue with his analysis. >>>>> When collections are written out to an LCIO file, extended class >>>>> information is lost. For instance, HelicalTrackHits become plain >>>>> old TrackerHits. So that means that if there is code that depends >>>>> on these specific classes being present, it will fail if you read >>>>> back the LCIO file in a second job. >>>>> It must be done ?inline? within the same job when the >>>>> HelicalTrackHits are >>>>> >>>>> accessible. He was trying to pick out 2 track events, write them >>>>> to an LCIO file, and then read them back for analysis. But this >>>>> doesn?t work for this very reason. LCIO doesn?t know anything >>>>> about HelicalTrackHits. It only understands basic TrackerHits. >>>>> >>>>> I don?t know of any way to get around this problem, other than >>>>> running >>>>> >>>>> everything in the same job. Because there is no way in LCIO to >>>>> tell what specific class you used to generate the objects in the >>>>> first place. Do you usually just run everything including your >>>>> plotting in the ?evio 2 lcio? >>>>> step? >>>>> >>>>> --Jeremy >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>> >> -------------------------- Prof. Yuri Gershtein [log in to unmask] http://physics.rutgers.edu/~gershtein (732)445-5500 x1794 W316 Serin Building Department of Physics and Astronomy 136 Frelinghuysen Rd Rutgers University Piscataway, NJ 08854 ######################################################################## 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