Anne-Laure Pequegnot | 17 Dec 17:29 2014
Picon
Picon

Is there a Morphing for histograms (not PDF)?

Dear ROOT experts,

I was looking at your RooMomentMorph and RooIntegralMorph classes of RooFit. I also looked at your documentation about the interpolation between multi-dimensional histograms using a new non-linear moment morphing method: http://arxiv.org/pdf/1410.7388.pdf

I want to use these morphing methods for my proper analysis. I have histograms template as an input, which represent my signal shape, normalized to the number of expected events. The problem is that these histograms can have negative bins and cannot be converted into RooHistPdf, and so unfortunately I can't use the RooMomentMorph class.

Do you know some tools that can do the morphing as RooMomentMorph but on histograms? If so, can these tools do the interpolation with multiple morphing parameters?

Thanks a lot for your precious help!

Best regards,

Anne-Laure PEQUEGNOT

-- Anne-Laure PEQUEGNOT PhD Student, Particules Physics -- Institut de Physique Nucléaire de Lyon Building Paul Dirac, office 116 4, Rue Enrico Fermi 69622 Villeurbanne Cedex Tel. : +33 4 72 44 82 70
Pere Mato Vila | 16 Dec 18:50 2014
Picon
Picon

HSF SLAC Workshop Reminder

Dear colleague, 

The upcoming workshop of the HEP Software Foundation is an important milestone for this initiative. I remind you that it will take place very soon, January 20-21 at SLAC. See the workshop announcement, http://hepsoftwarefoundation.org/workshop-slac-jan-2015, for the details. A draft agenda is available at http://indico.cern.ch/e/HSFWorkshop2.

If you are interested to participate, please register asap using the SLAC workshop registration page. Since we are stressing the bottom-up style of the initiative by asking the opinion and suggestions to the whole community,  you are also encouraged to submit a contribution proposal using the abstract submission page. If you are not directly involved, please forward this announcement to all the people that may be interested. We would like to see as many software projects and software developers represented: the Foundation will be what people makes of it!

Also, two mailing list related to these activities have been created and we'd like to see as many people as possible registering to them.
- HEP Software and Computing (general list): https://groups.google.com/forum/#!forum/hep-sw-comp

Cheers, 

  Pere Mato & Torre Wenaus  (for the HSF Startup Team) 
Nevzat Guler | 16 Dec 06:06 2014

Transparent filling with TGraphErrors

Dear ROOTers,

I am not able to use SetFillColorAlpha(kBlue,0.4); with TGraphErrors. Is there anything equivalent to create transparent fill with TGraph?

Best regards,

- Nevzat

 

ROOT Version   5.34/13  30 November 2013



Nicholas Devenish | 9 Dec 20:51 2014
Picon

Rint fix in 5.34 series?

I've run into an error that appears to have been fixed for root 6, but not applied to the 5.34 patches; It causes custom dictionary generation to fail when built with cmake.

Namely the fix -Rint -> -lRint in RootConfiguration.cmake in:

https://github.com/root-mirror/root/commit/e4f4d6d5d5a7cc9140a07ee319613583a013a80d#diff-23f7e69bebd4294cd5b13338332bf0e5


Was this purposefully left out of the 5.34 releases? And how likely to make it into any (potential) 5.24/24 is it?


Thanks,


Nick

Marcelo Zimbres | 9 Dec 19:49 2014
Picon

Realtime use of TMultiGraph

Hi,

I have been using TGraph to visualize realtime data. It has the very
nice feature of recalculating the axis range as new points are set via
 g->SetPoint(i, t, f); However, when I try to put my three graphs
together in one TMultiGraph, the axis range shown on the screen is not
recalculated for each point set, so after a few seconds, I cannot see
the data plotted anymore as the axis are static now. Do you know how
to get axis range to be recalculate on each draw of the TMultiGraph?

Best regards,
Marcelo

Indranil Das | 2 Dec 03:13 2014
Picon
Picon

Object/key not shown in ".ls" however visible in TBrowser

Dear experts,

I encounter a problem when tried to extract a TCanvas named "systematics_J/#psi" from file. It is visible in TBrowser but not in ".ls" command. Please find the file at,
/afs/cern.ch/user/i/idas/public/JpsiSimpleFit_C0MUL_pp8TeV_JPsi_C0MUL_T0_extended_ptrap_hTotDimuHisto.root

A snapshot is attached.

How to get the canvas using root macro ?

regards

Indra
Francesco Perfetto | 1 Dec 06:44 2014
Picon
Picon

Re: Superimpose two his to after projectt3d

Thanks Carolina,

but don’t work.

The problem is that  TH1 *hpyx2 = h2->Project3D("yx”); 
Draw on canvas and the previous plot is lost.

Regards
Francesco.

After then TH1 *hpx2 = h25->Project3D("z"); replaces the previous plot, eventhough hpx2->Draw();
Il giorno 01/dic/2014, alle ore 01:16, Carolina Garcia <caro24garcia <at> gmail.com> ha scritto:


Hi,

The last line might be    hpyx2->Draw( "same"); }



Carolina

On Sunday, November 30, 2014, Francesco Perfetto <francesco.perfetto <at> na.infn.it> wrote:
> Hi Rooters,
> if possible I need your help.
> I would like superimpose two or more histos after I do a Project3d(“yx”) on a TH3I, but the second plot (and so following) is not overlaid to the previous one.
> I do:
> {
>    TCanvas *c = new TCanvas();
>    TFile file("picchi/all_files_CPU_ACQUISITION_new_version_decode_2.root");
>    TTree *T = (TTree*)file.Get("testtree”); 
> // In the tree there is the TH3I h3_photon
>    TH3I *h3_photon = 0;
>    testtree->SetBranchAddress("h3_photn",&h3_photon);
>    testtree->GetEntry(106432); // get the entry
>    
>    TH3 *h1  = (TH3*) h3_photon;
>    h1->GetZaxis()->SetRange(59,59);
>    TH1 *hpyx1 = h1->Project3D("yx");
>    hpyx1->Draw(“colz");
>    TH3 *h2 = (TH3*)fh.Get("test_h3");
>    h2->GetZaxis()->SetRange(63,63);
>    TH1 *hpyx2 = h2->Project3D("yx”); 
> After this line new plot is done that replace the previous one 
>    hpyx2->Draw(“colz same");
> }
>
> Thanks in advance for your help.
> Francesco.

Francesco Perfetto | 1 Dec 00:26 2014
Picon
Picon

Superimpose two his to after projectt3d

Hi Rooters,

if possible I need your help.
I would like superimpose two or more histos after I do a Project3d(“yx”) on a TH3I, but the second plot (and so following) is not overlaid to the previous one.

I do:

{
   TCanvas *c = new TCanvas();
   TFile file("picchi/all_files_CPU_ACQUISITION_new_version_decode_2.root");
   TTree *T = (TTree*)file.Get("testtree”); 

// In the tree there is the TH3I h3_photon

   TH3I *h3_photon = 0;
   testtree->SetBranchAddress("h3_photn",&h3_photon);

   testtree->GetEntry(106432); // get the entry
   
   TH3 *h1  = (TH3*) h3_photon;
   h1->GetZaxis()->SetRange(59,59);
   TH1 *hpyx1 = h1->Project3D("yx");
   hpyx1->Draw(“colz");

   TH3 *h2 = (TH3*)fh.Get("test_h3");
   h2->GetZaxis()->SetRange(63,63);
   TH1 *hpyx2 = h2->Project3D("yx”); 

After this line new plot is done that replace the previous one 

   hpyx2->Draw(“colz same");
}


Thanks in advance for your help.
Francesco.
Nevzat Guler | 27 Nov 16:44 2014

Vertical separation between legend items

Dear Rooters,

I am trying to find a way to arrange vertical distance between legend items, for esthetic purposes.

I tried the function SetEntrySeparation() but did not observe any effect from it (see below).

I appreciate any suggestion.

Best regards,

Nevzat Guler

 

Version   5.34/13  30 November 2013

P.S. An example code snipped:

        if( stTar == "ND3" || "NH3") TLegend* leg2 = new TLegend(0.60,0.20,0.98,0.65);

        if( stTar == "N")            TLegend* leg2 = new TLegend(0.60,0.20,0.98,0.65);

        leg2->SetFillStyle(0);

        leg2->SetBorderSize(0);

        leg2->SetTextFont(13);

        if( stTar == "N") leg2->AddEntry(TG_gammaE94,"E94","p");

        if(stTar == "ND3") leg2->AddEntry(TG_maid1,"MAID","l");

        leg2->AddEntry(TG_lensk,"Lensky et al.","f");

        leg2->AddEntry(TG_meiss,"Bernard et al.","f");

        leg2->AddEntry(TG_kao1,"Kao #it{O(p^{4})}","l");

        leg2->AddEntry(TG_kao2,"Kao #it{O(p^{3})}","l");

        leg2->AddEntry(TG_fsppownglmod1,"Model","l");

        leg2->SetEntrySeparation(0.01);

        leg2->Draw();

 

Chris Jones | 26 Nov 17:49 2014
Picon
Picon

ROOT 6.02.02 fails to build on OSX 10.10

Hi,

Full build and configure logs attached, but in short

> cmake ../root-6.02.02 -Dgnuinstall=ON -DCMAKE_INSTALL_PREFIX=/Users/chris/Projects/install
> make -j 4

fails with

[ 57%] Generating etc/allDict.cxx.pch
ERROR in cling::CIFactory::createCI():
   resource directory 
/Users/chris/Projects/install/etc/root/cling/lib/clang/3.6.0 not found!
input_line_3:1:10: fatal error: 'RtypesCore.h' file not found
#include "RtypesCore.h"
          ^
[ 57%] Building CXX object 
math/mathcore/CMakeFiles/MathCore.dir/src/TDataPointN.cxx.o
lookup.type.by.name.file:1:1: error: unknown type name 'Double32_t'
Double32_t
^

  *** Break *** segmentation violation

Seems again ROOT was not fully tested with the GNU install option 
enabled prior to release ;)

cheers Chris

Attachment (configure.log.bz2): application/x-bzip, 4846 bytes
Attachment (build.log.bz2): application/x-bzip, 21 KiB
Alessandro Bertolin | 24 Nov 14:26 2014
Picon
Picon

is EvaluateMVA also available for "BDTF"/"BDTMitFisher" ?

Dear roottalk,

  I have a question about the TMVA package
in TMVAClassification.C I see that I can train a BDT "flavour" called
"BDTF" with corresponding BookMethod string "BDTMitFisher"
in TMVAClassificationApplication.C instead I see that I can EvaluateMVA
with "BDT method", "BDTD method" or "BDTG method"
so apparently "BDTF"/"BDTMitFisher" is not available anymore for the 
application step ...
am I missing something ?
am I doing something wrong ?
thanks a lot for your time and patience,

        Alessandro

Alessandro Bertolin
Istituto Nazionale di Fisica Nucleare
Sezione di Padova
Via Marzolo 8, 35131 Padova - Italy
Tel: +39 049 9677094 (office)
E-mail: alessandro.bertolin <at> pd.infn.it
Home page: http://www.pd.infn.it/~bertolin/TOP.html


Gmane