Understanding control file structure

To run analysis there is a simple command

ana control_file root_out

Control file tells to ana program what to do. Results (histograms and some parameters) are saved into root out file.

There are several tags one can use in control file. So quite complicate analysis could be programmed through it. The simplest way of using control file is to Analise processed or slimmed data. Consider example control.dat

#This is a simple analysis of run 3395. Run duration 86100 s.
/unix/nemo1/reco7.0/data/3395.root data 86100 0 D
/unix/nemo1/reco7.0/mc/mo100gs/mo100_2b2n_101_a.root mo2b2n 0.129 150540 B

• each line represents one analysis component to load
• comment lines in the control file should start from # symbol
• line begins from ROOT tree filename, then the component name to refer during analysis, activity of component in Bq (or time in s for raw data), number of MC events generated (0 if automatic normalisation wanted). Last letter is tag.
  • B = background. Normally used for every MC components.
  • D = rawdata. Note there could be only one rawdata component. Also analysis without data makes little sense. So one component with D tag is obligatory!
  • S = signal. Don't be confused with the word signal, it could be any component you want to fit (not only 2b2n!). This tells do not pay much attention to activity of this component. Instead, it will be fit at the end. There could be several signals, then TFractionFitter ROOT method used to fit data simultaneously with all of them.
  • L = component to put limit at.

  one can put option E after tag to tell absolute error in activity. This error wil be used during the fit.

Run the example with

ana control.dat save.root

Following picture will be drawn as a result Fig. 1.

Figure 1: Total energy for events with two scintillators. Crosses - raw data. MC prediction for mo2b2n shown for comparison.

Looping through all the trees is time consuming, so if you want to play with different activities values, or fit components in a different way, or draw another histogram etc. you do not need to perform analysis again. Following example shows how to reuse saved histograms:

#This reuse saved hystograms to fit 2b2n signal
save.root data 0 0 DR
save.root mo2b2n -1 150540 SR

• line begins with a filename, a ROOT file with histograms saved after previous call to ana
• component name should be exactly the same as it was during analysis. ana program will look for this component in the ROOT file and load all its histos.
• It is not necessary to specify time or number of MC events anymore, since one is stored in the save.root file. But, if you do put positive value, it will be used instead of stored one!
• One can reuse stored activity, putting some negative value in the control file. If zero or positive value provided, it will be used as an activity.
• All the same tags used, just R added to tell that this data is already READY and resulting histos should be read from the file.

Run the example with

ana control.dat rooty.root

Note to use different (from save.root) output filename!

One can combine different data sources:

#This reuse saved hystograms and adds new component
save.root data 86100 0 DR
save.root mo2b2n 0.129 150540 BR
/unix/nemo1/reco7.0/mc/swire_bi214/swire_bi214_636.root bi214sw 0.702 500000 B E 0.070

Other useful, but not necessary for first try control file options are component dependencies, and multiple channels analysis.

If component activity related to each other (e.g. same decay chain), one can specify them as dependent:

#This reuse saved hystograms and adds new component
save.root data 86100 0 DR
save.root mo2b2n 0.129 150540 BR
/unix/nemo1/reco7.0/mc/swire_bi214/swire_bi214_636.root bi214sw 0.702 500000 B
/unix/nemo1/reco7.0/mc/swire_pb214/swire_bi214_230.root pb214sw 0.702 1000000 B D bi214sw 1.

This tells analysis program to ignore activity for pb214sw component, and use the same activity as for bi214sw, multiplied by factor 1.

Another example:

#This reuse saved histograms and adds new component
save.root data 86100 0 DR
save.root mo2b2n 0.129 150540 BR
/unix/nemo1/reco7.0/mc/swire_bi214/swire_bi214_636.root bi214sw 0.702 500000 S
/unix/nemo1/reco7.0/mc/swire_pb214/swire_bi214_230.root pb214sw 0.702 1000000 S D bi214sw 1.

Though there are two signal components specified in control file, they are dependent. So fit will be performed, taking into account this dependency, with only one free variable, bi214sw activity.

Yet another useful option is E - error in activity. If activity od background component is known with some error. This error will be stored used in fit error calculation.

#Specify activity error
save.root data 86100 0 DR
save.root mo2b2n 0.129 150540 BR
/unix/nemo1/reco7.0/mc/swire_bi214/swire_bi214_636.root bi214sw 0.702 500000 B E 0.011
/unix/nemo1/reco7.0/mc/swire_pb214/swire_bi214_230.root pb214sw 0.702 1000000 B E 0.011 D bi214sw 1.

If there is a need to fit some component in several channels, there is a multiple channel analysis implemented. By default all components belong to channel 0. But user can explicitly order different channel. E.g. search for decay to excited states in 2e1g and 2e2g channels.

#Multiple channels analysis
2e1g.root data 86100 0 DR       C 0 eeg
2e1g.root mo2b2nx 0.0 1000000 SR C 0 eeg
2e2g.root data 86100 0 DR       C 1 eegg
2e2g.root mo2b2nx 0.0 1000000 SR C 1 eegg

Tag C tells that there is a channel definition after that. Channel number and channel name should be provided.

Note, that ana is not a user friendly program! This means it doesn't check if your instructions in control file have sense or not. Specifying nonsense (like channel 0 with different names, or tag mo2b2nx as signal in channel 0, and as background in channel 1) most probably will result to program crash.