Skip to content

Latest commit

 

History

History
115 lines (89 loc) · 4.87 KB

README.md

File metadata and controls

115 lines (89 loc) · 4.87 KB
Raw

BenderScript

BenderScript is a kind of steroid-enhanced and doped GaudiPython session. BenderScript shares all pros and contras with GaudiPython. It definitely provides very simple and efficient way for exploring the data, in particular very simple way for investigation of the content for the input data and TES. One can easily loop over events, data containers, make simple calcualtions, apply large part of LoKi functors, use some part of DaVinci machinery, etc. Howver there is large part of tasks that are very difficult in BenderScrip:

  • large part of DaVinci tools is not working. It includes all tools and functionality related to e.g. actions with the associated best primary vertex. These manipulations are very complicated or some of them (e.g. PV-refit) is practically impossible to perform in a correct way without proper DaVinci context.
  • Many tools do not work without the significant efforts. The results of tools often are very difficult to save to be used/reused e.g. for subsequent processing (e.g. save manually created particles and vertices to TES and to output file)
  • Some very important and popular LoKi functors do not operate (mainly due to the issues from the previous item), e.g BPV* or *DV or DTF_* - those are not easy to use from the plain command line. There are some alternatives , but not for all functors.

See also a little bit more detailed TWiki-based tutorial on BenderScript

bender script

BenderScript session is started using the command bender. There are many command line options and keys for this script. {% challenge "Challenge" %} Invoke this command using -h option. Is the displayed help clear enough? {% endchallenge %} Many of the options/keys can be deduced from the input data itself, in particular from the names of centrally produced files.

Input data

It is mandatory to supply bender session with input data. There are two ways to do it

  1. supply data files as command like parameters
lb-run Bender/prod bender /lhcb/LHCb/Collision12/PSIX.MDST/00035290/0000/00035290_00000221_1.psix.mdst
  1. import the python file with data in a form of Gaudi Configurables, e.g. with IOHelper().Input = [ ...] 
lb-run Bender/prod bender -i data.py

where data.py is

from GaudiConf import IOHelper
IOHelper().inputFiles([
    'root://eoslhcb.cern.ch//eos/lhcb/grid/prod/lhcb/LHCb/Collision12/PSIX.MDST/00035290/0000/00035290_00000221_1.psix.mdst',
], clear=True)

Xml-catalogs can be supplied either via import from data.py or explicitely via -x-key

lb-run Bender/prod bender -i data.py -x xml_catalog.xml

Many of the options/keys (DataType, Simulation, Turbo, 'RootInTES', InputType , ... ) can be deduced from the input data itself, in particular from the names of centrally produced files.

The basic action at command prompt

The incomplete list of the basic commands at the command prompts is :

  • ls
  • get
  • run
  • skip
  • rewind

Do not forget that bender session is just a python session, therefore one always can invoke the commands dir and help.

{% challenge "Challenge" %} Start bender session and check the description for these method. Is it clear enough? {% endchallenge %}

More advanced commands are:

  • seekForData
  • seekStripDecision
  • seekForODIN
  • seekForEvtRun
  • seekForAlgDecision
  • seekForVoidDecision

{% challenge "Challenge" %} Start bender session and check the description for these method. Is it clear enough? {% endchallenge %}

The functions that help coding the actual scripts:

  • irun
  • execute

Histograms

Of course one can book & ROOT histograms at the command line

h1 = ROOT.TH1D('h1','histo',10,0,1000) 
for i in irun ( ... ) :
     data = get ( .... ) 
     h1.Fill ( len (  data )

Of course at the end one needs to save this histogram into some ROOT-file

As viable and practical alternative one can rely here on Gaudi-histograms:

h1 = book ( 'some_path_here' , 'title' , 100 , 0 , 100 )
...
h1.fill( x )

In this case all created histograms will be automaticlaly saved into the outout file with historgams, that is defined via the command line option --histofile, e.g. bender ... --histofile DVhistos.root

N-tuples/Trees

Again, oen can rely on bare ROOT (e.g. PyROOT) here to book/fill/save the tuples, but one can also use the functionality from Gaudi:

t = nTuple('QQQ','MyTuple')
data = ... 
for b in data :
   t.column_float ( 'pt' , PT ( b ) )
   t.column_int   ( 'id' , int ( ID ( b ) ) )
   t.column_float ( 'mass' , M ( b ) )
   t.write()

In this case all created histograms will be automaticlaly saved into the outout file with historgams, that is defined via the command line option --tuplefile, e.g. bender ... --tuplefile DVtuples.root