CustomAnalysis < DetDev

DetDev Web>EUTelescopeAndIBLTestbeam>CustomAnalysis (2013-09-26, KennyWraight) (raw view)
---+!! %MAKETEXT{"EUTelescope And IBL Testbeam pages" args="<nop>%WEB%"}%

Notes on customising analysis of IBL testbeam data using =tbmon=.

---++ %MAKETEXT{"Analysis Cuts"}%

Various cuts are available in each of the analysis processes. Common cuts to all analyses which ccan be used involve the fundamental objects themselves such as analysis events and DUTs.

<b>Proximity</b>
 
When a DUT is defined the last two numbers in the return function for FE-I4 and FE-I3 DUTs denote the allowed proximity of tracks; x (cols) and y (rows), respectively and both in microns. This is used for the definition of efficiency, which is the ratio between the number of reconstructed tracks that hit the detector in the proximity of a fired pixel to the total number of reconstructed tracks that hit the detector. The proximity region is defined as a x-y distances in the makeduts.h file in the core/include directory of tbmon.

e.g.
<code>
DUT* m20 =  makeFEI4("MICRON01", dutnumber, 2, 250.0, 50.0);
</code>

#Chi2Anchor
<b>Chi Squared</b>
The quality of fitting is defined using a member function of the =config= class,  =Chi2Builder=. To check the chi2 distribution of the events see the =maxcellres_DUT_chi2= histogram.

e.g.
<code>
config.addBuilder(new Chi2Builder(15.0));
</code>

#Lv1Anchor
<b>Level 1 (out-of-time)</b>
To mitigate the effects of out-of-time hits a level-1 cut is made using the DUT object method =lv1Range=. To check the chi2 distribution of the events see the =maxcellres_DUT_lv1= histogram.

e.g.
<code>
m20->lv1Range(5, 9);
</code>

---++ %MAKETEXT{"Analysis Processes"}%

Assuming you are in =PATH_TO_ANALYSIS= directory (i.e. something like =path_to_tbmon/tbmon/trunk=) ...

To add new code to analyses you'll need to create/edit files in =PATH_TO_ANALYSIS/analysis/= part of the file structure. This is done in the standard header/source code way. A header (i.e. =.h= file) must be in the =analysis/include= directory with all declarations (classes, member functions, variables etc.). The corresponding source (i.e. =.cc= file) file must be in the =analysis/src= directory with all the analysis functionality.

If you are creating a new anlysis you'll need to tell the rest of the software about it. Once you have created =someanalysis.h= and =someanalysis.cc= files the code must be introduced in the =driver.cc= file in the =PATH_TO_ANALYSIS= directory. The analysis must be added to the set of included files at the beginning of the =driver= file (e.g. =#include "someanalysis.h"=) and added to the list of possible analyses in the =allAnalyses= function (e.g. =config.addAnalysis(new <nop>SomeAnalysis</nop>, "someanalysis", dut);=).

The last thing to do is mention the new analysis in the =Makefile= so that it is compiled along with the rest of the software. You must add a =.o= mention (e.g. =someanalysis.o= to the =ANALYSIS= section of the Makefile.

After any changes the code must be recompiled: 

<code> &gt;<cite> make </cite></code>

Then, supposing you have gone through all the steps required before running analysis algorithms* ([[DetDev.EUTelescopeAndIBLTestbeam#AnalysisAnchor][here]]) you can run your custom analysis like any other task: 

<code> &gt;<cite> ./tbmon -s/r/l RUN_NUMBER(s)/RUN_LIST -a ANALYSIS_TASK -c CONFIGURATION_SET </cite></code>

* Changes made only to =analysis= files, rather than =core= or =eventbuilders=, will not require rerunning of hotpixelfinder, checkalign, getetacorr, checkalign steps.

---++ %MAKETEXT{"Standard Plots"}%
Here are some standard plots used for analysis.

maxcellres_DUT_lv1 -- used for optimising level1 cuts (see [[#Lv1Anchor][above]])

maxcellres_DUT_chi2 -- used for fitting tracks to clusters ( see [[#Chi2Anchor][above]])

efficiency_DUT_effDis -- spectrum of pixel efficiency 

efficiency_DUT_effMap -- 2D map of pixel efficiency 

clustercheker_DUT_unmatchClusterSize -- size of unmatched clusters

clustercheker_DUT_matchClusterSize -- size of matched clusters

clustercheker_DUT_matchClusterSizeX/Y -- size of X/Y matched clusters
 
residuals_DUT_cluSizeX/Y -- size of residuals in X/Y, used to compare <latex [attr="value"]* > \sigma </latex> of fitted Gaussian (measured resolution) to compare to <latex [attr="value"]* > \frac{pitch}{\sqrt[2]{12}} </latex> (ideal resolution).

-- Main.KennyWraight - 2013-06-11
Topic revision: r4 - 2013-09-26 - KennyWraight
DetDev