Documentation / User's manual in C++ :

Chapter II. The DataServer module

Chapter II. The DataServer module

Chapter II. The DataServer module

Abstract

This chapter is dealing with the DataServer module which is the spine of the Uranie project. In this module crucial classes are defined to handle data and transmit them to other modules, along with basic statistics implementation.
The source files are in souRCE.git/dataSERVER/souRCE and the corresponding namespace is URANIE::DataServer.

Table of Contents

II.1. Introduction

II.2. The TAttribute class

II.2.1. Nature of the attribute
II.2.2. List of variable information
II.2.3. Examples of use of the class TAttribute
II.2.4. Adding TAttribute when data are already available
II.2.5. Introducing the TStochasticAttribute classes

II.3. Data handling

II.3.1. Main format of input/output
II.3.2. Import data from an ASCII file
II.3.3. Import data from a TNtuple/TDSNtuple/TTree
II.3.4. Adding attributes to a TDataServer
II.3.5. Merging two DataServer
II.3.6. Pattern selection
II.3.7. Export to an ASCII file

II.4. Statistical treatments and operations

II.4.1. Normalising the variable
II.4.2. Computing the ranking
II.4.3. Computing the elementary statistic
II.4.4. The quantile computation
II.4.5. Correlation matrix

II.5. Visualisation dedicated to uncertainties

II.5.1. Histogram
II.5.2. Box-and-whisker("boxplot")
II.5.3. CDF, CCDF curves
II.5.4. Graph 2D with contour levels
II.5.5. Graph 2D "profile"
II.5.6. Graph 2D "Tufte"
II.5.7. Graph 2D "pairs"
II.5.8. Graph "CobWeb"
II.5.9. QQ plot
II.5.10. PP plot

II.6. Combining these aspects: performing PCA

II.6.1. PCA usage within Uranie

II.1. Introduction

The DataServer module is the spine of the Uranie platform as it is where the data are stored, the attributes (name given to the variable in Uranie) are gathered and important basic mathematical operations are performed. Objects and methods will indeed need a TDataServer to retrieve, process and transmit the results of their own operations.

Since it is used by all other technical libraries (Sampler, Launcher, ...), this library is the core library of Uranie (as shown in Figure I.1). The already discussed TDataServer contains two main objects (has shown in Figure II.1):

the header: represented in Uranie by the object TDataSpecification. Different information related to the variables (called in "attributes" in Uranie language and represented by TAttribute objects) are specified in Section II.2.
the data matrix: represented by the object TDSNtupleD (class derived from the ROOT class TTree). For an advanced use of this object, the reader can refer to chapter XII "Trees ROOT" in ROOT's user manual.

Figure II.1. Diagram of the class TDataServer

The chapter presentation will be articulated as follows. First of all, the way variable are handled will be introduced in Section II.2. This is a needed step, as all TDataServer have to be filled with attributes for being able to move along. The creation of a TDataServer (with many construction, from ASCII to ROOT files) is discussed in Section II.3 before discussing the basic statistical treatment (in Section II.4) and the dedicated visualisation tool (in Section II.5).

Tip

An important point to know for the use of Uranie: all options are not case sensitive. Actually, the treatment of these options is taken anyway as lowercase. The following instructions are equivalent:


tdsGeyser->draw("x1","","Nclass=Sturges");
tdsGeyser->draw("x2","","nclass=Sturges");
tdsGeyser->draw("x2","","nclass=sturges");
tdsGeyser->draw("x2","","NCLASS=STURGES");


I.3. The Python Interface		II.2. The `TAttribute` class