Chapter 6. Reconstruction tutorial

Table of Contents

6.1. Installation
6.2. Connecting the chain

The reconstruction software takes the output from the digitization (or, when the time is there, from the real detector), which you can best think of as a stream of bits that comes off the electronics of the detector, and turns it into physics quantities. There is only a fine line between reconstruction and analysis,[13] and this tutorial will partly blend both stages. Otherwise, you can consider the physics quantities that result from the recontruction phase, typically stored in the form of ntuples and histograms, as the products to be fed into the analysis.

This tutorial demonstrates how to run the reconstruction within the context of the Atlas DC2 software and how to tune it to your analysis needs. It is assumed that ASK is properly set up on your system, and that you are working with v2.0; there may be subtle changes in futureversions.

The text in this tutorial is a bit more elaborate than the other parts of the DC2 chain, because it was historically standalone. Also, it used to be given with the GUI, and this is still the case. If you prefer (and you should), then you can find the CLI steps outlined in sidebars.

6.1. Installation

If this is your first ASK based tutorial, or if you want to start clean, refer to Chapter 1, Creating a workspace for details on how to setup a workspace. On the other hand, if you have just completed Chapter 4, Simulation tutorial for which you already created a clean environment, you can run from that same environment.

Start the ASK graphical interface (by executing the ask --gui command on your shell (or with ask --gui --detach if you work on a small laptop screen); and click on the tab labeled with Athena (if it is not already selected).

[Tip]Tip

The release screen is usually your starting point for new (local) packages. Both creation and checking out of packages is done from there. If your work directory does not contain any packages, then ASK will start on the release screen. Otherwise, it will start on the Athena screen.

In order to deal with changes and defects in releases, the reconstruction comes with a shell script that performs some additional setup and copies the most useful options scripts into your run directory.

With reference to the following screen shot, enter the name of the script "RecExCommon_links.sh" in the scripts text box. Press the get script button to retrieve the script.

After the setup script has been retrieved, go to the tab labeled run-time and select the RunTime package (if it is not already selected). Select the share/RecExCommon_links.sh script from the installation files listing. With this script selected (compare with the screen shot below), press the source button to finish the installation.

Finally, run an example job. Go back to the Athena screen by clicking on the tab labeled Athena, and select the run/myTopOptions.py options file (see screen shot below). Press run to run ATHENA with this script.

After the run, you will find a log file called AskRun.log in the package run directory (that is: Reconstruction/RecExample/RecExCommon/RecExCommon-*/run from under your installation directory). Any other output files of the reconstruction job are located in this same directory and in particular, there should be a ntuple.root file. Open the file with ROOT, locate the combined ntuple ("t3333"), and display a few parameters. If you have run the generators tutorial, verify that the truth information is still the same.



[13] Especially in the python world, where interactive access to data and algorithms allows you to rerun parts of the reconstruction during analysis.