Getting started with DELPHI data

Documentation Guide

This quick start guide is meant as a guide for the very first steps to get going with DELPHI software and data access.

Overview and requirements
- Before you start
Accessing the software stack
- Docker
- CVMFS
Documentation
Examples

Overview

The DELPHI stack consists of the following modules:

DELPHI DST analysis framework, also referred to as dstana. Please take a look at the Skelana framework for detailed information.
MonteCarlo production, includes simulation, reconstruction and short DST creation. Note that the DST format depends on the year. During LEP1, mainly short DST or long DST formats were used, while during LEP2 the extendent short DST format was the default. Note that the reconstruction code returns results in full DST format. Analyses should always use short or extended short DST format, because these generally contain fixes which were applied after the reconstruction itself.
Event reconstruction from raw data, using the DELPHI event server.
The graphical DELPHI Event display, also referred to as delgra.

Before you start ...

Please read and accept the DELPHI data access rules. They can as well be viewd at the the DELPHI web pages.

Accessing the software stack

There are two possible ways to access the software stack.

Docker

A container image is available which ships with all the modules installed. Please take a look at the DELPHI docker guide.

CVMFS

Binaries are also available from CVMFS, for a variety of different Linux flavors. Instructions can be found in our dedicated CVMFS guide.

Documentation

DELPHI manuals and notes are available from https://cds.cern.ch/

Here is a selection for getting started:

Source code

The sources are available from https://gitlab.cern.ch/delphi. Some modules still requires CERN authentication. The plan is to release the software in the near future.

The DELPHI stack is mostly written in Fortran, with some bits written in C. Only gfortran and GNU gcc are supported. We use the gfortran version which comes with the supported operating system. Recommended compiler and linker flags are set by the environment and should be used when building software linking with the DELPHI libraries.

Getting help

The collaboration main contact for data preservation is the mailing list DELPHI-data-preservation-board@cern.ch. Support can only be given on a best effort basis. Suggestions and feedback is of course welcome!

Examples

Some basic examples of how to run the software stack and perform various tasks can be found in the /cvmfs/delphi.cern.ch/examples tree.

In the following, we will

Create some Monte Carlo events and run simulation, reconstruction and DST production on them
- First, we will show how to do so interactively
- Then, how to do this on the batch farm at CERN
Show how to read the result from an analysis job

About simulation

Simulation is reconstruction code is supported for all the years 1992 and later. The code differs a bit for each of the years.

The DELPHI simulation code comes with a few build-in generators. Typically, generators are run externally though.
The script runsim is used to do the detector simulation, reconstruction and short DST creation steps. Use -gext to process an external file from some generator.

There is a dedicated guide available with more details about how to run DELPHI simulations.

Creating Monte Carlo samples interactively

Using internal generators

For creating a few events with a build-in generator run

$ runsim -VERSION v94c -LABO CERN -NRUN 1000 -EBEAM 45.625 -igen qqps -NEVMAX 10

This will create 10 Z-> qqbar events at a beam energy of 45.625 GeV (ECM 91.25 GeV), using the build-in QQPS generator and the detector setup of 1994.

It will as well pass the events through the detector simulation, and create a short DST file named simana.sdst which can be used for analysis, scanning etc.

Using external generators

In this case the generator is run externally and the output is written to a file in a specific format. This can then be passed through the detector simulation with

$ runsim -VERSION 94c -LABO CERN -NRUN 1000 -EBEAM 45.625 -gext generated.lund -NEVMAX 10

Again, this will create a short DST file name simana.sdst. Note that for the LEP 2 phase, a sol called extended short DST file will be created, which is typically called simana.xsdst instead.

Running an analysis job on the result

A dedicated guide is availabe with more details about how to run an analysis job for DELPHI.

The DELPHI analysis framework is called skelana. The way to write analysis code is by overriding one or more of the following Fortran functions:

USER00 for the initialisation at the start of the run
USER01 for the event selection. It should return 0 to skip the current event and 1 to read it.
USER02 is called for each selected event.
USER99 is called at the end of the run

An example can be found in the examples folder on the DELPHI container or on CERN gitlab. Take a look e.g. at the dump folder there.

Using nicknames

DELPHI data is organised in data sets which are identified via a nickname. When using opendata, each nickname comes with a corresponding DOI which you can quote.

To analyse data, use the nicknames which you can find at http://delphiwww.cern.ch/offline/data/castor/html. In this case, the PDLINPUT file created by the script above should contain the keyword PDL, followed by the nickname, e.g.