EUDAQ Software User Manual

Document technical information

Format pdf
Size 913.2 kB
First found Jun 9, 2017

Document content analysis

Language
English
Type
not defined
Concepts
no text concepts found

Persons

Organizations

Places

Transcript

EUDAQ User Manual
EUDET
EUDAQ Software User Manual
EUDAQ Development Team
As of 28th June 2014
for EUDAQ version v1.3.0
This document provides an overview of the EUDAQ software, the data
acquisition framework used by the EUDET JRA1 beam telescope. It describes
how to install and run the DAQ system and use many of the included utility
programs, and how users may integrate their systems into the EUDAQ
framework by writing their own Producer and DataConverterPlugin, thus
allowing them to take advantage of the EUTelescope analysis framework.
1
EUDAQ User Manual
Contents
Contents
1. Introduction
1.1. Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2. Directory Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4
4
5
2. Installing EUDAQ
2.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2. CMake . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3. C++11 compliant compiler . . . . . . . . . . . . . . . .
2.4. Downloading the source code . . . . . . . . . . . . . . .
2.5. Configuring via CMake . . . . . . . . . . . . . . . . . . .
2.6. Compilation on Linux/OSX . . . . . . . . . . . . . . . .
2.7. Setup and Compilation on Windows using Visual Studio
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
7
7
7
7
9
9
11
11
3. Running EUDAQ
3.1. Preparation . . . .
3.2. Processes . . . . .
3.3. Running the DAQ
3.4. Other Utilities . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4. Writing a Producer
4.1. Configuration . . . . . . . . . . . . . . . . .
4.2. Receiving Commands . . . . . . . . . . . . .
4.3. Sending Data and the RawDataEvent class .
4.4. Log Messages . . . . . . . . . . . . . . . . .
4.5. Interfacing Python-Code via the PyProducer
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
13
13
14
19
21
. . . . . .
. . . . . .
. . . . . .
. . . . . .
Interface
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
32
32
32
33
35
35
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5. Data Conversion
36
5.1. StandardEvent and StandardPlane . . . . . . . . . . . . . . . . . . . . . 36
5.2. LCIO and LCEvent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.3. DataConverterPlugin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
6. Other Parts of the Framework
6.1. FileWriter . . . . . . . . .
6.2. FileReader . . . . . . . . .
6.3. PluginManager . . . . . .
6.4. OptionParser . . . . . . .
6.5. Timer . . . . . . . . . . .
6.6. Utils . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
7. Reporting Issues
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
43
43
43
44
45
48
48
50
8. Developing and Contributing to EUDAQ
51
8.1. Regression Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
8.2. Commiting Code to the Main Repository . . . . . . . . . . . . . . . . . . 51
2
EUDAQ User Manual
A. Source Code
A.1. Example
A.2. Example
A.3. Example
A.4. Example
Contents
Config File . . . . .
Producer . . . . . .
DataConverterPlugin
Reader . . . . . . . .
B. Introduction to the build
B.1. MSBUILD . . . . . .
B.2. Project Files . . . . .
B.3. Known Problems . .
.
.
.
.
.
.
.
.
system and
. . . . . . .
. . . . . . .
. . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
53
53
54
58
61
project files on Windows
63
. . . . . . . . . . . . . . . . . . . . . . 63
. . . . . . . . . . . . . . . . . . . . . . 63
. . . . . . . . . . . . . . . . . . . . . . 66
C. Online Monitor Configuration Settings
C.1. Configuration Sections Overview . . . . . .
C.2. Configuration options in [General] . . . . .
C.3. Configuration options in [Correlations] . .
C.4. Configuration options in [Clusterizer] . . .
C.5. Configuration options in [HotPixelFinder]
C.6. Configuration options in [Mimosa26] . . .
C.7. Configuration Example . . . . . . . . . . .
Glossary
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
67
67
67
67
67
67
67
68
69
3
EUDAQ User Manual
1. Introduction
1. Introduction
The EUDAQ software is a data acquisition framework, written in C++, and designed to
be modular and portable, running on Linux, Mac OS X, and Windows. It was written
primarily to run the EUDET Pixel Telescope[1], but is designed to also be generally
useful for other systems.
The hardware-specific parts are kept separate from the rest, so that the common parts
can still be used independently. These include software for accessing the trigger logic
unit (TLU) and the the EUDET data reduction board (EUDRB) used by the EUDET
beam telescope.
The data files generated by the DAQ can be easily converted to the Linear Collider
I/O (LCIO) format, allowing the data to be analysed with the EUTelescope[2] analysis
package.
1.1. Architecture
It is split into a number of different processes (see Figure 1), each communicating using
TCP sockets. A central Run Control provides an interface for controlling the whole DAQ
system; other processes connect to the Run Control to receive commands and to report
their status.
Key:
Application
Thread
Buffer
1
Data Collector
Signal
/data
Listening
Socket
Command
Cmd
Receiver
Data
Storage
Data
Merge
Data
Sender
M
Monitor
Cmd
Receiver
Data
Receiver
GUI
Logging
N
Producer
Data
Server
N
M
Data
Sender
Hardware
Polling
Cmd
Receiver
Log
Server
Cmd
Server
1
Hardware
Run Control
Cmd
Receiver
Log
Receiver
Cmd
Sender
GUI
Storage
Figure 1: Schematic of the DAQ architecture.
4
1
Logger
GUI
EUDAQ User Manual
1. Introduction
Each piece of hardware that produces data (e.g. the TLU, the telescope, or a device
under test (DUT)) will have a Producer process. This will configure the hardware, read
out the data and send it to the Data Collector.
The Data Collector receives all the data streams from all the Producers, and combines
them into a single stream that is written to disk. It usually writes the data in a native
raw binary format, but it can be configured to write in other formats, such as LCIO.
The Logger receives log messages from all other processes, and displays them to the user,
as well as writing them all to file. This allows for easier debugging, since all log messages
are stored together in a central location.
A Monitor reads the data file and generates online-monitoring plots for display. In the
schematic it is shown to communicate with the DataCollector via a socket, but it actually
just reads the data file from disk (this may be changed in the future).
1.2. Directory Structure
The EUDAQ software is split into several parts that can each be compiled independently,
and are kept in separate subdirectories. The general structure is outlined below:
• main contains the main EUDAQ library with the parts that are common to most of
the software, and several command-line programs that depend only on this library.
All definitions in the library should be inside the eudaq namespace. It is organised
into the following subdirectories:
– lib/src contains the library source code,
– exe/src contains the (command line) executables source code,
– include contains the header files inside the eudaq subdirectory (to match the
namespace),
• gui contains the graphical programs that are built with Qt, such as the RunControl
and LogCollector.
• producers contains all (user-provided) producers shipped with the EUDAQ distribution, for example:
– tlu and eudrb contain the parts that depend on the TLU and EUDRB
respectively.
– vme contains a wrapper for the VME driver for the EUDRB.
– depfet, fortis, taki. . . contain the code for third-party producers that have
been used with the telescope.
• extern stores external software that is not part of EUDAQ itself, but that is needed
by EUDAQ in some cases, such as the ZestSC1 driver for the TLU and the Tsi148
VME driver.
• bin and lib contain the compiled binaries (executables and libraries) generated
from the other directories.
• conf contains configuration files for running the beam telescope.
5
EUDAQ User Manual
1. Introduction
• data and logs are directories for storing the data and log files generated while
running the DAQ.
• doc contains documentation, such as this manual.
Each directory containing code has its own src and include subdirectories, as well as a
local CMakeLists.txt file containing the rules for building that directory using CMake.
Header files usually have a .hh extension so that they can be automatically recognised
as C++ (as opposed to C), and source files have either .cc for parts of a library or .cxx
for executables.
6
EUDAQ User Manual
2. Installing EUDAQ
2. Installing EUDAQ
2.1. Overview
EUDAQ has relatively few dependencies on other software, but some features do rely
on other packages. To configure the EUDAQ build process, the CMake cross-platform,
open-source build system is used. To compile EUDAQ from source code requires a
compiler that implements the C++11 standard. The libusb library is only needed to
communicate over USB with a TLU[3]. The VME driver is only needed for reading out
EUDRBs[4] via VME with a Motorola MVME6100 single board computer. The other
dependencies are only needed for running the DAQ, and not for the common library (for
example if you only want to perform data analysis, or write a custom Producer to run in
the EUDET telescope, but not run the whole DAQ yourself).
2.2. CMake
In order to automatically generate configuration files for the build process of EUDAQ
both compiler and platform independent, the CMake build system is used.
CMake is available for all major operating systems from http://www.cmake.org/cmake/
resources/software.html. On most Linux distributions, it can usually be installed via
the built-in package manager (aptitude/apt-get/yum etc.) and on OSX using packages
provided by e.g. the MacPorts or Fink projects.
2.3. C++11 compliant compiler
The compilation of the EUDAQ source code requires a C++11 compliant compiler and
has been tested with GCC (at least version 4.6), Clang (at least version 3.1), and MSVC
(Visual Studio 2012 and later) on Linux, OS X and Windows.
If you are using Scientific Linux, please install the Developer Toolset available e.g. from
http://linux.web.cern.ch/linux/devtoolset/ to get access to a GCC version which
fully implements C++11.
2.3.1. libusb
In order to communicate with a TLU, the libusb library is needed. Therefore, if you want
to compile the TLU subdirectory, you should make sure that libusb is properly installed.
On Mac OS X, this can be installed using Fink or MacPorts. If using MacPorts you may
also need to install the libusb-compat package. On Linux it may already be installed,
otherwise you should use the built-in package manager to install it. Make sure to get the
development version, which may be named libusb-devel instead of simply libusb. On
Windows, libusb is only needed if compiling with cygwin, in which case you should use
the cygwin installer to install libusb. Otherwise libusb is not needed, as the included
ZestSC1 libraries should work as they are.
7
EUDAQ User Manual
2. Installing EUDAQ
2.3.2. ZestSC1 drivers and TLU firmware files
Additonally to the libusb library, the TLU producer requires the ZestSC1 driver package
and the FPGA firmware bitfiles; these are available to download via AFS from DESY.
If AFS is accessible on the machine when CMake is run, the necessary files will be
installed automatically; otherwise, manually copy full folder with sub-directories from
/afs/desy.de/group/telescopes/tlu/ZestSC1 and /afs/desy.de/group/telescopes/tlu/tlufir
into the ./extern subfolder in your EUDAQ source directory.
2.3.3. VME driver
In order to communicate with the EUDRB boards a VME library is needed. A kernel
module is included for the Tsi148 VME bridge, for use on a Motorola MVME6100, in
the extern/Tsi148 subdirectory. Installation of this module is beyond the scope of this
document.
The vme subdirectory includes code for accessing the VME bus with the Tsi148 module.
In principle other VME bridges could be used, you just need to write a C++ class that
inherits from the VMEInterface class and implements the necessary methods (look at
the TSI148Interface class for an example).
2.3.4. Qt
The graphical interface of EUDAQ uses the Qt graphical framework. In order to compile
the gui subdirectory, you must therefore have Qt installed. It is available in most Linux
distributions as the package qt4-devel, but make sure the version is at least 4.4, since
there are a few issues with earlier versions.
If the included version is too old, or on other platforms, it can be downloaded from
http://qt.nokia.com/downloads. Select the LGPL (free) version, then choose the
complete development environment (it may also work with just the framework, but this
is untested). Make sure the QTDIR environment variable is set to the Qt installation
directory, and the $QTDIR/bin directory is in your path.
If you are using OSX, the easiest way to install Qt is using the packages provided by the
MacPorts project (http://www.macports.org/).
2.3.5. Root
The online monitor, as well as a few command-line utilities (contained in the root
subdirectory), use the Root package for histogramming. It can be downloaded from
http://root.cern.ch or installed via your favorite package manager. Make sure Root’s
bin subdirectory is in your path, so that the root-config utility can be run. This can
be done by sourcing the thisroot.sh (or thisroot.ch for csh-like shells) script in the
bin directory of the Root installation:
source /path/to/root/bin/thisroot.sh
8
EUDAQ User Manual
2. Installing EUDAQ
2.3.6. LCIO / EUTelescope
To enable the writing of LCIO files, or the conversion of native files to LCIO format, eudaq
must be linked against the LCIO and EUTelescope libraries. Detailed instructions on
how to install both using the ilcinstall scripts can be found at http://eutelescope.
web.cern.ch/content/installation.
The EUTELESCOPE and LCIO environment variables should be set to the installation
directories of EUTelescope and LCIO respectively. This can be done by sourcing the
build env.sh script as follows:
source /path/to/Eutelescope/build_env.sh
2.4. Downloading the source code
The EUDAQ source code is hosted on github. The recommended way to obtain the
software is with git, since this will allow you to easily update to newer versions. The
latest version can be checked out with the following command:
git clone https://github.com/eudaq/eudaq.git eudaq
This will create the directory eudaq, and download the latest development version into
it. If you already have a copy installed, and want to update it to the latest version, you
do not need to clone the repository again, just change to the eudaq directory use the
command:
git pull
to update your local copy with all changes commited to the central repository.
Alternatively you can also download a zip file from https://github.com/eudaq/eudaq/
archive/master.zip.
For production environments (e.g. testbeams) we strongly recommend to use the latest
release version. Use the command git tag in the repository to find the newest version
and type e.g.
git checkout tags/v01-01-00
to change to version 1.1.0.
2.5. Configuring via CMake
CMake supports out-of-source configurations and builds – just enter the ’./build’ directory
and run CMake, i.e.
cd build
cmake ..
9
EUDAQ User Manual
2. Installing EUDAQ
CMake automatically searches for all required packages and verifies that all dependencies
are met using the CMakeLists.txt script in the main folder. By default, only the
central shared library, the main executables and (if Qt4 or Qt5 have been found) the
graphical user interface (GUI) are configured for compilation. You can modify this
default behavior by passing the BUILD [name] option to CMake where [name] refers to
an optional component, e.g.
cmake -D BUILD_gui=OFF -D BUILD_tlu=ON ..
to disable the GUI but enable additionally the TLU producer and executables.
The corresponding settings are cached, so that they will be again used next time CMake
is run.
Some of the optional packages and producers include:
main : The common library, and some command-line programs that depend on only this
library
tlu : The TLU library, and the command-line programs that depend on it. Requires
libusb, ZestSC1 drivers, and the TLU firmware files.
gui : The graphical parts of the DAQ, such as the Run Control and Log Collector.
Require Qt to be installed.
onlinemon : The Root Online Monitor. Requires Root to be installed.
nreader : The native reader Marlin processor used for data conversion into LCIO by
EUTelescope. Requires LCIO and EUTelescope to be installed.
manual This manual compiled from its LATEXsources. Requires a working LATEXinstallation.
The producers are stored in the ./producer subdirectory and include: altro, altroUSB,
depfet, eudrb, fortis, mimoroma, mvd, pixelmanproducer, and taki. These are usercontributed producers for specific detectors inside the EUDET telescope. They should
not be compiled unless needed.
A short description of selected producers:
producers/eudrb : The code for accessing EUDRB boards over VME. Depends on the
vme library which will be automatically built when eudrb is enabled. This should
only be compiled on an MVME6100 single-board computer, as it is only compatible
with the Tundra Tsi148 VME bridge, and PPC processors.
To install the binaries and the library outside the source tree, you need to set the
INSTALL PREFIX option, e.g.
cmake -D INSTALL_PREFIX=/usr/local ..
to install the executables into the bin and the library into lib subdirectories of
/usr/local.
If you ever need to, you can safely remove all files from the build folder as it only contains
automatically generated files. Just run
10
EUDAQ User Manual
2. Installing EUDAQ
cd build
rm -rf *
to start from scratch.
2.6. Compilation on Linux/OSX
You should just have to run the command:
make install
from the top EUDAQ directory to compile the common library, along with some commandline programs (the contents of the ./main/exe subdirectory). If other parts are needed,
you can specify them as arguments to the CMake command during the configuration
step.
The executable binaries and the common shared library will be installed by default
into the bin and lib directories in the source tree, respectively. If you would like to
install into a different location, please set the respective parameter during the CMake
configuration.
2.7. Setup and Compilation on Windows using Visual Studio
This section gives a short overview on the steps needed to compile the project under
Windows (tested under Windows 7, 32-bit). For a more detailed introduction to the
Windows build system and Visual Studio project files see the appendix B on page 63.
• Prerequisites:
– Download Qt4 or Qt5:
– Download and install the pthreads library (pre-build binary from ftp://
sources.redhat.com/pub/pthreads-win32) into either c:
pthreads-w32 or ./extern/pthreads-w32
– Download Visual Studio Express Desktop (e.g. 2013 Version): http://www.
microsoft.com/en-us/download/details.aspx?id=40787
• Start the Visual Studio Developer Command Prompt from the Start Menu entries for
Visual Studio (Tools subfolder) which opens a cmd.exe session with the necessary
environment variables already set. If your Qt installation has not been added to the
global %PATH% variable, you need to execute the qtenv2.bat batch file (or similar)
in the Qt folder, e.g.
C:\Qt\Qt5.1.1\5.1.1\msvc2012\bin\qtenv2.bat
Replace ”5.1.1” with the version string of your Qt installation.
11
EUDAQ User Manual
2. Installing EUDAQ
• Now clone the EUDAQ repository (or download using GitHub) and enter the build
directory on the prompt, e.g. by entering
cd c:\Users\[username]\Documents\GitHub\eudaq\build
• Configuration: Now enter
cmake ..
to generate the VS project files.
• Compile by calling
MSBUILD.exe EUDAQ.sln /p:Configuration=Release
or install into eudaq
bin by running
MSBUILD.exe INSTALL.vcxproj /p:Configuration=Release
• This will compile the main library and the GUI; for the remaining processors,
please check the individual documentation.
Note on “moc.exe - System Error: The program can’t start because MSVCP110.dll is
missing from your computer ” errors: when using Visual Express 2013 and pthreads-w32
2.9.1, you might require “Visual C++ Redistributable for Visual Studio 2012”: download
(either x86 or x64) from http://www.microsoft.com/en-us/download/details.aspx?
id=30679 and install.
12
EUDAQ User Manual
3. Running EUDAQ
3. Running EUDAQ
This section will describe running the DAQ system, mainly from the point of view of
the EUDET JRA1 Pixel Telescope[5] with a DUT, although most of it should also be
applicable to the DAQ in general, even without the telescope.
All executable programs from the different subdirectories are placed inside the bin
subdirectory, and should be run from here.
They should all accept a -h (or --help) command-line parameter, which will provide a
summary of the different command-line options that can be used.
3.1. Preparation
Some preparation is needed to make sure the environment is set up correctly and the
necessary TCP ports are not blocked before the DAQ can run properly.
3.1.1. Directories
The DAQ expects two directories to exist, that it will use to store data files and log files.
They need not be real directories – they can be symbolic links to other directories if you
don’t want to store the files inside the EUDAQ installation.
First, inside the eudaq directory, there should be a directory (or symbolic link) called
data. This will contain the data files written by the Data Collector, as well as a file
containing the last run number, so that it will continue incrementing even when the DAQ
is restarted.
Secondly, there should be a directory (or symbolic link) called logs. This will be used
by the Log Collector to store log files containing all the log messages received.
3.1.2. Firewall
The different processes communicate between themselves using TCP/IP sockets. If a
firewall is running, it may block these connections, especially if the processes are running
on different computers. If all the processes will be run from the same computer, then
it is probably not necessary to do anything. If a port is blocked, you will see an error
message similar to the following when attempting to start some programs:
Are you sure the server is running? - Error 61 connecting to ←localhost:44000: Connection refused
The ports used may be configured on the command line, but the default values used are:
44000 : This is the port used to send commands from the Run Control.
44001 : This port is used to send data from the producers to the Data Collector.
44002 : This port is used to send log messages from all processes to the Log Collector.
13
EUDAQ User Manual
3. Running EUDAQ
If processes will be run on different computers, then these ports should be opened up in
the firewall. The method for doing this depends on the Operating System used, and is
outside the scope of this manual.
3.1.3. Environment
When a process connects to the Run Control, it must be told what addresses to use
to connect to the Log Collector and (if it is a Producer) to the Data Collector. The
Run Control will ask the Log and Data Collectors what address to report, and these
processes therefore need a way to determine what address they are listening on. There
is no completely fool-proof way of determining this, so they look at the environment
variable $HOSTNAME.
Usually this should be the DNS name of the machine it is running on, but in some cases
it may not work correctly. If this is the case, it may be necessary to set this variable
manually, either to the real host name, or the machine’s IP address, or (if all the processes
will be run on the same computer) it can be set to localhost.
Depending on the command shell used, the command to do this should be either
“export HOSTNAME=name” (for bash-like shells) or “setenv HOSTNAME name” (for csh-like
shells), where name is the name to use.
3.1.4. TLU permissions
If you are not using a TLU, or not running on Linux, you may skip this part.
On many Linux distributions, the device node used to communicate over the USB bus is
only accessible by the user root by default.
To have the system set the correct permissions when a TLU is connected, you need to
add a udev rule: as root user, create the file /etc/udev/rules.d/54-tlu.rules and
add the following lines:
# for Red Hat, e.g. SL5
SYSFS{idVendor}=="165d", SYSFS{idProduct}=="0001", GROUP="NOROOTUSB", ←MODE="0666"
if you are using a Red Hat-based distribution (such as Scientific Linux) or:
# for Debian
ACTION=="add", DRIVERS=="?*", ATTR{idVendor}=="165d", ←ATTR{idProduct}=="0001", MODE="0666"
in case you are using a debian-based distribution such as Ubuntu.
After replugging the TLU, the device should be accessible by all users.
3.2. Processes
The DAQ system is made up of a number of different processes that may all be run on
the same, or on different computers. They are each described below.
14
EUDAQ User Manual
3. Running EUDAQ
3.2.1. Run Control
There are two versions of the Run Control – a text-based version, and a graphical version
(see Figure 2). The graphical version is preferred, since it is the most used, and therefore
the most tested and complete. The executable is called euRun.exe, or on Mac OS X
it is an application bundle called euRun.app. The text-based version can be useful for
testing, the executable is TestRunControl.exe.
Figure 2: The Run Control graphical user interface.
Normally no command-line options should be needed, but it can be told to listen on a
non-standard port, (e.g. to run two copies on the same machine), with the -a hporti
option:
./euRun.app/Contents/MacOS/euRun -a 3000
This example is for Mac OS X, where the executable is inside an application bundle,
on other architectures it will be just euRun.exe. Note also that it is not recommended
to run two copies of the DAQ simultaneously, since it becomes difficult to keep them
completely separate as the Log and Data Collectors must also be run on different ports.
3.2.2. Log Collector
Running the Log Collector is optional. If it is run, then all log messages generated by all
other processes in the DAQ will be collected in one central location.
Like the Run Control, there are also two versions of the Log Collector. The graphical
version is called euLog.exe, or euLog.app on Mac OS X, and the text-based version is
called TestLogCollector.exe.
15
EUDAQ User Manual
3. Running EUDAQ
Figure 3: The Log Collector graphical user interface.
If it is being run on the same machine as the Run Control, it should not need any
command-line options. However, if it is run on a different machine, it must be told on
which machine the Run Control is running, using the -r hhostnamei option, e.g.:
./euLog.exe -r eudetmac001.cern.ch
It may also be told to listen on a non-standard port, using the -a hporti option, similar
to the Run Control.
3.2.3. Data Collector
The Data Collector is the process that collects all the raw data from the Producers,
merges all the connected incoming streams into a single data stream, and writes it to file.
Like the Log Collector, it should be told where to connect to the Run Control if it is not
running on the same machine, and it may also be told to listen on a non-standard port,
with the -r and -a options respectively, for example:
./TestDataCollector.exe -r eudet -a tcp://55001
It is also possible to run multiple Data Collector instances within one EUDAQ session.
This can be useful to reduce network traffic and e.g. write the output of one producer
to a locally attached disk. When running several Data Collectors simultaneously, Run
Controls assigns a Producer to a Data Collector by name: if the name of a Data Collector
matches that of a Producer, the latter will be given the address and port of the former.
There can be only one instance of an unnamed Data Collector which serves as the default
for any non-matching Producer; if no unnamed Data Collector is present, the first one
connecting will serve as the default.
The name of a Data Collector can be set with the -n option, for example:
./TestDataCollector.exe -n myproducer
Should you wish to run several instances of the Data Collector on one machine, you need
to make sure that they listen to different addresses using the -a option as described
above. Furthermore, you need to make each Data Collector write to a different file by
16
EUDAQ User Manual
3. Running EUDAQ
including the FilePattern option in the corresponding section of your configuration file
(also see section 3.3.3):
[DataCollector.myproducer]
FilePattern = "../data/run$6R_myproducer$X"
3.2.4. TestProducer
For testing purposes, you may use the Test Producer. This works similarly to a real
producer, but does not talk to any real hardware, instead providing a menu for the user
to manually send events (or see the ExampleProducer, below).
3.2.5. ExampleProducer
The ExampleProducer was written to illustrate the writing of a new Producer (see
section 4). However, it will actually generate some example data, and so can also be used
for testing purposes. It works more like a real Producer than the TestProducer, in that
it does not require user intervention to generate each trigger, and the data generated
emulates a simple (but realistic) sensor, and can be properly converted, and therefore
displayed in the Monitor.
3.2.6. TLUProducer
If you do not have a TLU in your setup, you may skip this part. Otherwise you should
run a TLUProducer, which will configure the TLU, and read out the timestamps and
send them to the Data Collector.
On the computer with the TLU connected, start the TLUProducer.exe program. If this
is not the same machine as the Run Control, use the -r option as for the Data and Log
Collectors. For example:
./TLUProducer.exe -r eudet.unige.ch:3000
If the TLUProducer fails to start, make sure the permissions are set up correctly (see
subsubsection 3.1.4).
3.2.7. EUDRBProducer
The EUDRB boards are used to read out the telescope sensors. The EUDRB Producer
is designed to run on a Motorola MVME6100 single board computer, using the Tundra
TSI148 VME bridge for communication with the EUDRBs.
If more than one EUDRBProducer is to be run, they must all have different names. The
name can be set with the -n hnamei option.
As with the other processes, the address of the Run Control should be set with the -r
option. An example is shown below:
./EUDRBProducer.exe -n EUDRB2 -r 192.168.1.1
17
EUDAQ User Manual
3. Running EUDAQ
3.2.8. Other Producer(s)
If you have a producer for your own hardware (see section 4), it should also have an
option to set the address of the Run Control.
3.2.9. OnlineMon
The OnlineMon reads the data file written by the Data Collector, and generates several
Root histograms that can be useful for online monitoring. Since it reads the native data
file directly, it must be run on the same machine as the Data Collector.
Figure 4: The OnlineMon showing correlation plots between different Mimosa26 planes
of the EUDET telescope.
The OnlineMon can be run in one of two modes: online or offline. In online mode, it
connects to the RunControl, so it will know when new runs are started, and it will
automatically open each new data file as it is created. In offline mode, there is no
RunControl, and it only analyses the data file it is given on the command line. An
example command line is:
./OnlineMon.exe -f 5432
This will run it in offline mode, opening the file corresponding to run 5432 (alternatively,
the full path to a file may be given). To run it in online mode, simply omit the -f option,
18
EUDAQ User Manual
3. Running EUDAQ
then the -r option may be used if the RunControl is running on a different computer or
using a non-standard port.
3.2.10. Python Interface and Wrapper for Core EUDAQ Components
A Python interface is provided for selected EUDAQ components: RunControl, DataCollector and a Producer, that can be extended on the Python side. The interface is realized
through the ctypes package that is part of every standard Python installation and
requires the numpy Python package to be installed. The interface code for all components
is located in the main/python directory.
To use the interface and access the components as Python objects, the wrapper must be
loaded inside your Python script:
#!/usr/bin/env python2
execfile('PyEUDAQWrapper.py') # load ctypes wrapper
prc = PyRunControl() # start run control with default settings
# wait for more than one active connection to appear
while prc.NumConnections < 2:
sleep(1)
prc.Configure("ExampleConfig") # load configuration file
while not prc.AllOk:
sleep(1) # sleep while waiting for all connected producers
prc.StartRun()
This little scripts creates a RunControl instance, sends a configuration to all connected
producers, waits for their reply, and starts a new run. Several more extensive examples
for using Python with EUDAQ are located in the python directory in the main EUDAQ
directory.
3.3. Running the DAQ
To start the DAQ, all the necessary processes must be started in the correct order. The
first process must be the Run Control, since all other processes will attempt to connect
to it when they start up. Then it is recommended to start the Log Collector, since any
log messages it receives may be useful to help with debugging in case everything does not
start as expected. Next, the Data Collector should be started. Finally all the Producers,
and if needed, the RootMonitor.
3.3.1. STARTRUN
The STARTRUN file, in the main eudaq directory (as opposed to the bin subdirectory where
the executables exist), is a shell script that can be customized to load the appropriate
processes for running the DAQ. This allows you to start all the processes necessary with a
single command. If starting processes on other computers via SSH, it is recommended to
set up SSH keys so that the processes may be started without having to type a password.
19
EUDAQ User Manual
3. Running EUDAQ
In the future the STARTRUN script may be replaced with a more intelligent version that
uses a configuration file generated by the config script to decide what to load.
3.3.2. Controlling the DAQ
Once all the processes have been started, the DAQ can be configured, and runs may be
started and stopped using the Run Control (see Figure 2).
First the appropriate configuration should be selected from the drop-down list (see
subsubsection 3.3.3 for creating and editing configurations), and the GeoID should be
verified (see subsubsection 3.3.4), before continuing.
Then the Config button can be pressed, which will send a configuration command (with
the contents of the selected configuration file) to all connected processes. The full contents
of the configuration file will also be stored in the beginning-of-run-event (BORE) of the
data file, so that this information is always available along with the data.
Once all connected processes are fully configured, a run may be started, by pressing
the Start button. Whatever text is in the corresponding text box when the button is
pressed will be stored as a comment in the data file. This can be used to help identify
the different runs later.
Once a run is completed, it may be stopped by pressing the Stop button. Runs will also
stop and restart automatically when the data file reaches a threshold in size (by default
this is 1 GB). This is because there is a file size limit of 2 GB for storage on the GRID,
and the processed files can grow bigger than the original native files. The threshold size
for restarting a run may be configured in the config file (see subsubsection 3.3.3).
At any point a message may be sent to the log file by filling in the Log text box and
pressing the corresponding button. The text should appear in the LogCollector window,
and will be stored in the log file for later access.
Once the run is stopped, the system may be reconfigured with a different configuration,
or another run may be started.
3.3.3. Config Files
The Config drop-down in the Run Control is populated from the files in the config
subdirectory. These are just text files in a specific format, containing name-value pairs
separated into different sections. See subsection A.1 for an example file.
Any text from a # character until the end of the line is treated as a comment, and
ignored. Each section in the config file is delimited by a name in square brackets (e.g.
[RunControl]). The name represents the type of process to which it applies; if there
are several such processes, then they can be differentiated by including the name after
a period (e.g. [Producer.Example]). Within each section, any number of parameters
may be specified, in the form Name = Value. It is then up to the individual processes
how these parameters are interpreted.
The entire contents of the config file will be sent to all processes during the configuration,
and each process will have the appropriate section selected. The file will also be attached
20
EUDAQ User Manual
3. Running EUDAQ
to the BORE, so that it is available with the data later, even if the original config file is
modified or deleted.
3.3.4. GeoID
The GeoID is a number representing the physical positioning of the telescope and DUT(s).
Each time a change is made to the telescope layout, this number should be incremented.
To change the number, double-click on it, and a window will appear with the new value.
By default it will increment the old value by one, so normally you should just click OK,
but if necessary you may edit the value first.
The GeoID is inserted into the config file when it is sent, so it is also stored in the data
file, and will be used to select the correct GEAR file for alignment during the data
analysis stage.
3.4. Other Utilities
There are a number of other utilities available that are not needed for running the DAQ,
but can be useful for other tasks such as debugging. The executables are all located in
the bin subdirectory. They should all accept a help (-h or --help) option, to print a
summary of the available options.
3.4.1. TLUControl
The TLUControl.exe program is a standalone program for running the TLU without
using the full DAQ. The most commonly used parameters are shown below. For each
option, the short (preceeded by one dash) and the long (preceeded by two dashes) option
names are shown (only one of the two forms should be used for each option, but long and
short options can be mixed together on the command line), along with any parameters
and their default value that will be used if the option is not specified.
-d --dutmask hmask = 0i : The DUT mask; this defines which DUT connections are
activated. It is a bit-mask, so 1 means connector 0, 2 means connector 1, etc..
-a --andmask hmask = 255i : The AND mask; this defines which external trigger inputs
are activated. It is a bit-mask, so 1 means channel 0, 2 means channel 1, etc.. The
specified channels are ANDed together, and used to generate a trigger signal.
-t --trigger hmsecs = 0i : Internal trigger period. If non-zero, the TLU will generate
internal triggers with the specified period in milliseconds. If set to zero, the internal
trigger is off.
-i --dutinputs hvalues = ""i : Input mode select. A sequence of comma-separated
strings specifying which connectors to use for the DUT inputs. Valid values are
RJ45, LEMO, HDMI, and NONE.
21
EUDAQ User Manual
3. Running EUDAQ
-u --wait-for-user : Pause the program after the TLU is configured, before starting
triggers. The default is to not wait for the user.
Other parameters available are as follows:
-o --ormask hmask = 0i : The OR mask; this defines which external trigger inputs are
activated. It is a bit-mask, so 1 means channel 0, 2 means channel 1, etc.. The
specified channels are ORed together, and used to generate a trigger signal.
-v --vetomask hmask = 0i : The VETO mask; this defines which external trigger inputs
are activated. It is a bit-mask, so 1 means channel 0, 2 means channel 1, etc.. The
specified channels are used to veto the generation of a trigger if they are active.
-w --wait hms = 1000i : Wait time. This is the time to wait between updates.
-n --notimestamp : Indicates that the timestamp buffer should not be read out.
-q --quit : Quit the program after configuring the TLU.
-s --save-file hfilename = ""i : The filename to save trigger numbers and timestamps
-p --strobeperiod hcycles = 1000i : Period for timing strobe (in TLU clock cycles).
-l --strobelength hcycles = 100i : Length of ‘on’ time for timing strobe (in TLU
clock cycles).
-b --dutveto hmask = 0i : Mask for enabling veto of triggers (‘backpressure’) by rasing
DUT CLK.
-hm --handshakemode hnohandshake = 0i : In this mode the TLU issues a fixed-length
pulse on the trigger line (0 = no handshake).
-pw --powervctrl hmV = 800i : [obsolete but provided for backward compatibility,
please use -pv] Sets the Vcntl control voltage to all PMTs. The range of values is between 0 and 1000 (or 0 and 2000 if the TLU has been modified by cutting
LC1 and jumpering LO1 on the PMT Supply Daughterboard and specifying the -pm
1 option).
-pv --pmtvcntl hmV = 800i : Sets the Vcntl control voltage to all PMTs (see option
-pw for more details). Will override the value of -pw if it is specified. If neither -pw
or -pv is specified, the default value will be used (and can be overridden on an
individual PMT basis).
-p1 --pmtvcntl1 hmVi : Sets the PMT Vcntl voltage for PMT1 (Chan 0) only. If not
specified, the default or values specified by -pw or -pv (which will override -pw) is
used.
22
EUDAQ User Manual
3. Running EUDAQ
-p2 --pmtvcntl2 hmVi : Sets the PMT Vcntl voltage for PMT2 (Chan 1) only. If not
specified, the default or values specified by -pw or -pv (which will override -pw) is
used.
-p3 --pmtvcntl3 hmVi : Sets the PMT Vcntl voltage for PMT3 (Chan 2) only. If not
specified, the default or values specified by -pw or -pv (which will override -pw) is
used.
-p4 --pmtvcntl4 hmVi : Sets the PMT Vcntl voltage for PMT4 (Chan 3) only. If not
specified, the default or values specified by -pw or -pv (which will override -pw) is
used.
-pm --pmtvcntlmod hvalue = 0i : Specifies whether the TLU PMT Supply Daughtercard is modified (LC1 cut and LO1 jumpered) or not. A hvaluei of 0 specifies that
it is unmodified (and thus the Vcntl range is from 0mV to 1000mV), and a hvaluei
of 1 specifies that the TLU is modified (and thus the Vcntl range is from 0mV
to 2000mV). This feature is to accomodate newer Hamamatsu PMT models (e.g.
H10721) that require a control voltage range of, for instance, 500mV to 1100mV
that are being used in place of the older (discontinued, but what the TLU was
designed to accomodate and control) models that required a control voltage of
between 250mV and 900mV.
-f --bitfile hfilename = ""i : The bitfile containing the TLU firmware to be loaded.
-e --error-handler hvalue = 2i : Error handler setting. Setting to 0 indicates the
program should abort on an error. Setting it to a value greater than 0 indicates
the number of tries that should be attempted before generating an exception.
-r --fwversion hvalue = 0i : Specifies the firmware version to load (setting to 0
indicates the version should be chosen automatically).
-z --trace-file hfilename = ""i : The filename to save a trace of all USB accesses.
Prepend a dash (‘-’) to output errors only, or a plus (‘+’) for all data (including
block transfers).
An example use of the command is shown below:
./TLUControl.exe -t 200 -d 3 -i LEMO,RJ45 -u
Using options:
TLU version = 0 (auto)
Bit file name = '' (auto)
Trigger interval = 200 ms (5 Hz)
DUT Mask = 0x03 (3)
Veto Mask = 0x00 (0)
And Mask = 0xff (255)
Or Mask
= 0x00 (0)
DUT inputs = LEMO,RJ45
23
EUDAQ User Manual
3. Running EUDAQ
Strobe period = 0x0003e8 (1000)
Strobe length = 0x000064 (100)
Enable DUT Veto = 0x00 (0)
Save file = '' (none)
TLU Version = v0.2c
TLU Serial number = 0x062b (1579)
Firmware file = TLU2_Toplevel.bit
Firmware version = 65
Library version = 65
Press enter to start triggers.
TLU Started!
Status:
20,00,--,--,--,-- (0,0)
Scalers:
0, 0, 0, 0
Particles: 2
Triggers: 0
Entries:
0
TS errors: 0, 0 (redundancy, re-read)
Timestamp: 0x8d768 (579432) = 0.00150891
Time: 0.009 s, Freq: 0 Hz, Average: 0 Hz
0,
1,
2,
3,
4,
0x27fb479 (41923705) = 0.109174, diff=41923705
0x7139ab9 (118725305) = 0.309174, diff=76801600
0xba780f9 (195526905) = 0.509174, diff=76801600
0x103b6739 (272328505) = 0.709174, diff=76801600
0x14cf4d79 (349130105) = 0.909174, diff=76801600
Status:
20,00,--,--,--,-- (0,1)
Scalers:
0, 0, 0, 0
Particles: 7
Triggers: 5
Entries:
5
TS errors: 0, 0 (redundancy, re-read)
Timestamp: 0x1726fa48 (388430408) = 1.01152
Time: 1.023 s, Freq: 4.92913 Hz, Average: 4.88442 Hz
5,
6,
7,
8,
9,
0x196333b9 (425931705) =
0x1df719f9 (502733305) =
0x228b0039 (579534905) =
0x271ee679 (656336505) =
0x2bb2ccb9 (733138105) =
Status:
20,00,--,--,--,-- (0,1)
Scalers:
0, 0, 0, 0
Particles: 12
1.10917,
1.30917,
1.50917,
1.70917,
1.90917,
24
diff=76801600
diff=76801600
diff=76801600
diff=76801600
diff=76801600
EUDAQ User Manual
3. Running EUDAQ
Triggers: 10
Entries:
5
TS errors: 0, 0 (redundancy, re-read)
Timestamp: 0x2e5bb708 (777762568) = 2.02538
Time: 2.037 s, Freq: 4.93259 Hz, Average: 4.90838 Hz
^CQuitting...
This sets up internal triggers at 5 Hz (200 ms period), and activates DUT inputs 0 and 1.
Input 0 is configured to use the LEMO connector, and input 1 to use the RJ45 connector.
The first part of the output just summarizes the input parameters. The next part shows
information about the version numbers of the TLU and the firmware.
It will then configure the TLU, and if the -u option is used, it will wait for the user to
press enter before continuing. The triggers are then enabled, and a summary of the status
is printed out periodically (by default every 1 second). The program can be stopped
cleanly by pressing Ctrl-C.
Each block of status output consists of:
• a list of triggers, if there were any since the last update (the first time there are
none), each showing:
– the trigger number,
– the timestamp of the trigger, in hex, decimal and converted to seconds,
– the difference since the last trigger.
• the status of the DUT connections (see below),
• the values of the scalers on the external trigger inputs,
• the number of “particles”, which means all the potential triggers (including those
that were vetoed),
• the number of triggers that actually got sent to the DUTs,
• the number of entries in the trigger buffer, this should be equal to the number of
triggers printed out at the top of the status block,
• the number of timestamp errors detected by redundancy, and by re-reading,
• the current timestamp value,
• the time since the run started, the current trigger frequency, and the average
frequency over the whole run.
In the example output this block is repeated three times, before Ctrl-C is pressed to
stop it. The status is of the DUT connections formatted as:
• two digits for each DUT connection consisting of:
– two hyphens (--) if the connection is inactive, else
– the first digit represents the inputs from the DUT; with the busy line in bit
0 and the clock line in bit 1 (note the clock input can float low or high if a
LEMO input is selected, as it is not connected),
25
EUDAQ User Manual
3. Running EUDAQ
– the second digit represents the state of the FSM, as defined in the TLU
manual[3] (0 is ready, 1 is waiting for busy high, 4 is waiting for busy low, 5
is DUT-initiated veto, and F is an error condition).
• then in parentheses:
– the veto state (software veto in bit 0, overall veto in bit 1),
– the DMA state (1 when a DMA transfer is taking place).
3.4.2. VMETest
The VMETest.exe program uses the EUDAQ VME library to perform VME accesses. It
can be useful for determining whether a VME card is responding at a particular address.
The available options are:
-b haddressi : The base address for the VME accesses. This value will be added to the
offsets specified in the commands to give the actual address used.
-s hbytesi : Sets the window size in bytes. This is the amount of memory that is
mapped into the VME address space. Any accesses outside this range will result in
an access violation.
-a hbitsi : The address bus width in bits. Valid values are 16, 24, 32 or 64.
-d hbitsi : The data bus width in bits. Valid values are 8, 16, 32 or 64.
-m hmodei : The VME access mode. Valid values are S (single accesses), B (BLT), M
(MBLT), 2 (2eVME), E (2eSST) and T (2eSSTB).
The options set up the mode for the VME accesses. Following the options, a number of
commands can be specified to perform actual reads or writes. The commands can be any
of the following:
rhoffseti : Reads a value from the specified offset, and displays the value read.
Rhoffseti,hwordsi : Performs a block read of the specified number of words, starting
from the specified offset.
whoffseti,hvaluei : Writes the specified value to the specified offset.
Whoffseti,hvalue1i[,hvalue2i...] : Performs a block write of the specified values,
starting at the specified offset.
Numerical arguments to either the options or the commands can be given either in
decimal, or in hexadecimal by prefixing them with 0x, as in C or C++. Note that the
options require a space between the option character and its argument, but the commands
must not have a space. For example:
./VMETest.exe -b 0x180000 -a 24 -d 16 w0x20,123 r0x10
This sets up a window starting at 180000 hex, in A24 address space with D16. It then
writes the value 123 to offset 32 (20 hex), and then reads the value at offset 16 (10 hex).
26
EUDAQ User Manual
3. Running EUDAQ
3.4.3. TestReader
The TestReader.exe program will read a native data file, and can display various pieces
of information from the file. Commonly used options are:
-b : Display the BORE.
-e : Display the end-of-run-event (EORE).
-d hrangei : Display the specified range of event numbers.
-p : Process the displayed events and display the corresponding StandardEvents.
-u : Dump the raw data for the displayed events.
-s : Try to resynchronize events based on the TLU event number. A full description of
this option is outside the scope of this manual (but if you don’t know what it is,
you probably don’t need it).
After the options a list of one or more filenames can be given. Any filenames that consist
only of numerical digits will be interpreted according to the input pattern (by default
this is “../data/run$6R.raw”, where $6R will be replaced with the run number padded
to 6 digits). For example:
./TestReader.exe -b -e -p -d 1-10,100,1000 example.raw 5432
This will display the BORE and EORE, and the events 1 to 10, 100 and 1000, processing them to also display the StandardEvents, from the files example.raw and
../data/run005432.raw.
3.4.4. Converter
The Converter.exe program will read a native data file, optionally select just a subset
of events from the file, and can then write it out to another file in either the same native
format, or a different format. The most commonly used options are:
-t htypei : The file type to write out. The available types are listed below.
-e hrangei : Select the specified range of event numbers.
-s : Try to resynchronize events based on the TLU event number (see TestReader in
subsubsection 3.4.3).
The available output file types are as follows:
native : The native EUDAQ binary file format, consisting of a serialised stream of
DetectorEvents, containing the raw data read out from the hardware.
27
EUDAQ User Manual
3. Running EUDAQ
standard : Like the native format, this is also a serialised stream, but in this case
it contains StandardEvents, in which the raw data has been converted into a
standard format.
lcio : The standard LCIO file format used by the analysis software. This type is only
available if EUDAQ was compiled with LCIO support.
root : A Root file containing a TTree with the hit pixel information.
text : A simple text based format (not yet implemented).
mimoloop : A text based format mimicking the output of the mimoloop program (from
Angelo Cotta Ramusino and Lorenzo Chiarelli at INFN Ferrara).
Although this program can be used to convert a native data file into LCIO format, the
more usual (and therefore better tested) way is to use the EUTelescope converter.
3.4.5. ClusterExtractor
This program can be used to quickly extract some clusters from raw data. It is not as
sophisticated as the EUTelescope package, which should be preferred for real analysis,
but it can be useful for doing quick checks. It will read a native data file, perform a
basic clustering, and then write these clusters to one text file per sensor plane. The most
commonly used options are:
-p hpixelsi : The cluster size in pixels. It should be an odd number, with 1 meaning
no clustering (just pixels over threshold), 3 meaning 3×3 pixel clusters, etc.
-n hadcsi : The noise level (sigma) in ADC units. This is used to scale the thresholds
in terms of the noise.
-s hthreshi : The threshold for seed pixels, in terms of the noise.
-c hthreshi : The threshold for the total charge of a cluster, in terms of the cumulative
noise of all the pixels in the cluster.
-w : Reports the cluster centre as the weighted average of the pixels, instead of the
position of the seed pixel.
An example use is:
./ClusterExtractor.exe -p 3 -n 3.5 -s 6 -c 10 -w 5432
This will generate a number of text files named runNNN eutel M.txt, where NNN is the
run number, and M is the sensor plane number. The format of the output text files is as
follows:
28
EUDAQ User Manual
2
182
241
3
111
5
113
7
252
95
2
153
120
1
67
1
141
2
240
170
3. Running EUDAQ
51487659237
126
125
51489095892
346
51491334074
171
51495330212
305
189
The first line contains the event number, the number of clusters, and the TLU timestamp.
Then for each cluster there is one line, containing the x and y coordinates of the cluster
centre, and the total charge in ADC units. The cluster lines are prepended with a space
to make it easier to scan the file by eye.
3.4.6. MagicLogBook
This program is designed to extract as much information as possible from data files and
log files, in order to reconstruct a log book. Despite its name, it is in fact not magical,
so it is preferable to keep a good log book during running, rather than relying on this
program to generate it later.
The available options are listed below:
-f hfieldsi : A list of fields to include in the output, in the form name=value, with
multiple fields separated by commas. If a predefined list is also specified these will
be appended to the list.
-s hseparatori : The separator to use between fields in the output. The default is a
tab character.
-h hstringi : A string that appears at the beginning of the header line (with the list of
field names), that can be used to differentiate it from the other lines. The default
is an empty string.
-p name : Use a predefined list of fields. Currently available values are normal and full.
-o hfilei : The output filename. By default the standard output is used.
The easiest method of running is to use a predefined list of fields. There are currently
two predefined lists available: normal and full. If neither of these are suitable, contact
the EUDAQ maintainer, as it may be possible to add more options.
The normal list includes:
• the run number,
• the config file name,
29
EUDAQ User Manual
3. Running EUDAQ
• the run start time,
• for the EUDRBs:
–
–
–
–
–
the mode,
the sensor type,
whether they are running unsynchronized,
the number of boards,
and the firmware version.
• and for the TLU:
–
–
–
–
the internal trigger interval,
the AND mask,
the DUT mask,
and the firmware version.
The full list includes all the values from the normal list, plus the number of events in
the run and the end of run time. This is because these values can only be known by
reading the whole data file to the end, which is slow, especially for large data files.
If necessary, other information is available using custom fields, although the syntax for
these is a bit complicated, since it is designed to be as flexible as possible at specifying
any information in the data file. In the future it may be redefined in order to simplify it
if possible. Therefore it is recommended to use a predefined list of fields where possible.
Custom fields are specified as a comma separated list of items in the form name=value,
with the name being what will appear on the header line of the output, and the value
specifying what exactly to extract from the file. The possible values are illustrated below,
although not exhaustively:
events∗ : The number of events in the run.
config : The configuration name, or:
config:section:key : The value of the key from the corresponding section in
the config (e.g. config:Producer.EUDRB:NumBoards).
bore, tlu, eudrb, eore∗ : Something from the BORE, the TLUEvent or EUDRBEvent
subevents of the BORE, or the EORE, respectively:
bore:.Run : The run number
bore:hnamei : Otherwise, if the second part does not start with a period, the value
of the tag hnamei is used (e.g. tlu:DutMask or eudrb:MODE).
log : Something from the log file (not implemented yet).
∗
items marked with an asterisk require reading the whole data file, and are therefore
slow, especially when large data files are involved.
Note that the EUDRBEvent is now deprecated, having been replaced by the RawDataEvent,
but there is currently no way to specify this.
The MagicLogBook command is used as follows:
30
EUDAQ User Manual
3. Running EUDAQ
./MagicLogBook.exe -p normal ../data/*.raw
This will produce an output similar to the following:
Run
6371
6372
6373
6374
6375
6376
6379
Config
Mode Det Start
U P Trg AND
eudet-beam
2009-07-29 07:44:39.535 1 6
0 0xf
eudet-beam
2009-07-29 08:03:05.079 1 6
0 0xf
eudet-m26test
2009-07-30 09:57:45.157 1 6 255 0xff
eudet-m26test
2009-07-30 10:00:45.205 1 6 255 0xff
eudet-m26test
2009-07-30 10:05:38.625 1 6
1 0xff
eudet-m26test
2009-07-30 10:10:00.107 1 6
1 0xff
eudet-m26test
2009-07-30 10:13:05.322 1 6
1 0xff
DUT
0x10
0x10
0x12
0x12
0x12
0x12
0x12
Tfw Efw
241
241
241
241
241
241
241
Note that the header row has been modified slightly to fit into the page width: the U
should be UnSync, P should be Planes, Trg should be TriggerInterval, Tfw should be
TLUfw, and Efw should be EUDRBfw. The columns Mode, Det and EUDRBfw are missing
from the output due to the fact that this information is now stored in a RawDataEvent,
which is not currently accessible with this version of the program.
3.4.7. FileChecker
This is a small utility that reads raw data files and checks if all events are readable,
can be syncronised using the TLU trigger id and lists which type of subevents the file
contains.
It should be called with list of file paths or run numbers. For any argument that consist
only of numerical digits the file path is constructed by substituting $6R in the input
pattern (defaults to “../data/run$6R.raw”) with the run number padded to 6 digits.
For example:
./FileChecker.exe {6045..6050}
This would produce the following output.
run
-----6045
6046
6047
6048
6049
6050
valid
----true
true
true
true
false
false
num_events
---------13131
1
14674
7776
0
-1
contains
------------------------MUPIX4,NI,TLU
MUPIX4,NI,TLU
MUPIX4,NI,TLU
MUPIX4,NI,TLU
errors
----------------------
no events in the file.
read error.
3.4.8. Others
Some programs that are less used (or recently added) may not be described here. If they
look interesting, you can find out more about them by running them with the help (-h
or --help) option, or by examining the source code.
31
EUDAQ User Manual
4. Writing a Producer
4. Writing a Producer
In order to integrate a DUT fully into the DAQ, it needs its own Producer. A Producer
is both a CommandReceiver and a DataSender, meaning it receives commands from
Run Control, and it also sends events to the Data Collector. A base class is provided
that users may inherit from, to make this as easy as possible. For example code, see
subsection A.2.
4.1. Configuration
The Configuration class is a way of storing configuration information in a way that
is easily accessible, and can be saved to or loaded from a human-readable file (see
subsubsection 3.3.3), and can be sent over the network. It is defined in the following
header:
#include "eudaq/Configuration.hh"
The configuration consists of a number of sections, each of which contains a list of
name-value pairs. The values are stored as strings, but they can be converted to/from
arbitrary types. Methods are provided to load from or save to file, to set the current
section, and to set or get configuration values. An example use is shown below:
std::ifstream infile("../conf/ExampleConfig.conf");
eudaq::Configuration config(infile, "Producer.Example");
int param = config.Get("Parameter", 0);
std::cout << "Loaded config, param = " << param << std::endl;
config.Set("Parameter", param+1);
config.Set("OtherParam", "something");
std::ofstream outfile("Test.conf");
config.Save(outfile);
This creates a configuration loaded from the file ../conf/ExampleConfig.conf, selecting
the Producer.Example section. It then gets an integer parameter from the configuration
and displays it. Then it modifies the value of the parameter and sets another parameter,
before writing the configuration to the file Test.conf.
A configuration object will be received by the Producer during the configuration, as
described in subsubsection 4.2.1.
4.2. Receiving Commands
Whenever a command is received from the Run Control, a corresponding member function
of the Producer will be called by the code in the base classes. In order to react to a
command, the necessary code is simply put inside the corresponding method. The
Producer base class is declared by including the following header file:
#include "eudaq/Producer.hh"
32
EUDAQ User Manual
4. Writing a Producer
4.2.1. OnConfigure
This method is called whenever a configure command is received from the Run Control.
The method signature is:
virtual void OnConfigure(const eudaq::Configuration & config);
As a parameter, it receives the configuration chosen in the Run Control. Information
may be extracted from the configuration in order to set up the hardware.
4.2.2. OnStartRun
This is called on the start of each run. The method signature is:
virtual void OnStartRun(unsigned param);
As a parameter, it receives the run number of the started run. The Producer must send
a BORE, and then prepare for reading out events from the hardware.
4.2.3. OnStopRun
This is called at the end of the run. The method signature is simply:
virtual void OnStopRun();
Care should be taken that there are no more events pending to be read out. Once all
data events have been sent, an EORE should also be sent, to signal to the DAQ that the
Producer has ended the run successfully.
4.3. Sending Data and the RawDataEvent class
Events may be sent to the DAQ using the Producer’s SendEvent() method that has
the following signature:
void SendEvent(const Event &);
It takes as a parameter an object derived from the eudaq::Event base class that will be
serialised and sent to the Data Collector. In practice it will usually be of concrete type
RawDataEvent.
The RawDataEvent is a generic container for blocks of raw bytes, used to encapsulate the
data read out from the sensor electronics and send it to the DAQ. Each RawDataEvent
may contain any number of raw data blocks. By convention each block usually corresponds
to one sensor, but this is not required; it is up to each Producer how the raw data are
encoded, since it is up to the corresponding DataConverterPlugin how they are decoded.
The RawDataEvent class is defined in the following header file:
#include "eudaq/RawDataEvent.hh"
The class is described in more detail below.
33
EUDAQ User Manual
4. Writing a Producer
4.3.1. Constructing
A RawDataEvent is constructed as follows:
RawDataEvent event("EXAMPLE", run, event);
Where "EXAMPLE" is a string unique to the particular producer that will be used to
select the correct converter during decoding. The run and event parameters are the run
number and event number respectively.
As well as normal data events, the producer must also send a BORE and EORE at the
beginning and end of a run respectively. These are just normal RawDataEvent objects,
but with a particular flag set. The RawDataEvent has factory methods to simplify these
cases:
RawDataEvent::BORE("EXAMPLE", run);
RawDataEvent::EORE("EXAMPLE", run, event);
These methods return a RawDataEvent that may be either be sent directly to the DAQ,
or be modified first, e.g. by setting tags as described below in subsubsection 4.3.3.
4.3.2. Adding Data
Once a RawDataEvent has been constructed, data blocks may be added either using a
vector:
std::vector<unsigned char> buffer = ...;
event.AddBlock(id, buffer);
or using a pointer to a block of memory, and a length in bytes:
unsigned char * buffer = ...;
event.AddBlock(id, buffer, len);
Where id is an integer used to differentiate the different blocks. Usually it can just be 0
for the first block and increment by 1 for the following blocks. And buffer contains the
actual data for the block. If the buffer is a vector, the whole length is used, if it is a
pointer, then the length must be specified.
The type of the vector or pointer need not be unsigned char, since these methods are
in fact template methods that can take a vector of any basic type, but if larger types
are used, care must be taken about endianness, since the buffer will be converted to
unsigned char according to the endianness of the machine it is running on. Therefore
if the producer may run on different architectures steps should be taken to ensure that
any endianness issues are handled correctly.
4.3.3. Tags
The RawDataEvent (in fact any type that descends from the Event base class) may
also have tags set. These are name-value pairs containing extra information that does
not easily fit in the usual raw data. This is used particularly in the BORE to include
34
EUDAQ User Manual
4. Writing a Producer
information about the particular run that may be useful for the decoding later. A tag
may be set as follows:
event.SetTag("Temperature", 42);
The value corresponding to the tag can be set as an arbitrary type (in this case an
integer), it will be converted to a string internally.
4.4. Log Messages
A method is provided for sending log messages to the central Log Collector. To use it
the follwing header should be included:
#include "eudaq/Logger.hh"
This defines the following macros for sending log messages, listed in decreasing order of
severity:
EUDAQ USER : A user-generated message (e.g. from the RunControl Log button).
EUDAQ ERROR : Something that has gone wrong and should probably be looked into.
EUDAQ WARN : A warning that something may not be quite right.
EUDAQ INFO : An message generated during normal running containing information that
may be useful to the user.
EUDAQ EXTRA : Some extra information that may be less useful in normal running.
EUDAQ DEBUG : Information for debugging purposes that will normally be hidden.
They are used as follows:
EUDAQ_ERROR("No keyboard detected: press F1 to continue.");
The messages will be sent to the central Log Collector if it is connected, otherwise they
will be displayed on the local terminal. The log level can be changed in the following
way:
EUDAQ_LOG_LEVEL("WARN");
Any messages lower than the specified level will just be ignored. This can be useful to
filter out unimportant messages and, for example, just display error messages.
4.5. Interfacing Python-Code via the PyProducer Interface
As described in section 3.2.10, a Python interface is provided for selected EUDAQ
components including a Producer. This basic implementation can be extended on the
Python-side as demonstrated by the example python/example-producer.py.
35
EUDAQ User Manual
5. Data Conversion
5. Data Conversion
Data are stored on disk, by default, in a native binary format, containing the raw data as
read out by the various Producers. It is basically the same format used for serialising the
data over the socket connection to the Data Collector. To be useful, this data must be
converted into a standardised format so that the monitoring and analysis software does
not depend on particularities of the individual sensors, but can be applied generically
to any sensor. Two different formats are used for this. The first is the StandardEvent
type, an internal class that does not depend on any external libraries, and is used by
the online monitoring, and many of the utility programs of the framework. The second
type is the LCIO standard format from the linear collider community, used by the full
analysis software.
5.1. StandardEvent and StandardPlane
The StandardEvent is a class designed to represent pixel sensor data in a reasonably easy
to use way, but still be flexible enough to store the data from a wide range of different
sensors as completely as possible. Each StandardEvent represents one event of data from
the whole telescope and any DUTs, so a run will consist of a sequence of StandardEvents.
It inherits from the Event base class, meaning that it has a run number, an event number,
an optional timestamp, and may also contain tags (see subsubsection 4.3.3). It also has
an array of StandardPlanes, each representing one sensor plane of the telescope or DUT.
Each StandardPlane contains the charge values from the pixels of one sensor, and may
contain several frames in cases where the sensor is read out multiple times per event. It
also has the concept of a “result” frame, which is calculated from the one or more of the
source frames according to different rules that may be specified with flags. The result
frame contains only one charge value per pixel, with a positive signal, and is what will be
used for the analysis. It may consist of either differences between the original frames (e.g.
in the case of correlated double sampling (CDS)), a sum of all original frames, or specific
parts of the different frames selected according to the pivot information. Flags may be
set to select which of the different methods is used. It may also contain a submatrix
number per pixel, which can be used to differentiate different parts of the sensor, so
that they may be analyzed separately later, and a pivot boolean (true or false) per pixel,
which can be used to indicate whether the pixel was sampled before or after the trigger,
and is used to determine which parts of the sensor to combine when the FLAG NEEDCDS
flag is set.
Both the StandardEvent and the StandardPlane classes are defined in the following
header file:
#include "eudaq/StandardEvent.hh"
In general, a user should not need to construct a StandardEvent object, but should
create one or more StandardPlanes, that will be added to a given StandardEvent.
36
EUDAQ User Manual
5. Data Conversion
5.1.1. Constructor
The StandardPlane constructor has the following signature:
StandardPlane(unsigned id, const std::string & type,
const std::string & sensor = "");
Where id is an arbitrary numerical identifier for the plane that can be used to differentiate
between different planes of the same type, type is the type of the Producer that generated
the frame (should be the same as that in the Producer and the DataConverterPlugin),
and sensor is the name of the sensor, in the case that the Producer can read out more
than one type of sensor.
5.1.2. SetSizeRaw and SetSizeZS
Once a StandardPlane has been constructed, the size should be set. There are two
methods for doing this, depending on whether the data are stored in raw or zerosuppressed mode. In raw mode all pixels are stored, whether they have a signal or not.
In zero-suppressed mode, only those with a signal above a certain threshold are stored,
along with their coordinates, and any below the threshold are suppressed.
The signature of the SetSizeRaw method is:
void SetSizeRaw(unsigned w, unsigned h, unsigned frames = 1, int flags = 0);
Where w is the full width of the sensor (in the x-direction, usually columns) in pixels, h
is the full height of the sensor (in the y-direction, usually rows) in pixels, frames is the
number of frames, and flags may be a combination of the following values, separated
by a bitwise OR (i.e. |):
FLAG NEEDCDS : Indicates that the data are in 2 or 3 frames and that neighbouring frames
should be subtracted to produce the result.
FLAG NEGATIVE : Indicates that the charge values are negative, so should be negated to
produce the result.
FLAG ACCUMULATE : Indicates that all frames should be summed to produce the result.
FLAG WITHPIVOT : Indicates that pivot information is stored per pixel, and should be
used for constructing the result.
FLAG WITHSUBMAT : Indicates that submatrix information is stored per pixel.
FLAG DIFFCOORDS : Indicates that each frame can have different coordinates, in the case
of zero-suppressed data, otherwise all frames will share the same coordinates.
The signature of the SetSizeZS method is a follows:
void SetSizeZS(unsigned w, unsigned h, unsigned npix,
unsigned frames = 1, int flags = 0);
37
EUDAQ User Manual
5. Data Conversion
Where all parameters are the same as in SetSizeRaw, but there is an extra parameter
(npix) that specifies how many pixels to preallocate. If the number of pixels above
threshold is known, this may be used to allocate them all at once. If not, then this
parameter may be set to zero, and pixels can be allocated as needed (but note that this
way may be slower, since memory will need to be reallocated for each new pixel).
5.1.3. SetPixel and PushPixel
Once the size has been set, the values of the pixels can then be loaded into the
StandardPlane. There are two methods for doing this: SetPixel, that sets the value of
an already allocated pixel, and PushPixel that allocates space for a new pixel and sets
that.
The signatures of SetPixel are as follows:
void SetPixel(unsigned index, unsigned x, unsigned y, unsigned pix,
bool pivot = false, unsigned frame = 0);
void SetPixel(unsigned index, unsigned x, unsigned y, unsigned pix,
unsigned frame);
where index is the index of the pixel to set, x and y are the coordinates of the pixel, and
pix is the charge value for the pixel. The value of the pivot, and the frame number may
optionally be set also, if relevant. Note that if only the pivot is set, care should be taken
that it is of type bool to avoid accidentally setting the frame instead.
The signatures of PushPixel are as follows:
void PushPixel(unsigned x, unsigned y, unsigned pix,
bool pivot = false, unsigned frame = 0);
void PushPixel(unsigned x, unsigned y, unsigned pix,
unsigned frame);
where all parameters are the same as in SetPixel. The only difference being the lack of
an index parameter, since this will always be the newly allocated pixel.
5.1.4. Setting other information
Other than the pixel values, the StandardPlane also stores some other information that
should be set if applicable:
void SetTLUEvent(unsigned ev);
This sets the trigger ID as read out from the TLU. If it was read out and stored, it
should be set using this method to allow cross checks in the analysis.
void SetPivotPixel(unsigned p);
This sets the value of the pivot pixel (or pivot row etc. – the value is arbitrary). It is
only here to allow cross-checks in the analysis; if the pixels are to be combined using
the pivot information, then it should also be set in the per-pixel pivot values. The value
38
EUDAQ User Manual
5. Data Conversion
here cannot be used for that purpose since the order of reading out the pixels is not in
general known.
void SetFlags(FLAGS flags);
Some flags may be set after calling SetSizeRaw or SetSizeZS, but this is not possible
with the flags FLAG WITHPIVOT, FLAG WITHSUBMAT or FLAG DIFFCOORDS since these flags
affect how memory is allocated by those methods.
5.1.5. Adding to the StandardEvent
Once the plane has been constructed and filled, it may be added to a StandardPlane
using the following method:
StandardPlane & AddPlane(const StandardPlane &);
This will copy the plane into the list of StandardPlanes stored by the StandardEvent.
It will return a reference to the copy of the plane, that can be used to make further
modifications if necessary.
5.1.6. Extracting information
The StandardEvent inherits the following methods from the Event base class:
unsigned GetRunNumber() const;
unsigned GetEventNumber() const;
uint64_t GetTimestamp() const;
T GetTag(const std::string & name, T def) const;
allowing access to the run number, event number, timestamp (if set) and any tags (where
T is an arbitrary type). It also has the following methods to access the StandardPlanes
that it contains:
size_t NumPlanes() const;
const StandardPlane & GetPlane(size_t i) const;
These return the number of planes stored, and a reference to a particular plane, respectively. The individual planes can then be examined using the following methods:
const std::string & Type() const;
const std::string & Sensor() const;
unsigned ID() const;
unsigned TLUEvent() const;
unsigned PivotPixel() const;
These return the type of the plane (i.e. the type of Producer / DataConverter that
generated it), the type of sensor for the plane (in the case that the plane type can hold
different types of sensor data), the ID of the plane (used to differentiate different planes
of the same type), the TLU trigger ID for the plane (if it was read out and stored) and
39
EUDAQ User Manual
5. Data Conversion
the value of the pivot pixel (or pivot row) for the plane. Further information about the
plane is available in:
unsigned
unsigned
unsigned
unsigned
unsigned
unsigned
XSize() const;
YSize() const;
NumFrames() const;
TotalPixels() const;
HitPixels() const;
HitPixels(unsigned frame) const;
These return the full width and height of the sensor in pixels, the number of frames
stored for the plane, total number of pixels for the plane (i.e. full width × height), the
number of pixels over threshold (for zero-suppressed data) in the result frame, and the
number of pixels over threshold in a particular source frame.
Note that for the HitPixels method, there are two versions; the first takes no parameter
and returns the number of hit pixels in the result frame, while the second takes the frame
number as a parameter and returns the number of hit pixels in that frame from the
underlying source data. Normally the first version would be used, unless access is needed
to the raw data from the sensor. Similarly, the other methods for accessing the data all
have two versions:
double GetPixel(unsigned index) const;
double GetX(unsigned index) const;
double GetY(unsigned index) const;
const std::vector<pixel_t> & PixVector() const;
const std::vector<coord_t> & XVector() const;
const std::vector<coord_t> & YVector() const;
These return the charge value, the x coordinate and the y coordinate of a particular pixel
(for the first three methods), or a vector of these values for all pixels in the frame (for
the final three methods.
Here, coord t and pixel t are both double, even though the values stored are usually
integers. This is in order to make the StandardPlane as general as possible, allowing it
to store, for example, clusters with non-integer coordinates instead of pixels, and it also
makes it easier to pass the values directly into Root histograms without first having to
convert them to double. All the above methods also have a version taking the frame
number (as the second parameter if they already have one parameter), which returns the
information from the underlying source frame instead of the result frame.
5.2. LCIO and LCEvent
Due to time constraints, the LCIO format is not yet described in this manual. If you
need to write a converter to LCIO, first check whether a newer version of this manual is
available, otherwise you can look at the other converters that are already implemented,
and if that is not enough, seek the help of an expert.
40
EUDAQ User Manual
5. Data Conversion
5.3. DataConverterPlugin
In order to allow different DUTs to easily incorporate their data into the monitoring
and analysis chain, the DataConverterPlugin system was developed. This allows all the
conversion code for each producer to be kept in one file, with the necessary parts being
called automatically as needed. This section describes how to write a new converter
plugin, to use existing converter plugins see subsection 6.3.
Writing a converter plugin for a new producer involves defining a new class that derives
from the DataConverterPlugin base class and implementing a few methods. Each
converter plugin contains a unique string that defines which type of RawDataEvents it
is able to convert. This is the same string that is set in the RawDataEvent when it is
created by the relevant producer. The DataConverterPlugin class is defined in the
following header:
#include "eudaq/DataConverterPlugin.hh"
The methods to be implemented are described below, and a full example is provided in
subsection A.3.
5.3.1. Constructor
The constructor should call the DataConverterPlugin constructor, and pass as a parameter the string representing the type of RawDataEvent this plugin can convert. A single
static instance of the converter should then be defined, and instantiated in the source
file. This is illustrated below:
class ExampleConverterPlugin : public eudaq::DataConverterPlugin {
ExampleConverterPlugin() : eudaq::DataConverterPlugin("EXAMPLE") {
// constructor...
}
// more methods...
static ExampleConverterPlugin m_instance;
};
ExampleConverterPlugin ExampleConverterPlugin::m_instance;
this will cause the constructor to be called during initialization of the program, and the
DataConverterPlugin constructor will automatically register the plugin and make it
available in the PluginManager.
5.3.2. Initialization
Every time a new run is started, the Initialize method will be called. It has the
following signature:
virtual void Initialize(const Event & ev, const Configuration & c);
41
EUDAQ User Manual
5. Data Conversion
It receives as parameters the BORE, and the configuration used for the run. The plugin
may extract any tags from the BORE, or other information from the configuration, and
store it in member variables for use during decoding.
5.3.3. GetTriggerID
Since each producer that reads out the trigger ID from the TLU stores it differently in
the raw data, there is no general way to extract this information. The GetTriggerID
method remedies this, by providing a generic interface to access the trigger ID. The
signature is as follows:
virtual unsigned GetTriggerID(const Event & ev) const;
It receives the Event as a parameter, from which it should extract the TLU trigger ID,
and return it as an unsigned integer.
5.3.4. GetStandardEvent
This method should extract the sensor data from the RawDataEvent input parameter,
and fill in the StandardEvent by adding the appropriate number of StandardPlanes
(one per sensor plane). The method signature is:
virtual bool GetStandardSubEvent(StandardEvent & out,
const Event & in) const;
It should return true if it successfully updated the StandardEvent, or false to indicate
an error.
5.3.5. GetLCIOEvent
Similar to GetStandardEvent, the GetLCIOEvent method converts a RawDataEvent into
a standardized format, in this case LCIO. The signature is:
virtual lcio::LCEvent * GetLCIOEvent(const Event * ev) const;
It receives the RawDataEvent as a parameter, and should return a pointer to a new
LCEvent if the conversion is successful. In the event of an error, it should return a null
pointer.
42
EUDAQ User Manual
6. Other Parts of the Framework
6. Other Parts of the Framework
The EUDAQ framework contains a number of other parts that may be useful. Those
that have not already been described in previous sections will be outlined below.
6.1. FileWriter
The FileWriter part of the framework allows different file formats to be written, using a
common interface, using a plugin-like system to define new file types. The FileWriter
class defines the interface that each type must implement, and the FileWriterFactory
allows code that writes data files to select any available file type, and write it in a generic
way, without needing to know details about the particular file format. A number of
different file types are already implemented, for a list with descriptions, see page 27
The easiest way to make use of the different FileWriters, is to use the Converter.exe
program (see subsubsection 3.4.4).
The FileWriter base class is defined in the following header:
#include "eudaq/FileWriter.hh"
In order to implement a new FileWriter, a new class must be written, inheriting from
the FileWriter base class, and implementing the following methods:
virtual void StartRun(unsigned);
virtual void WriteEvent(const DetectorEvent &);
virtual uint64_t FileBytes() const;
The StartRun method is called at the start of each new run with the run number as
a parameter. This allows a new file to be opened, and any header information to be
written if necessary. Then the WriteEvent method is called for each event to be written.
Here the DetectorEvent can be decoded and processed and the necessary data written
to file. The FileBytes method should return the number of bytes written to the file.
However, it is optional, and may simply return zero if the actual size is not easily known.
6.2. FileReader
Although tools are provided to access the information in the native data files, and
to convert them to other formats (such as LCIO, for analysis with the EUTelescope
package), in some cases it may be preferable to access the native data directly. For this,
the FileReader class is provided, allowing a custom program to be written to access a
native file and process it as desired.
The constructor takes as an argument the name of the file to be opened, and will read
the first event from the file (which should be the BORE). The NextEvent() method can
then be called to advance through the file. It can optionally take as a parameter the
number of events to skip, and will return true as long as a new event was read. The
currently loaded event can be accessed with the GetDetectorEvent() method.
43
EUDAQ User Manual
6. Other Parts of the Framework
The basic usage is shown below, while a more complete example is available in subsection A.4:
#include "eudaq/FileReader.hh"
#include <iostream>
int main(int argc, char ** argv) {
if (argc != 2) {
std::cerr << "usage: " << argv[0] << " file" << std::endl;
return 1;
}
eudaq::FileReader reader(argv[1]);
std::cout << "Opened file: " << reader.Filename() << std::endl;
std::cout << "BORE:\n" << reader.GetDetectorEvent() << std::endl;
while (reader.NextEvent()) {
std::cout << reader.GetDetectorEvent() << std::endl;
}
return 0;
}
This will open the file specified on the command line, and print out a summary of all the
events in there. Be aware that running it as it is may generate a large amount of output,
especially with large data files.
6.3. PluginManager
The PluginManager handles the different DataConverterPlugins, allowing raw data
stored in a RawDataEvent to be easily converted to a StandardEvent or LCEvent without
having to know the details of all the detector types in there. It is defined in the following
header:
#include "eudaq/PluginManager.hh"
In order to convert the events correctly, the plugins must have access to the information
in the BORE. Therefore, before any events may be converted, and for each data file, the
PluginManager must be initialized as follows:
eudaq::PluginManager::Initialize(bore);
The PluginManager will take care of passing the relevant parts of the BORE to the
appropriate DataConverterPlugins. The DetectorEvents can then be converted as
follows:
eudaq::StandardEvent sev = eudaq::PluginManager::ConvertToStandard(dev);
The PluginManager will take care of splitting the DetectorEvent into its constituent
subevents, and passing them all to the appropriate DataConverterPlugins to be inserted into the returned StandardEvent. For a slightly more complete example of the
PluginManager, see the ExampleReader in subsection A.4.
44
EUDAQ User Manual
6. Other Parts of the Framework
6.4. OptionParser
The OptionParser is used to simplify parsing of command-line options. It provides a
way to specify which arguments a program accepts, with the types, default values and
descriptions, so that the help text can be automatically generated, and therefore is always
in sync with the code, and all command line programs can have a uniform interface.
All programs using the OptionParser will automatically provide a -h (and --help)
option to display the help text, as well as a -v (and --version) option to display the
program version, unless the program explicitly overrides these options with other ones
with the same names.
The OptionParser is the class that handles the actual parsing of the command line. The
signature of the constructor is as follows:
OptionParser(const std::string & name, const std::string & version,
const std::string & desc="", int minargs = -1, int maxargs = -1);
The first three arguments are the program name, version and (optionally) description,
and these are optionally followed by two numbers specifying the number of arguments
expected after the command line options. The default value of -1 for the minimum means
no arguments are allowed, and for the maximum means that an arbitrary number may
be given (i.e. there is no explicit maximum).
If the automatically generated help text is not sufficient, extra text may also be given to
display at the end of the help text, by passing it to the following method:
void OptionParser::ExtraHelpText(const std::string & text);
This can be used to provide extra information about the options to the program.
Once an OptionParser object has been constructed, the different options may be specified.
There are two types: OptionFlag, which specifies a simple option with no argument,
and the template Option<T>, which specifies an option taking an argument of type T.
The OptionFlag constructor has the following signature:
OptionFlag(OptionParser & op, const std::string & shortname,
const std::string & longname, const std::string & desc = "");
where op is a reference to the OptionParser object created previously, that will do the
actual parsing of the command line. It then takes two names: a short version (usually
a single character) that is used with a single hyphen, and a long version that must be
preceded by two hyphens on the command line. Finally, a description may be given that
will be displayed in the help text.
The Option constructor has the following two signatures, one for normal types, the other
for vectors of another type:
Option<T>(OptionParser & op, const std::string & shortname,
const std::string & longname, const T & deflt = T(),
const std::string & argname = "", const std::string & desc = "");
Option<std::vector<T> >(OptionParser & op, const std::string & shortname,
const std::string & longname, const std::string & argname = "",
45
EUDAQ User Manual
6. Other Parts of the Framework
const std::string & sep = "", const std::string & desc = "");
where, in both cases, the first three arguments are as for OptionFlag. The first constructor
then takes a default value that will be used in the case the option is not specified on
the command line, a name for the argument to the option (to be used in the help text),
and a description of the option. The vector version also takes an argument name and a
description, but no default value (the default is always an empty vector), instead it takes
a separator, which is the string used to separate multiple elements of the vector on the
command line. By default (or if an empty string is specified), a comma will be used.
Once all the options have been specified, the command line can be parsed, which is done
by calling the following method of the OptionParser object:
OptionParser & OptionParser::Parse(const char ** args);
as an argument it takes the list of arguments from the command line (by convention
usually called argv). If there is an error during parsing, an exception may be thrown;
this should be handled by the HandleMainException method as described below.
Afterwards the values of the options can be accessed using their Value() method. The
IsSet() method is also available to tell whether an option has been set on the command
line (for OptionFlags this will hold the same value as the Value() method).
Finally, the OptionParser has a HandleMainException method that provides a way to
catch any unhandled exceptions, and either display help if it is a problem with parsing
the command line, or otherwise display a standard text informing the user of a problem.
It will also catch exceptions of type MessageException and display the message, without
treating it as an error, so this can be used to exit the program with a message to the
user. It is recommended to put the main program inside a try block, then call the
HandleMainException method from a catch(...) block, after any other exceptions
have been handled (if necessary).
An example use is shown below, illustrating most of what is described above:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
#include "eudaq/OptionParser.hh"
#include "eudaq/Utils.hh"
#include <iostream>
int main(int /*argc*/, char ** argv) {
eudaq::OptionParser op("Example", "1.0", "An example program", 0);
eudaq::OptionFlag test(op, "t", "test", "Enable test");
eudaq::Option<double> example(op, "e", "example", 3.14, "value",
"Example parameter");
eudaq::Option<std::vector<int> > another(op, "a", "another", "values", ";",
"Example vector");
op.ExtraHelpText("Some more information about this");
try {
op.Parse(argv);
std::cout << "Test: " << (test.IsSet() ? "Enabled\n" : "Disabled\n")
<< "Example: " << example.Value() << "\n"
46
EUDAQ User Manual
6. Other Parts of the Framework
17
<< "Another: " << eudaq::to_string(another.Value(), ", ")
18
<< std::endl;
19
if (op.NumArgs() == 0) {
20
throw(eudaq::MessageException("No arguments were given"));
21
}
22
for (unsigned i = 0; i < op.NumArgs(); ++i) {
23
std::cout << "Argument " << (i+1) << ": " << op.GetArg(i) << std::endl;
24
}
25
} catch(...) {
26
return op.HandleMainException();
27
}
28
return 0;
29 }
Running this program produces the following output:
./OptionExample.exe -h
Example version 1.0
An example program
usage: ./OptionExample.exe [options] [0 or more arguments]
options:
-t --test
Enable test
-e --example <value> (default = 42)
Example parameter
-a --another <values> (default = )
Example vector
Some more information about this program.
./OptionExample.exe
Test: Disabled
Example: 42
Another:
No arguments were given
./OptionExample.exe -t -e 2.718 -a 1;2;3 foo bar
Test: Enabled
Example: 2.718
Another: 1, 2, 3
Argument 1: foo
Argument 2: bar
47
EUDAQ User Manual
6. Other Parts of the Framework
6.5. Timer
The Timer class wraps the underlying operating system’s timer functions, making them
easier to use in a platform independent way. Whenever a Timer object is created, it will
record the current time. Then at any time in the future, the elapsed time in seconds
may be accesses with the Seconds() method.
There is also a Stop() method to stop the timer counting, so any subsequent calls to
Seconds will return the same value, and a Restart() method to reset the timer’s start
time to the current time and start counting again. An example use is shown below:
#include "eudaq/Timer.hh"
Timer t;
function_a();
cout << "Function A took " << t.Seconds() << " seconds." << endl;
t.Restart();
function_b();
cout << "Function B took " << t.Seconds() << " seconds." << endl;
// wait 3 microseconds
t.Restart();
while (t.Seconds() < 3e-6) {
// do nothing
}
This shows a timer being used to measure the execution time of two functions, and to
wait for a small delay. Usually to wait for a delay, it is preferable to use sleep (or mSleep,
see subsubsection 6.6.4), but in most operating systems the minimum delay for a sleep
is around 20 ms (even when using usleep which has microsecond resolution) so if the
delay must be shorter, a busy loop like above is needed.
6.6. Utils
The Utils package is a collection of useful functions and classes too small to merit their
own individual files. It is used by including the header:
#include "eudaq/Utils.hh"
Some of the most useful parts are described here.
6.6.1. to string
This is a template function that takes (almost) any type and returns the value converted
to a string. An optional second argument specifies the minimum number of digits to use
(padding with zeroes if necessary).
int value = 123;
strfunction(to_string(value));
strfunction(to_string(value, 6));
48
EUDAQ User Manual
6. Other Parts of the Framework
This will pass first the string "123", and then the string "000123" to the function
strfunction.
6.6.2. from string
This template function is the inverse of to string. It takes as arguments a string and a
default value of type T, and returns an object of type T initialised from the string. If it
is not possible to convert the string to the required type, the default value is returned
instead.
std::string value = "456";
intfunction(from_string(value, 0));
This will call intfunction with the integer value 456.
6.6.3. hexdec
This is a class to facilitate printing numbers in both hexadecimal and decimal. It is
used similarly to to string, but when printed, it will display the value in hexadecimal,
followed by the value in decimal in parentheses. The hexadecimal values will be padded
to the full width of the type, unless a second argument is given specifying the minimum
number of hex digits to display.
short value = 789;
cout << hexdec(value) << endl
<< hexdec(value, 0) << endl;
This will display:
0x0315 (789)
0x315 (789)
If the result is required in a string, instead of being printed, this can be achieved with
to_string(hexdec(value)).
6.6.4. mSleep
This is a wrapper around the operating system’s sleep/usleep (or equivalent) function.
It takes as an argument the number of milliseconds to sleep. The advantage of this
function is that it will work on Linux, Mac OS X and Windows, as it will automatically
call the correct underlying function.
49
EUDAQ User Manual
7. Reporting Issues
7. Reporting Issues
The github server, on which EUDAQ is hosted, provides a system for reporting bugs
and for requesting new features. It is accessible at the following address: https:
//github.com/eudaq/eudaq/issues.
Here you may submit new reports (you are required to register first to do this), or follow
the status of existing bugs and feature requests. This is recommended over (or at least, as
well as) sending an email to the developers, as it ensures a record of the issue is available,
and others may follow the progress.
50
EUDAQ User Manual
8. Developing and Contributing to EUDAQ
8. Developing and Contributing to EUDAQ
8.1. Regression Testing
If a CMake version later than 2.8.0 is found and Python is installed together with the
numpy package, several regression tests are made available that can be executed through
CTest. The tests are based on the Python wrapper around EUDAQ components as
described in section 3.2.10. Run the tests by typing
cd build
cmake ..
ctest
This starts the script etc/tests/run dummydataproduction.py which runs a short
DAQ session with instances of RunControl, DataCollector and a (dummy) Producer and
compares the output to a reference file stored on AFS at DESY. If your system is set up
correctly, you have access to the reference file, and the basic components of the EUDAQ
library work, all tests should pass. To see the output of failing tests, you can add the
--output-on-failure parameter to the CTest command.
These basic tests can easily be extended to test other parts of the core framework or of
your own producer. Take a look at the etc/tests/testing.cmake CMake script and
the central CMakeLists.txt file where it is included for an example of how to set up
tests with CTest.
The automated nightly tests are defined in CMake scripts located in etc/tests/nightly
and are executed by the scripts run nightly.sh and run nightly.bat for Unix and
Windows platforms, respectively. In addition to the dummy run described above, the
nightly tests check out all changes from the central repository, build the full code base,
and submit all results to the CDash webserver hosted at DESY: http://aidasoft.desy.
de/CDash/index.php?project=EUDAQ
8.2. Commiting Code to the Main Repository
If you would like to contribute your code back into the main repository, please follow the
“fork & pull request” strategy:
• Create a user account on github, log in
• “Fork” the (main) project on github (using the button on the page of the main
repo)
• Either : clone from the newly forked project and add ’upstream’ repository to local
clone (change user names in URLs accordingly):
git clone https://github.com/hperrey/eudaq eudaq
cd eudaq
git remote add upstream https://github.com/eudaq/eudaq.git
51
EUDAQ User Manual
8. Developing and Contributing to EUDAQ
• or if edits were made to a previous checkout of upstream: rename origin to upstream,
add fork as new origin:
cd eudaq
git remote rename origin upstream
git remote add origin https://github.com/hperrey/eudaq
git remote -v show
• Optional: edit away on your local clone! But keep in sync with the development in
the upstream repository by running
git fetch upstream
git pull upstream master
# download named heads or tags
# merge changes into your branch
on a regular basis. Replace master by the appropriate branch if you work on a
separate one. Don’t forget that you can refer to issues in the main repository
anytime by using the string eudaq/eudaq#XX in your commit messages, where XX
stands for the issue number, e.g.
[...]. this addresses issue eudaq/eudaq#1
• Push the edits to origin (our fork)
git push origin
(defaults to git push origin master where origin is the repo and master the
branch)
• Verify that your changes made it to your github fork and then click there on the
“compare & pull request” button
• Summarize your changes and click on “send”
• Thank you!
52
EUDAQ User Manual
A. Source Code
A. Source Code
This section contains example code to illustrate the concepts in the manual, when they
are too long to be included in the main section.
All files are also present in the EUDAQ distribution; so if possible those versions should be
used, since they may be more up to date than the manual.
A.1. Example Config File
Latest version available at:
https://github.com/eudaq/eudaq/blob/master/conf/ExampleConfig.conf
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
# This is an example config file, you can adapt it to your needs.
# All text following a # character is treated as comments
[RunControl]
RunSizeLimit = 1000000000
[DataCollector]
FilePattern = "../data/run$6R$X"
[LogCollector]
SaveLevel = EXTRA
PrintLevel = INFO
[Producer.Example]
Parameter = 123
[Producer.TLU]
AndMask = 0xf
OrMask = 0
VetoMask = 0
DutMask = 20
TriggerInterval = 0
TrigRollover = 0
#DUTInput3=LEMO
53
EUDAQ User Manual
A. Source Code
A.2. Example Producer
Latest version available at:
https://github.com/eudaq/eudaq/blob/master/main/exe/src/ExampleProducer.cxx
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
"eudaq/Producer.hh"
"eudaq/Logger.hh"
"eudaq/RawDataEvent.hh"
"eudaq/Timer.hh"
"eudaq/Utils.hh"
"eudaq/OptionParser.hh"
"eudaq/ExampleHardware.hh"
<iostream>
<ostream>
<vector>
// A name to identify the raw data format of the events generated
// Modify this to something appropriate for your producer.
static const std::string EVENT_TYPE = "Example";
// Declare a new class that inherits from eudaq::Producer
class ExampleProducer : public eudaq::Producer {
public:
// The constructor must call the eudaq::Producer constructor with the name
// and the runcontrol connection string, and initialize any member ←variables.
ExampleProducer(const std::string & name, const std::string & runcontrol)
: eudaq::Producer(name, runcontrol),
m_run(0), m_ev(0), stopping(false), done(false),started(0) {}
// This gets called whenever the DAQ is configured
virtual void OnConfigure(const eudaq::Configuration & config) {
std::cout << "Configuring: " << config.Name() << std::endl;
// Do any configuration of the hardware here
// Configuration file values are accessible as config.Get(name, default)
m_exampleparam = config.Get("Parameter", 0);
std::cout << "Example Parameter = " << m_exampleparam << std::endl;
hardware.Setup(m_exampleparam);
// At the end, set the status that will be displayed in the Run Control.
SetStatus(eudaq::Status::LVL_OK, "Configured (" + config.Name() + ")");
}
// This gets called whenever a new run is started
// It receives the new run number as a parameter
54
EUDAQ User Manual
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
A. Source Code
virtual void OnStartRun(unsigned param) {
m_run = param;
m_ev = 0;
std::cout << "Start Run: " << m_run << std::endl;
// It must send a BORE to the Data Collector
eudaq::RawDataEvent bore(eudaq::RawDataEvent::BORE(EVENT_TYPE, m_run));
// You can set tags on the BORE that will be saved in the data file
// and can be used later to help decoding
bore.SetTag("EXAMPLE", eudaq::to_string(m_exampleparam));
// Send the event to the Data Collector
SendEvent(bore);
// At the end, set the status that will be displayed in the Run Control.
SetStatus(eudaq::Status::LVL_OK, "Running");
started=true;
}
// This gets called whenever a run is stopped
virtual void OnStopRun() {
std::cout << "Stopping Run" << std::endl;
started=false;
// Set a flag to signal to the polling loop that the run is over
stopping = true;
// wait until all events have been read out from the hardware
while (stopping) {
eudaq::mSleep(20);
}
// Send an EORE after all the real events have been sent
// You can also set tags on it (as with the BORE) if necessary
SendEvent(eudaq::RawDataEvent::EORE("Test", m_run, ++m_ev));
}
// This gets called when the Run Control is terminating,
// we should also exit.
virtual void OnTerminate() {
std::cout << "Terminating..." << std::endl;
done = true;
}
// This is just an example, adapt it to your hardware
void ReadoutLoop() {
// Loop until Run Control tells us to terminate
55
EUDAQ User Manual
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
while (!done) {
if (!hardware.EventsPending()) {
// No events are pending, so check if the run is stopping
if (stopping) {
// if so, signal that there are no events left
stopping = false;
}
// Now sleep for a bit, to prevent chewing up all the CPU
eudaq::mSleep(20);
// Then restart the loop
continue;
}
if (!started)
{
// Now sleep for a bit, to prevent chewing up all the CPU
eudaq::mSleep(20);
// Then restart the loop
continue;
}
// If we get here, there must be data to read out
// Create a RawDataEvent to contain the event data to be sent
eudaq::RawDataEvent ev(EVENT_TYPE, m_run, m_ev);
for (unsigned plane = 0; plane < hardware.NumSensors(); ++plane) {
// Read out a block of raw data from the hardware
std::vector<unsigned char> buffer = hardware.ReadSensor(plane);
// Each data block has an ID that is used for ordering the planes ←later
// If there are multiple sensors, they should be numbered ←incrementally
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
A. Source Code
// Add the block of raw data to the event
ev.AddBlock(plane, buffer);
}
hardware.CompletedEvent();
// Send the event to the Data Collector
SendEvent(ev);
// Now increment the event number
m_ev++;
}
}
private:
// This is just a dummy class representing the hardware
// It here basically that the example code will compile
// but it also generates example raw data to help illustrate the decoder
56
EUDAQ User Manual
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
A. Source Code
eudaq::ExampleHardware hardware;
unsigned m_run, m_ev, m_exampleparam;
bool stopping, done,started;
};
// The main function that will create a Producer instance and run it
int main(int /*argc*/, const char ** argv) {
// You can use the OptionParser to get command-line arguments
// then they will automatically be described in the help (-h) option
eudaq::OptionParser op("EUDAQ Example Producer", "1.0",
"Just an example, modify it to suit your own needs");
eudaq::Option<std::string> rctrl(op, "r", "runcontrol",
"tcp://localhost:44000", "address",
"The address of the RunControl.");
eudaq::Option<std::string> level(op, "l", "log-level", "NONE", "level",
"The minimum level for displaying log messages locally");
eudaq::Option<std::string> name (op, "n", "name", "Example", "string",
"The name of this Producer");
try {
// This will look through the command-line arguments and set the options
op.Parse(argv);
// Set the Log level for displaying messages based on command-line
EUDAQ_LOG_LEVEL(level.Value());
// Create a producer
ExampleProducer producer(name.Value(), rctrl.Value());
// And set it running...
producer.ReadoutLoop();
// When the readout loop terminates, it is time to go
std::cout << "Quitting" << std::endl;
} catch (...) {
// This does some basic error handling of common exceptions
return op.HandleMainException();
}
return 0;
}
57
EUDAQ User Manual
A. Source Code
A.3. Example DataConverterPlugin
Latest version available at:
https://github.com/eudaq/eudaq/blob/master/main/lib/src/ExampleConverterPlugin.cc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
#include "eudaq/DataConverterPlugin.hh"
#include "eudaq/StandardEvent.hh"
#include "eudaq/Utils.hh"
// All LCIO-specific parts are put in conditional compilation blocks
// so that the other parts may still be used if LCIO is not available.
#if USE_LCIO
# include "IMPL/LCEventImpl.h"
# include "IMPL/TrackerRawDataImpl.h"
# include "IMPL/LCCollectionVec.h"
# include "lcio.h"
#endif
namespace eudaq {
// The event type for which this converter plugin will be registered
// Modify this to match your actual event type (from the Producer)
static const char* EVENT_TYPE = "Example";
// Declare a new class that inherits from DataConverterPlugin
class ExampleConverterPlugin : public DataConverterPlugin {
public:
// This is called once at the beginning of each run.
// You may extract information from the BORE and/or configuration
// and store it in member variables to use during the decoding later.
virtual void Initialize(const Event & bore,
const Configuration & cnf) {
m_exampleparam = bore.GetTag("EXAMPLE", 0);
#ifndef WIN32 //some linux Stuff //$$change
(void)cnf; // just to suppress a warning about unused parameter cnf
#endif
}
// This should return the trigger ID (as provided by the TLU)
// if it was read out, otherwise it can either return (unsigned)-1,
// or be left undefined as there is already a default version.
virtual unsigned GetTriggerID(const Event & ev) const {
static const unsigned TRIGGER_OFFSET = 8;
// Make sure the event is of class RawDataEvent
58
EUDAQ User Manual
if (const RawDataEvent * rev = dynamic_cast<const RawDataEvent *> ←(&ev)) {
// This is just an example, modified it to suit your raw data format
// Make sure we have at least one block of data, and it is large ←enough
if (rev->NumBlocks() > 0 &&
rev->GetBlock(0).size() >= (TRIGGER_OFFSET + sizeof(short))) {
// Read a little-endian unsigned short from offset TRIGGER_OFFSET
return getlittleendian<unsigned short> ←(&rev->GetBlock(0)[TRIGGER_OFFSET]);
}
}
// If we are unable to extract the Trigger ID, signal with (unsigned)-1
return (unsigned)-1;
43
44
45
46
47
48
49
50
51
52
53
54
55
56
A. Source Code
}
// Here, the data from the RawDataEvent is extracted into a ←StandardEvent.
// The return value indicates whether the conversion was successful.
// Again, this is just an example, adapted it for the actual data layout.
virtual bool GetStandardSubEvent(StandardEvent & sev,
const Event & ev) const {
// If the event type is used for different sensors
// they can be differentiated here
std::string sensortype = "example";
// Create a StandardPlane representing one sensor plane
int id = 0;
StandardPlane plane(id, EVENT_TYPE, sensortype);
// Set the number of pixels
int width = 100, height = 50;
plane.SetSizeRaw(width, height);
// Set the trigger ID
plane.SetTLUEvent(GetTriggerID(ev));
// Add the plane to the StandardEvent
sev.AddPlane(plane);
// Indicate that data was successfully converted
return true;
}
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78 #if USE_LCIO
79
// This is where the conversion to LCIO is done
80
virtual lcio::LCEvent * GetLCIOEvent(const Event * /*ev*/) const {
81
return 0;
82
}
83 #endif
84
59
EUDAQ User Manual
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105 }
A. Source Code
private:
// The constructor can be private, only one static instance is created
// The DataConverterPlugin constructor must be passed the event type
// in order to register this converter for the corresponding conversions
// Member variables should also be initialized to default values here.
ExampleConverterPlugin()
: DataConverterPlugin(EVENT_TYPE), m_exampleparam(0)
{}
// Information extracted in Initialize() can be stored here:
unsigned m_exampleparam;
// The single instance of this converter plugin
static ExampleConverterPlugin m_instance;
}; // class ExampleConverterPlugin
// Instantiate the converter plugin instance
ExampleConverterPlugin ExampleConverterPlugin::m_instance;
// namespace eudaq
60
EUDAQ User Manual
A. Source Code
A.4. Example Reader
Latest version available at:
https://github.com/eudaq/eudaq/blob/master/main/exe/src/ExampleReader.cxx
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
#include
#include
#include
#include
"eudaq/FileReader.hh"
"eudaq/PluginManager.hh"
"eudaq/OptionParser.hh"
<iostream>
static const std::string EVENT_TYPE = "Example";
int main(int /*argc*/, const char ** argv) {
// You can use the OptionParser to get command-line arguments
// then they will automatically be described in the help (-h) option
eudaq::OptionParser op("EUDAQ Example File Reader", "1.0",
"Just an example, modify it to suit your own needs",
1);
eudaq::OptionFlag doraw(op, "r", "raw", "Display raw data from events");
eudaq::OptionFlag docon(op, "c", "converted", "Display converted events");
try {
// This will look through the command-line arguments and set the options
op.Parse(argv);
// Loop over all filenames
for (size_t i = 0; i < op.NumArgs(); ++i) {
// Create a reader for this file
eudaq::FileReader reader(op.GetArg(i));
// Display the actual filename (argument could have been a run number)
std::cout << "Opened file: " << reader.Filename() << std::endl;
// The BORE is now accessible in reader.GetDetectorEvent()
if (docon.IsSet()) {
// The PluginManager should be initialized with the BORE
eudaq::PluginManager::Initialize(reader.GetDetectorEvent());
}
// Now loop over all events in the file
while (reader.NextEvent()) {
if (reader.GetDetectorEvent().IsEORE()) {
std::cout << "End of run detected" << std::endl;
// Don't try to process if it is an EORE
break;
}
61
EUDAQ User Manual
43
44
45
46
47
48
A. Source Code
if (doraw.IsSet()) {
// Display summary of raw event
//std::cout << reader.GetDetectorEvent() << std::endl;
try {
// Look for a specific RawDataEvent, will throw an exception if ←not found
const eudaq::RawDataEvent & rev =
reader.GetDetectorEvent().GetRawSubEvent(EVENT_TYPE);
// Display summary of the Example RawDataEvent
std::cout << rev << std::endl;
} catch (const eudaq::Exception & e) {
std::cout << "No " << EVENT_TYPE << " subevent in event "
<< reader.GetDetectorEvent().GetEventNumber()
<< std::endl;
}
49
50
51
52
53
54
55
56
57
58
}
59
60
if (docon.IsSet()) {
61
// Convert the RawDataEvent into a StandardEvent
62
eudaq::StandardEvent sev =
63
eudaq::PluginManager::ConvertToStandard(reader.GetDetectorEvent());
64
65
// Display summary of converted event
66
std::cout << sev << std::endl;
67
}
68
}
69
}
70
71
} catch (...) {
72
// This does some basic error handling of common exceptions
73
return op.HandleMainException();
74
}
75
return 0;
76 }
62
EUDAQ User ManualB. Introduction to the build system and project files on Windows
B. Introduction to the build system and project files on
Windows
B.1. MSBUILD
This is the program that processes the project (solution) files and feeds it to the compiler
and linker. If you have a working project file it is more or less straight forward. It has a
very simple syntax:
MSBUILD.exe MyApp.sln /t:Rebuild /p:Configuration=Release
myApp.sln is the file you want to Process. The parameter /target (short /t) tells
msbuild what to do in this case rebuild. You have all the options you need like: clean,
build and rebuild. You can also specify your own targets. With the “parameter property”
switch you can change the properties of your Project. Let’s say you want to compile
EUDAQ, you go in the build folder where the solution (sln) file is and type:
MSBUILD.exe EUDAQ.sln /p:Configuration=Release
One thing one has to keep in mind is that there are some default configurations. The
default is a debug build for x86. If you want to have it different then you need to specify
it in the command line. And one thing you want to have is a release build! With the /p
switch you can overwrite properties like in this case the configuration. But you could
also overwrite the compiler version it should use. Let’s say you want to use VS 2013 then
you have to specify it by writing:
MSBUILD.exe EUDAQ.sln /p:PlatformToolset=v120 /p:Configuration=Release
But be careful when changing the compiler settings. It is possible that some then link
against an incompatible version of your external libraries.
B.2. Project Files
Project files are the Visual Studio equivalent to Makefiles. The Project files have a very
easy syntax but a complicated mechanism behind it. Making changes to an existing file
is very easy. Writing a new one from scratch is expert level. But also, in most cases,
pointless because CMake does it for you. Therefore usually one gets a finished Project
file that was auto created by CMake and one just wants to make some minor changes to
it, therefore it is enough to know where one can tweak around.
Please remember to adjust the CMake files when you are done accordingly, so that your
changes are reproduced and not overwritten on the next CMake run.
Let’s start easy and assume you want to change the output directory. You can do this
by adding the following line to the corresponding Property group.
<PropertyGroup ←Condition=” '$( Configuration ) | $ ( Platform )'=='Release |Win32' ”>
<OutDir>. . \ . . \ Windows Binaries\</OutDir>
63
EUDAQ User ManualB. Introduction to the build system and project files on Windows
</PropertyGroup>
Or let’s say you want to change the compiler version. You can do this by changing the
platform toolset to the version you need. You can find this option in
<PropertyGroup ←Condition=” '$( Configuration ) | $ ( Platform )'=='Debug|Win32' ” ←Label=”Configuration”>
...
<PlatformToolset>v110</PlatformToolset>
</PropertyGroup>
V110 stands for Visual Studio 2012. V120 stands for VS 2013 and so on. The next
interesting switches are in here:
<ItemDefinitionGroup Condition=” '$( Configuration ) |$( Platform)'=='Debug|Win32' ”>
<ClCompile>
<PrecompiledHeader></PrecompiledHeader>
<WarningLevel>Level3</WarningLevel>
<Optimization>Disabled</Optimization>
<PreprocessorDefinitions>
WIN32;
DEBUG;
CONSOLE;
%(PreprocessorDefinitions )
</PreprocessorDefinitions>
<AdditionalIncludeDirectories>
. . \ . . \ main\include ;
. . \ . . \ extern\pthread−win32\include ;
. . \ . . \ tlu\include ;
. . \ . . \ extern\ZestSC1\windows 7\Inc ;
. . \ . . \ extern\libusb−win32−bin −1.2.6.0\ include
</AdditionalIncludeDirectories>
</ClCompile>
<Link>
<SubSystem>Console</SubSystem>
<GenerateDebugInformation>true</GenerateDebugInformation>
<AdditionalLibraryDirectories>
. . \ . . \ extern\libusb−win32−bin −1.2.6.0\ l i b \msvc\;
. . \ . . \ extern\ZestSC1\windows 7\ l i b \x86\
</AdditionalLibraryDirectories>
<AdditionalDependencies>
ZestSC1 . l i b ; libusb . l i b ; kernel32 . l i b ; user32 . l i b ; gdi32 . l i b ; winspool . l i b ; comdlg32 . l i b ;
advapi32 . l i b ; shell32 . l i b ; ole32 . l i b ; oleaut32 . l i b ; uuid . l i b ; odbc32 . l i b ;
odbccp32 . l i b ;%(AdditionalDependencies )
</AdditionalDependencies>
</Link>
</ItemDefinitionGroup>
An Item definition Group is the place where you define your items. One can compare
Items to a struct in C++; it is an object that contains different types of information.
The Condition statement works like an IF in C++.
64
EUDAQ User ManualB. Introduction to the build system and project files on Windows
In this article you can find all the possibilities you have: http://msdn.microsoft.
com/de-de/library/7szfhaft.aspx In the next line you are defining an item called
“CLCompile” and you give it the some attributes like “PreprocessorDefinitions” or
“AdditionalIncludeDirectories”. This Object contains all the information that gets sent
to the compiler. That means all the compiler flags are set here. The actual files are
included later in the project file. So for now you have only defined how you want to
compile your files but not what files you want to compile. AdditionalIncludeDirectories
does exactly what you think it does. It understands all relative paths and path with
environment variables exactly as it should. Next thing is “PreprocessorDefinitions”. It
also works exactly as you think it does. That means you can either define just names for
your #ifdef statements in the code or you can define macros like
<PreprocessorDefinitions>
SOMEVALUE=3;
WIN32;
DEBUG;
CONSOLE;
%(PreprocessorDefinitions )
</PreprocessorDefinitions>
Then you can call in your code SOMEVALUE and it will be 3. I do not know if it is
possible to define macro function like
#define x_square(x) x*x.
<AdditionalDependencies>
$(myFancyLibPath) \∗. l i b ;
odbccp32 . l i b ;%(AdditionalDependencies )
</AdditionalDependencies>
And it will link against all *.lib files in this directory.
Next thing you need to know is where to put your files you want to compile. Somewhere below the ItemDefinitionGroup there is an ItemGroup which contains the Include
statements. It looks like this:
<ItemGroup>
<ClCompile Include=”src\someFile . cc”/>
<ClCompile Include=”src\someOtherFile . cc”/>
...
<ClCompile Include=”src \∗.cpp”/>
...
</ItemGroup>
Here you can either put individual files or groups of files in. But be careful that you
don’t include the same file twice. There is also an ItemGroup which contains the include
files. This one seems to be more important for the IDE of VS so that it shows the header
files in the Solution Explorer.
65
EUDAQ User ManualB. Introduction to the build system and project files on Windows
A typical use case is that you wrote your own Data Converter Plugin. This file needs to
be mentioned here!
What you won’t find in the project file is the section that passes the files to the compiler.
This part is hidden behind the following import statement:
<Import Project=”$(VCTargetsPath)\Microsoft .Cpp. targets”/>
It is usually not required to modify this file. But if you want to view it you can find it in
this folder:
C:\Program Files (x86)\MSBuild\Microsoft.Cpp\v4.0\
This file is written neither to be very clear nor understandable, so better check out the
documentation pages such as:
http://msdn.microsoft.com/en-us/library/dd293626.aspx
B.3. Known Problems
• The environment variables are pulled in as properties therefore they can be overwritten in the project file or in the “vcxproj.user” file. So if for example your
QT Project won’t compile and keeps complaining about not finding the correct
directory make sure you are not overwriting the QTDIR environment Variable with
a Property.
66
EUDAQ User Manual
C. Online Monitor Configuration Settings
C. Online Monitor Configuration Settings
C.1. Configuration Sections Overview
we have the following Section Keywords, to be put in [].
• [General]
• [Correlations]
• [Clusterizer]
• [HotPixelFinder]
• [Mimosa26]
C.2. Configuration options in [General]
SnapShotDir string
Stores the location of snapshots from the online monitor
SnapShotFormat string
Which Format to use for the snapshots, e.g. ”.pdf”
C.3. Configuration options in [Correlations]
MinClusterSize int
Which minimum cluster size to use for the correlation plots
DisablePlanes int,int,int
List of planes to disbale, separates by a ”,”
C.4. Configuration options in [Clusterizer]
C.5. Configuration options in [HotPixelFinder]
HotPixelCut float
Cut above which a pixel is considered ”hot”
C.6. Configuration options in [Mimosa26]
Mimosa26 max sections int
Number of section of the Mimosa 26 chip, default is 4
Mimosa26 section boundary int
Number of pixels in a Mimosa26 section, default is 288
67
EUDAQ User Manual
C. Online Monitor Configuration Settings
C.7. Configuration Example
[General]
SnapShotDir = "/scratch/eudet/EUDAQ/bin/"
SnapShotFormat = ".pdf"
[Correlations]
MinClusterSize = 2
DisablePlanes = 2,3
[Clusterizer]
[HotPixelFinder]
HotPixelCut = 0.05
[Mimosa26]
Mimosa26_max_sections = 4
Mimosa26_section_boundary = 288
68
EUDAQ User Manual
Glossary
Glossary
BORE beginning-of-run-event, basically a run header.
CDS correlated double sampling, when two frames are acquired, one before and one
after the trigger, and then subtracted to get the actual signal.
DUT device under test.
EORE end-of-run-event, basically a run trailer.
EUDRB the EUDET data reduction board, a VME readout board for pixel sensors.
LCIO Linear Collider I/O, the file format used by the analysis software.
TLU trigger logic unit.
Acknowledgements
This work is supported by the Commission of the European Communities under the 6th
Framework Programme “Structuring the European Research Area,” contract number
RII3–026126.
References
[1] P. Roloff, “The EUDET high resolution pixel telescope”, Nucl. Instrum. Meth., A604,
(2009), 265–268.
[2] A. Bulgheroni, “EUTelescope, the JRA1 tracking and reconstruction software: a
status report”, EUDET-Memo-2008-48.
URL http://www.eudet.org/e26/e28/e615/e835/eudet-memo-2008-48.pdf
[3] D. G. Cussans, “Description of the JRA1 Trigger Logic Unit (TLU), v0.2c”, EUDETMemo-2009-04.
URL http://www.eudet.org/e26/e28/e42441/e57298/EUDET-MEMO-2009-04.pdf
[4] A. Cotta Ramusino, “The EUDET Data Reduction Board (EUDRB)”, EUDETMemo-2008-38.
URL http://www.eudet.org/e26/e28/e615/e814/eudet-memo-2008-38.pdf
[5] EUDET JRA1 Group, “EUDET Pixel Telescope Data Taking Manual - Updated
Version 2009”, EUDET-Memo-2009-03.
URL http://www.eudet.org/e26/e28/e42441/e67493/EUDET-MEMO-2009-03.pdf
69

Similar documents

×

Report this document