org.lcsim ILC Reconstruction and Analysis

1
org.lcsim ILC Reconstruction and Analysis

• Tony Johnson
• SLAC
• March 2005

2
org.lcsim ILC Reconstruction and Analysis

• Earlier Packages
• hep.lcd Reconstruction and Analysis
• LCIOPlugin
• Goals of org.lcsim
• Components
• Conditions Handling
• Geometry Access
• Event Access
• Reconstruction
• Analysis Tools
• JAS Plugin
• Accessing Code
• Status and Plans

3
hep.lcd Overview

• Java based analysis reconstruction and analysis
  package
• In use for 5 years
• Runs standalone or inside JAS
• Full Reconstruction (trackingcalorimetry)
• Analysis Tools (including WIRED event display)
• Fast MC
• Smeared tracks and calorimetry clusters
• Physics Tools (Vertex Finding, Jet Finding)
• Beam Background Overlays
• Contributed Area
• Analysis Utilities and sample analyses provided
  by users

4
hep.lcd Physics Utilities

• Physics Utilities
• 4-vector, 3-vector classes
• Event shape/Thrust finder
• Jet Finders
• Many kT algorithms implemented (e.g. Jade and
  Durham )
• Extensible to allow implementation of other
  algorithms
• Topological Vertexing
• Implementation of SLDs ZVTOP
• Event Generators Framework
• Diagnostic Generator
• User-defined particle mix, momenta and vertices.
• Framework for reading stdhep files
• Generator interface for direct connection to
  (fortran) event generators
• PYTHIA, HERWIG, ISAJET,
• Beam background overlay

5
Full Reconstruction

• Tracking
• Hit Smearing/Efficiency (detector simulation puts
  out perfect hits)
• Random Background overlay
• Track Finding
• Full pattern recognition in the Central Barrel
  region
• Tuned for Large and Silicon Detectors
• Track Fitters
• SLD Weight Matrix Fitter
• Can do Single Detector or Combined fit (e.g.
  VTXTPC)
• Calorimetry
• Cluster Finding
• Several Clustering Algorithms Currently
  Implemented
• Cluster QA package developed
• Cluster Refinements
• Combine HAD EM clusters
• Endcap Barrel overlap region
• Identify cluster as originating from
  charged/neutral hadron, EM

6
Recent improvements to hep.lcd

• Ported to work with JAS3 (T. Johnson)
• New digitization routines being added (N.Sinev)
• Cluster identification code (G. Bower)
• Cluster diagnostic package (R. Cassell)
• Improved Stepper (C. Milstene)
• Forward Tracking (N. Graf)
• Flavour Tagging (D. Jackson)
• Particle ID (B. Wilson, S. Rolnick)
• Support for Non-Projective geometries (G. Lima)
• Garfield track-finder (E. von Toerne et al)

7
hep.lcd analysis tools

• Tutorial http//jas.freehep.org/jas3/Tutorial/ind
  ex.html

8
hep.lcd package pros-and-cons

• Features
• Can run standalone or inside JAS
• Runs on almost any platform
• No dependencies on pre-installed software or
  network connections. Can install and run in
  minutes from Snowmass 2001 CD
• Problems
• IO based on old LCD SIO format
• No way to write output from reconstruction
• Dependent on old geometry description
  (detector.ini) limited to cylindrical dectectors

9
LCIO plugin for JAS3

• Works with any LCIO file
• Diagnostic tools allow to step through and view
  events

10
LCIO plugin for JAS3

• Event Analysis

11
Using WIRED with LCIOPlugin

• LCIO files can (but dont necessarily) contain
  positions of all hits, but no detector geometry
  info except detector name

• LCIOPluginWIRED can view hits for any file, but
  these are hard to visualize without detector
• LCIOPlugin can read external HepRep2 file
  describing detector geometry
• Location currently needs to be specified
• HepRep file (xml) can be generated
• by hand (quite easy for simple geometry)
• by Geant4 using HepRep2 writer (perhaps with some
  filtering to remove excess details)
• By program using Java or C HepRep API

12
LCIOPlugin with WIRED4
13
LCIOPlugin with WIRED4
14
LCIO Plugins pros-and-cons

• Features
• Not dependent on any detector specific
  assumptions
• Simple environment for debugging problems with
  LCIO and writing simple analysis routines
• Detector independent event display
• Problems
• No connection to reconstruction or analysis tools
• Detector geometry must be specified by hand
• Existing LCIO API does not emphasize
  user-friendliness

15
org.lcsim Reconstruction and Analysis package
16
org.lcsim Goals

• Retain core functionality from hep.lcd package
• Full suite of reconstruction and analysis tools
  available to all LCIO users
• Update to use LCIO for IO and as basis for
  simulation, raw data and reconstruction event
  formats
• Isolate users from raw LCIO structures
• Update/simplify framework using experience from
  hep.lcd
• Provide good tutorial documentation
• Internationalization
• Try to make package independent of detector,
  geometry assumptions so can work with any
  detector
• Read properties of detectors at runtime
• Update to Java 1.5
• Many improvements since hep.lcd framework was
  created.
• Ability to run standalone or in JAS3

17
org.lcsim Conditions Database

• Goal Using detector name in LCIO file
• Provide access to a extensible set of conditions
  including
• Detector Geometry
• Algorithm Specific Constants
• E.g. FastMC smearing parameters
• Dont make assumptions about format of data
• Dont rely on internet access, or local database
  installation
• Make it easy for users to define new detectors

18
org.lcsim Conditions DatabaseImplementation

• Detector Constants stored in .zip file
• Typically contains
• Compact geometry file
• Set of (ascii) constants for standard algorithms
• Could additionally contain
• Arbitrary files (xml, ascii, binary) needed by
  other algorithms
• Other geometry formats (HepRep, LCDD)
• Links to database (e.g. for access to LCCD).
• To define a new detector just create a new .zip
  file.

19
org.lcsim Conditions DatabaseImplementation

• org.lcsim locates file based on detector name by
• Looking in alias.properties file in
• Users home directory
• Built-in to org.lcsim software
• On www.lcsim.org website
• If translated name is a URL (file, ftp, http)
• Read .zip file at specified location
• Search for .zip file in
• /.lcsim/detectors
• Built-in to org.lcsim software
• In http//www.lcsim.org/detectors
• Downloaded files are cached locally for future
  use

20
org.lcsim Conditions DatabaseAccessing constants

• Constants are accessed by name (actually path in
  zip file)

• Later on if constants change, listener will be
  called

21
Geometry Access

• Geometry Access is build on top of conditions
  database
• Framework reads and decodes compact geometry
  description.
• User can iterate over detectors or access
  detector by name or id
• Provides access to detector properties ( layers
  etc).
• Shape of detector
• Framework builds IDDecoder objects for each hit
  collection.
• ID decoder unpacks IDs in LCIO files
• Provides access hit position, detector layer etc.
• Can provide additional utilities, for instance
  list of neighboring cells in calorimeter

22
org.lcsim Drivers, Event Access

• Reconstruction and Analysis Code is written by
  extending Driver class.
• Most code extends only a single method
• process(EventHeader event)
• which passes in EventHeader through which all
  event data, conditions, and geometry is accessed.
• EventHeader maintains features which users liked
  from old hep.lcd framework, but adds
  compatibility with LCIO events
• All LCIO data is accessible.
• Arbitrary user data, either collections or single
  objects can be added to event.
• Event can be written out
• Currently only objects understood by LCIO are
  written out
• MCParticles, Hits, Clusters, Tracks,
  Reconstructed Particles (can include Jets,
  Vertices)
• Maybe be able to use LCIO GenericObjects to
  write out more in future.

23
Reconstruction Clustering

• Several Clustering Algorithms now included
• Cluster Cheater
• uses MC information
• Nearest Neighbor Clusterer
• Cone Clusterer
• Algorithms can be run in parallel for comparison

24
org.lcsim IO

• Can read LCIO files, stdhep files (for FastMC)
• Can write LCIO files

25
org.lcsim JAS Plugin

• The org.lcsim can be used standalone, or inside
  JAS3. Same code can be used in both places, so
  easy to move back and forth
• E.g. develop in IDE and run in JAS3
• E.g. develop in JAS3 and run in batch
• JAS3 plugin adds
• Example Analysis Code
• Event browser (similar to LCIOPlugin)
• Easy viewing of AIDA plots
• WIRED event display integration

26
org.lcsim Examples
27
org.lcsim Examples
28
org.lcsim Plot Viewing
29
Org.lcsim WIRED event display
30
org.lcsim Status

• Physics Utilities - done
• stdhep reader
• 3, 4-vector utilities
• diagnostic generator
• Jet finder, event shape utilities
• Conditions framework done
• Ability to read detector constants from zip
  file
• To define new detector just create new zip file
  and place on web
• File is read and cached locally
• Ability to read compact geometry file
• Driver framework done
• Fast MC done
• IO Framework done
• Event Access working, still being improved
• Event Display interface mostly done
• Reconstruction
• Clustering done
• Tracking TRF - in progress, very nearly done
• Vertex finding needs port from hep.lcd

31
Accessing Code

• To use you can just download JAS3, and then use
  the JAS3 Plugin Manager to install the org.lcsim
  Plugin
• http//confluence.slac.stanford.edu/display/ilc/In
  stallingJAS3
• To add new software to the framework check out
  the software from CVS and build using maven
• http//confluence.slac.stanford.edu/display/ilc/Bu
  ildingorg.lcsimsoftware

32
org.lcsim To Do List

• Improve Documentation/Tutorials
• Continue to (work with others to) migrate
  reconstruction algorithms from hep.lcd
• Strongly encourage people to work in CVS
• If we make changes to event access etc. we can
  update all code in CVS using refactoring tools
• Remove divergences from LCIO standard
• 100 compatible with LCIO file format, philosophy
• Attempts to make event access easier should be
  merged back into LCIO standard interfaces?
• LCIOPlugin and org.lcsim plugin should be merged
• Test ability to interoperate with other software,
  e.g. slic, marlin, mokka, jupiter, brahms,
  lelaps, etc.
• Produce new CD, hopefully well before Snowmass
  meeting.
• Online weekly clinic using VRVS/ES net?