Title: LAT Science Tools: Status
1LAT Science Tools Status Concept
- Seth Digel
- W. W. Hansen Experimental Physics Laboratory
- Stanford University
- digel_at_stanford.edu
2Status of Science Tools
- Planning for the standard analysis environment is
fairly advanced - the set of analysis tools, utilities, and
databases to be developed jointly with the GLAST
Science Support Center - concept is to include everything that a guest
investigator (or LAT Co-I) will need for routine
high-level analysis of LAT data - Initial definition was undertaken by the SSC-LAT
science tools software working group - convened by Jonathan Ormes and Peter Michelson in
early spring - SSC representation David Band (science lead),
Jay Norris (technical management), Bob Schaefer
(software lead) - LAT Toby Burnett (simulation software
architect), Seth Digel (coordinator for science
tools planning), Richard Dubois (SAS subsystem
lead) - meets weekly via VRVS (http//www-glast.slac.stanf
ord.edu/ScienceTools/slwg/meetings)
3Science Tools status (2)
- Requirements summaries and use cases
(step-by-step descriptions of analysis
procedures) were solicited from the LAT
collaboration and the SSC to verify the
completeness of the standard analysis environment
(http//www-glast.slac.stanford.edu/ScienceTools/t
ool_defs) - A science tools software workshop was held June
12-14 at SLAC - goals were to review and refine the definition of
the standard analysis environment, and to
organize for detailed definition of the
requirements - 30 attendees (LAT SSC)
- http//www-glast.slac.stanford.edu/ScienceTools/wo
rkshops/june02 - The workshop agenda was organized to have
significant time devoted to breakout sessions for
detailed discussions - Conveners of the breakout sessions, selected by
the SSC-LAT WG for their scientific and software
backgrounds, sheperded the sessions along,
reported on the progress, and are on the hook to
organize the detailed requirements documents
4Science Tools status (3)
- The development areas and their organizers are
- Databases and related utilities Bob Schaefer
(SSC), Karl Young (LAT) - Analysis tools David Band (SSC), Seth Digel
(LAT) - Likelihood analysis Pat Nolan (LAT)
- Pulsar-related Masa Hirayama (SSC)
- Source identification Isabelle Grenier (LAT)
- Gamma-ray burst-related David Band (SSC)
- Observation simulation Toby Burnett (LAT)
- User interface Jim Chiang (SSC), Heather Kelly
(LAT)
5Science Tools status (4)
- Current work
- defining the requirements in detail
- defining software development standards
- assembling a development schedule based on
priorities and dependencies - prototyping some tools and databases
- Must ensure that resources available (FTE-years)
are consistent with needs so far no significant
mismatch is indicated - Next milestone is an internal review (with
external reviewers) of the requirements and
development plan, in September
6Where do the requirements come from?
- The requirements summaries do not distinguish
them, but several sources of requirements can be
identified - Some come from the mandate in the AO that a
single analysis environment be developed for the
LAT team and for external users supported by the
SSC - relates to integration of tools and the database
architecture - Some come from a formal agreement between the LAT
team and NASAs HEASARC (High Energy Astrophysics
Science Archive Research Center), literally the
ultimate destination of the data and analysis
tools - e.g., the clear separation of the databases from
the analysis tools, the requirements that the
tools read and write FITS files, and the use of
CALDB for response functions - Other kinds of requirements
- Functional requirements for what a tool must
accomplish - Interface requirements for how a tool receives
input and delivers output - Performance requirements for how a rapidly a tool
must function
7Concept of the Analysis Environment
- The plan for the high-level analysis environment
treats the LAT as an astronomical instrument - Implicit in this assumption is that the
high-level analysis can proceed with the
instrument response functions - IRFs are defined from beam tests and MC
instrument simulations and monitored in flight - Also implicit is that background rejection is
well in hand (i.e., meets performance
requirements) - Routine scientific analysis will not require
modeling the charged particle background or
making cuts to select good times by geomagnetic
quantities
8Components of the Analysis Environment
Flow is generally left to right, from Level 1 to
Level 2
Image/plot display (U8)
Pulsar ephem (D5)
Source model def. tool (U7)
Event display
Level 0.5
Pulsar phase assign (A3)
Likelihood (A1)
Point source catalog (D3)
Data extract (U1)2
Level 1 (D1)1
GRB spectral/temporal analysis (A6)
Src. ID (A2)
Catalog Access (U5)
Exposure calc. (U3)
Pt.ing/livetime extractor (U2)
Pointing history (D2)1
GRB physical modeling (A7)
Astronomical catalogs (incl. D4)
IRFs (D6)
Map gen(U4)
Interstellar emission model (A8)
Pulsar period search (A5)
Pulsar profiles (A4)
1 High-level observation simulators (not shown)
generate alternate versions of these. 2 Data can
also be extracted (sub-selected) from
previously-extracted Level 1 data sets.
9Walk through
- Event display
- to be inherited from instrument simulation
- no path to higher-level analysis
- reference by event ID, useful for sanity check
10Walk through (2)
- Connection to Level 1 processing
- intermediate utilities stand between the Level 1
data and the analysis tools - exposure calculation is separate as well it
doesnt care about the source of the
pointing/livetime information - exposure calculation must match Level 1
extraction parameters
11Walk through (3)
- Known pulsars
- D5 represents timing information (radio)
maintained during the mission - after phase assignment, analysis can proceed as
usual, or periodicity tests applied
12Walk through (4)
- Other databases as input to high-level analysis
- Instrument response functions PSF, Aeff, Eng.
res. (RMF) - Interstellar emission model (tool or database?)
is unavoidable
13Walk through (5)
- Image/plot display
- Not actually free floating, resource for analysis
tools - Interstellar emission model (tool or database?)
is unavoidable
14Walk through (6)
- Source model definition tool
- specification of models (point sources, extended
sources, diffuse) for likelihood analysis and for
observation simulation (not shown) - probably should indicate that the source model
definitions can be written in files
15Walk through (7)
- Core analysis tools
- likelihood analysis characterization of point
and extended sources - map generation for visualization, export any
binning required for likelihood analysis will be
internal to analysis
16Walk through (8)
- Beyond Level 2
- LAT source catalog is related to likelihood
analysis, but not connected by a standard
analysis tool - astronomical catalogs may be external resources
17Toward the long-term schedule
- Milestones
- Review of standard analysis environment 2002/Q3
- SAS CDR 2003/Q1
- LAT CDR 2003/Q2
- Mock data challenge I 2003/Q4
- Mock data challenge II 2005/Q1
- Delivery of LAT 2005/Q3
- Review of standard analysis environment
- Requirements definitions for tools/databases,
management plan, schedule - SAS CDR
- Requirements definitions for additional (LAT
team-only) software (e.g., monitoring for
transients, inflight IRF monitoring,
high-resolution spectroscopy) - Progress on open issues, esp. Level 1 database
and exposure calc. - Definition of user interface
18Toward long-term schedule (2)
- LAT CDR
- Example tool (e.g., map generation) to exercise
programming infrastructure, utilities, and user
interface - Instrument response functions in CALDB
- Mock data challenge I
- Observation simulator (pointing history and
sources) - Level 1 database
- Core analysis tools, esp. likelihood, and
supporting utilities - Mock data challenge II
- All analysis tools and utilities (release 0)
- Delivery of LAT
- Entire standard analysis environment (release 1),
with documentation - Keep in mind that the standard analysis
environment will be developed jointly with the
SSC, not presented to the SSC in 2005/Q3