Science Tools Working Group
We met yesterday and most likely will next meet in 2 weeks.
The current version of ScienceTools remains v7r4p1. The next release will be at the end of the current 6-week build cycle, Oct. 13
Data products: No news. The status of the Science Data Products ICD was on the agenda for the Operations Technical Interchange Meeting held at Goddard this week. Julie is going to give a recap of this session at the Science Ops EVO Meetings tomorrow.
Databases and related utilities
No news.
Likelihood analysis
Jim has made gtcntsmap able to interpret energy bin definition files written by gtbindef so that it can make counts maps for specific user-specified energy ranges. The Diffuse group wants this functionality in order to make EGRET-like counts maps.
Jim has made gtlivetimecube run faster now for the case where a long pointing history file is provided as input but only a small part of the time range is actually requested.
Also, Jim has got gtmodelmap writing a non-empty GTI extension - it copies the GTI of the input counts file.
GRB tools
David described the functionality of gtburstfit. It is not a temporal-spectral model fitter but finds and fits temporal profiles of GRBs using binned data as input. Jeff Scargle apparently has figured out how to make it work with unbinned LAT data, too, but that is not yet implemented in the tool.
James is reworking gtrspgen to have it make interpolations, integrations, etc. over the response functions using cos(inclination angle) rather than just the inclination angle. This will make it consistent with how the response functions are handled, e.g., by gtlikelihood, and apparently will fix the discrepancies noted between Xspec and gtlikelihood analyses.
Pulsar tools
No news about the tools. Masa has updated the tutorials for the pulsar tools to be consistent with the new functionality of the user interface
Observation simulation
No news.
User interface and infrastructure
Toby gave a presentation (linked to the agenda page for the Science Tools meeting) summarizing how the IRFs have been defined from AllGamma data and specifications of event classification cuts, and described his proposal for how the process of producing the IRFs can be more streamlined. The IRF group, originally constituted for DC2, is being reformed (again chaired by Jim), as may already have been reported.
Jim reminded us that David has pointed out that the DATE-OBS and DATE-END time strings in our FITS files should be in the TT time system, to be consistent with the FITS convention. This is not a big deal but not something that can be done currently with the time handling classes in the astro package, which do not know about leap seconds. Toby is willing to work on this and David, Jim, Toby, me, and anyone else who is interested (let me know) will work out what specifically needs to be implemented.
David led a brief discussion about the desirability of making more of the science tools (optionally) produce graphical output. We were more time ocnstrained than usual during the meeting and will have to get back to this topic.
David is updating the Cicerone with input from Chris Shrader and Analia.
Source Catalog
The catalog group did not meet this week.
Science Tools Working Group
We met last week and will meet again Sept 27, 2006. The tentative agenda includes cleaning up any outstanding JIRA issues.
The current version of ScienceTools remains v7r4p1. An incremental update may or may not occur before the end of the next build cycle.
Data products: No news. The status of the Science Data Products ICD is on the agenda for the Operations Technical Interchange Meeting to be held at Goddard at the end of the month.
Databases and related utilities
No news.
Likelihood analysis
- I have implemented a speedup of the unbinned likelihood calculation that skips re-evaluation of contributions from sources that have no free model parameters. This should afford substantial time savings for various stages of the catalog analysis.
- The warnings based on the Poisson probability of the expected number of counts for the nominal energy ranges plotted for the counts spectra over the entire region-of-interest has been implemented by James and Analia and has been incorporated in gtlikelihood.
- The bugs reported last week regarding one-sided time range cuts and diffuse response column names have been fixed.
- gtcntsmap now supports different pixel sizes for the longitude and latitude coordinates of the counts maps that are created.
GRB tools
No news.
Pulsar tools
No news.
Observation simulation
Toby and I have modified the observationSim code so that gtobssim now uses the same scheme as Gleam for assigning MC_SRC_ID numbers to each event. This will ensure that event ids are the same from run to run, so long as the input source models are identical.
User interface and infrastructure
No news.
David is still looking for good citizens like you to read, comment on, and contribute to the Cicerone document. Several sections still need beefing up. He has made recent revisions to the pulsar analysis section.
Source Catalog
The catalog group met this week. Jean presented an update on his comparisons of the various source detection methods as applied to the DC2 test image data, and Gino presented some results of the PGWave package as applied to these data and the test pattern data.
Science Tools Working Group
We met this week and most probably will meet again in 2 weeks. This report covers the last 3 weeks, which were quiet ones for Science Tools.
The current version of the Science Tools remains v7r4p1. An incremental update may be in the works shortly.
Data products: No news. The status of the Science Data Products ICD is on the agenda for the Operations Technical Interchange Meeting to be held at Goddard at the end of the month.
Databases and related utilities
No news. The LAT Dataserver Working Group, a Service Challenge Subcommittee to be chaired by Julie, is going to start doing what committees do, "to help sharpen input to the Data Handling group for features and functionality to be built in to the LAT dataservers. It should result in a long term (1 year) plan."
Tom reports that a request for quotes is out for the production server systems to be used at the GSSC and the hardware is expected to be in place by the end of the year. My understanding is that this is for the LAT event/photon data; the other GLAST data products, which can be served using a standard Browse-type system will be handled on HEASARC systems.
Likelihood analysis
Jim has fixed, or is in the process of fixing, a couple of bugs: with how one-sided time-range cuts (say when only a lower-limit is specified) are handled and with how the diffuse response columns are named by gtdiffresp when event class cuts have been made.
In the near term he is also working on a method to speed up evaluations of the likelihood function by not recomputing parts that relate to sources whose parameters are not changing, such as sources in the border region around a region of interest. I think that the suggestion for doing this came from Jean. The implementation certainly has to be done carefully but could make a marked improvement for evaluation of models with a lot of static sources.
Analia is continuing to work on introducing a 'confidence level' for warning users when counts in each energy bin are significantly different from the model predictions. This would be evaluated on the summary counts spectra output from likelihood. If this turns out to be useful, presumably checks on other aspects (such as spatial agreement) could be added.
GRB tools
No news. We may have an update next time on whether a temporal-spectral model fitting tool will be developed.
Pulsar tools
Tom reported that he talked with David Smith at the collaboration meeting about the nuts and bolts of translating radio ephemerides into the D4 (pulsar database) format. Apparently at Bordeaux they have been working on some scripts to do this; when they are happy that the scripts are debugged, Tom will set up a system at GSSC to receive radio ephemerides and ingest them in to D4. I did not think to ask about details about how the ephemerides are do be delivered, if anyone is reading this. This also raised questions about how D4 is to be delivered to users in real life.
From Masa: "The pulsar tools now accept one more time format (MJD in addition to GLAST) and two more time systems (TAI and UTC in addition to TDB and TAI) for manual input of a time, such as a reference time (reftime) and an ephemeris epoch (ephepoch). What we plan to do in this build cycle are: cleaning up the codes, testing, identifying/fixing bugs,
and updating manuals."
Observation simulation
No news.
User interface and infrastructure
From James: "James began work to reconcile differences between the new parameter interface (Ape) and the Ciao tools parameter interface. The goal is to matching behavior with all major parameter-based astronomical software packages as a prelude for introducing Ape into the Science Tools."
Chuck has updated the 'splash page' for the science tools section of the user workbook and is considering coming up with an even more clear way for neophytes to find what they need in the workbook. He is also working with Vincent to incorporate documentation for gtsrcid.
David is still looking for good citizens like you to read, comment on, and contribute to the Cicerone document. Several sections need beefing up.
Source Catalog
The catalog group met last week and I don't recall why I report on it in these meetings. The presentations were a continuation of discussions of algorithms for source detection.