Science Tools Working Group
We met this week for the 20th and final time in 2006.
The version of ScienceTools remains v7r6p1, which includes the first version of the HANDOFF IRFS. The LATEST builds have somewhat updated parameterizations of the HANDOFF IRFS, probably more reliable at lower energies and extending to higher energies, with the same name.
Data products: No news per se. I'd like to get something like the Data Products Working Group going again to work out the many remaining details in the various science data products to be delivered to the GSSC, and also the products that will come out of, say, ASP.
Databases and related utilities
The SC1 data in the server at the GSSC are being delivered with incorrect GTIs; as far as I know, Stefan Funk was the first to notice. Part of the reason was a small coding error in the server. But Tom S. pointed out that the FT1 files delivered to the GSSC did not have correct GTIs either. So the extensions need straightening out, with gtmaketime. As of this writing, I think that the plan is to by one way or another update the GTIs of the files that the GSSC is serving.
Likelihood analysis
Jim made the evaluation of the 'goodness of fit' optional in gtlikelihood. Code that Analia and James added for this calculation is at least some of the time throwing an exception (while evaluating a gamma function). Jim has re-enabled the generation of scripts for running PyLikelihood without CMT. In the past this as not worked reliably on Windows, but the application that David L. has in mind is Linux.
Jean noticed that gtexpmap is running slowly with the HANDOFF response functions, like by a factor of 20-30 relative to DC2 response functions. Jim is working on optimizing the angular integrations of the HANDOFF response functions, which would speed up this and other Likelihood calculations. He expects to present something tomorrow at the IRF Working Group VRVS meeting - not the answer but the issues and some directions to be considered.
Related to source detection, David described plans at the GSSC to adapt/adopt code from Toby to estimate the flux that corresponds to a given value of TS within a certain range of time in a specified direction. The idea is to set this up as a Web-accessible service for people (eventually) preparing LAT GI proposals. Some other investigations, like Benoit's recent work on generating a sensitivity map using Jean's method, were also discussed.
GRB tools
David also described a tool (Web-based) to help people prepare proposals, using Webspec to generate simulated spectra for GRBS (LAT and GBM) and for point sources in general. This involves some assumptions, obviously. More details will come up at a future Science Tools meeting.
Pulsar tools
Masa is working on the issues described last time. He and James are also considering the design of the system that receives timing parameters from pulsar observers and updates the master file of ephemerides - issues are, e.g., logging, controlling new entries, and procedures for replacing entries.
Observation simulation
No new news.
User interface and infrastructure
James nearly has HOOPS retrofitted to use the new ape (replacement for PIL). From a user perspective the changes will be that environment variables will be handled automatically, and several PIL command-line problems (various annoyances documented in JIRA) will be resolved.
James and Toby will look into a problem Toby has experienced (Windows-only, apparently) using tip to write a FITS table.
Chuck reports that the Science Tools area of the User Workbook is up-to-date with all documentation that he's received. David will be working on some updates to Cicerone based on comments from the GUC.
Source Catalog
Did not meet this week.
Science Tools Working Group
Did not meet this week. We probably will meet next on December 12.
The version of ScienceTools remains v7r6p1, which includes the first version of the HANDOFF IRFS.
The feedback from the GLAST Users Committee 'Beta Test' of the science tools was positive overall. The members of the committee made a number of suggestions, which Chris Shrader is compiling, but none that suggested that the science tools were off track.
I gave a talk last week at the SC workshop on future directions for science tools development, which drew from contributions to Science tool development directions
Data products: No news, although there should be. FT1 probably needs a tweak in the definition of the LIVETIME column. FT2 needs a little pruning - the DEADTIME column is not used and would not be missed.
Databases and related utilities
No news. Actually, it looks like Jim has updated the dataSubselector package (gtselect) to merge range cuts in the event files that it processes. I think that this means that where possible, like when a succession of cuts on the same quantity have been applied to an events file, gtselect will merge the DSS specifications of those cuts.
Likelihood analysis
No direct news. Jim has updated Likelihood to have it check only that the energy range of an exposure map encloses the energy range of the event file. Formerly I think that it required exact agreement.
GRB tools
No news.
Pulsar tools
Masa has incorporated some of the feedback from the GUC review in the 'action items' on the Pulsar tools Web page; see the proposal for a new tool, gtptest and for 'Barycentering-on-the-fly' (which relates to JIRA issue PULS-10). gtbary changes every time-related quantity that it can in a FITS file, including of course the event times in the TIME column. Depending on how you look at it, the changes are not huge (light travel time to the sun - minutes) but they do make confusion possible down the line - say, if the response functions are changed at a certain time and the file is used for likelihood analysis.
Observation simulation
No direct news. Toby has implemented a trial versions of 'Sun' and 'Moon' reference frames for diffuse sources. These have not been extensively tested yet.
User interface and infrastructure
No new news.
Source Catalog
Met this week. We had some discussions about variability measures and Jean presented final results on detectability of closely-spaced sources and on localization accuracy of the source detection algorithms using the DC2 data.