Science Tools Working Group
We met this week and most probably will meet again in 2 weeks, after the collaboration meeting. This report covers the last 2 weeks.
The current version of the Science Tools is v7r4p1. It has a number of (incremental) changes from v7r4. Among these are a switch from the timeSystem classes to the timeCorrect classes for handling systems of time for the pulsar tools (Masa and James) and generalization of irfInterface to allow for time-dependence in the IRFs (Chiang).
Data products: No news. Masa pointed out that until recently the pulsar tools assumed that the reference time (the origin of MET) was MJD 51910 in the TT system. The actual reference time is MJD 51910 UTC, which amounts to a difference of about 64 seconds. This matters for pulsar timing; we can shift our origin of time all we want but we cannot change where the earth is at any given absolute time. The DC2 pulsar simulations worked because PulsarSpectrum must also have been assuming the same reference time. No need to panic, but I'd stick with v7r3 if you want to analyze DC2 pulsar data.
Databases and related utilities
No news. Tom pointed out that the GRT 5 test being run today, which includes sending an FT1 file to the GSSC, will include ingesting the file into the data server at the GSSC. The server runs probably as many validity checks on incoming files as reasonably possible.
Likelihood analysis
Jim reports that for the columns that gtdiffresp adds to FT1 files the delimiter between IRF name and diffuse source name is now "__" instead of "::". The change is backward compatible; the new naming scheme is consistent with HEASARC "recommendations" for naming columns in FITS binary tables.
At the request from a user, Jim added a hidden parameter (write_output_files) to the parameter file for gtlikelihood. Setting it to no disables writing output files including the new counts_spectra.fits. He also fixed a bug in the generation of counts_spectra.fits.
GRB tools
No news.
Pulsar tools
Masa reports that the pulsar tools now interpret the MJDREFI/F keywords. In their absence, the tools look for MJDREF, and failing that will look to see whether TELESCOP = GLAST. If that doesn't work, I guess the tools will complain and quit.
Observation simulation
No news, but we had some discussion of a tool that would be something like an extended version of ModelEditor that to the extent possible would translate between obsSim and likelihood-type XML model files. Such a tool would be useful; the translation aspect would not necessarily have to run in a GUI. The larger question seems to be about how to get it implemented via st_graph, which the GSSC is committed to support for distribution of the Science Tools to guest investigators.
User interface and infrastructure
No news. David reports that Cicerone still has some gaps, e.g., in pulsar tools and observation simulation. This is supposed to be in shape by early November for the GLAST Users Committee.
Source Catalog
The catalog group met this week and Jean presented results on understanding the reasons some DC2 sources were detectable but were not uncovered in the DC2 catalog analysis. I tallied the source lists that have come in for the comparison among source detection algorithms.
Science Tools Working Group
We met this week and most probably will meet again in 2 weeks, sharing the time slot with the Catalog group.
The current version of the Science Tools remains v7r4.
Data products: David pointed out that DATE-OBS and DATE-END (the human-readable start and stop times for the data in the file) in the FT1 files should be expressed in TT to be in conformance with FITS specifications. This is not a big change; none of the science tools interpret the date strings to find out times.
Databases and related utilities
No news.
Likelihood analysis
Jim reports that with Sabrina Casanova (LANL) some progress has been made on getting the EGRET response functions in shape to analyze EGRET data with the science tools, but a return visit from Sabrina probably will be needed. EGRET analysis will require something like an extended FT2 file that tallies the mode changes of the trigger telescopes. From a science tools perspective, most of the change is making irfInterface accept time as a parameter, and this capability may be useful to have ready anyway for LAT data analysis.
James and Analia have done some more work on implementing a 'badness of fit' indicator for gtlikelihood. They have been leaning toward flagging bins (energy ranges) where the deviation of the fit is greater than 2sigma from the data. James agreed to circulate a note about the options that they have considered. Intrinsically, likelihood analysis does not provide a goodness of fit, and certainly having an indication whether a given model just isn't working would be worthwhile.
James and Analia are continuing to work on the plotting in likelihood, most recently getting likelihood to set colors in the plots.
GRB tools
No news and I'm afraid still no update on the IRF normalization/gtrspgen issue reported by James a couple of weeks ago; I forgot to ask James at the meeting.
Pulsar tools
James and Masa report that they are commencing work on incorporating the MJDREF change in the pulsar tools. Most of the tools were not paying attention to the value of MJDREF at all - and so did not care when the definition changed. Now they will all read MJDREFI/MJDREFF. The most intricate work has been in making updated test data sets for gtpsearch. The MJDREF-related changes to the tools are expected to take a week or so - possibly longer than the next tag of science tools.
Over the next month or so (until the end of the current build cycle) Masa and James will continue to implement the changes to the interfaces to the science tools to allow multiple systems of time to be specified. This seems to be going well, although James mentioned a couple of bugs in the Windows versions that need to be tracked down.
Observation simulation
No news
User interface and infrastructure
No news.
Source Catalog
The catalog group did not meet this week.
Science Tools Working Group
Have not met for 3 weeks; most likely we will meet on August 9. Science tools development news is scant this week.
The current version of the Science Tools remains v7r4.
Data products: Jim implemented the MJDREF change described last week for FT1. He pointed out to me that this did not break gtbary (based on the code for the Chandra tool axbary) because internally it already checked for MJDREFI and MJDREFF keywords. gtpsearch may be broken by this change, but in general the tools seem remarkably insensitive to the disappearance of MJDREF. Jim also implemented other keywords that the HEASARC via David asked be included, among them redundant specification of the OBSERVER keywords; see this JIRA issue.
On Wednesday, David posted an updated definition version of the File Format Definitions for the Science Data Products that includes the additions and changes to FT1. FT2 is probably the next target for detailed scrutiny. David is soliciting comments on the data product definitions in general. Detailed-oriented good citizens are needed.
Databases and related utilities
No news.
Likelihood analysis
No specific news, but see Jim's recent summary of likelihood-related news since June.
GRB tools
No news, and I'm afraid no update on the IRF normalization/gtrspgen issue reported by James a couple of weeks ago; James is away this week.
Pulsar tools
From Masa: "Last week James and I finished converting all pulsar tools to use new time-handling classes. The latest tagged versions of the pulsar tools are now with the new infrastructure, yet their functionality didn't change since the latest release of Science Tools (although there are some minor differences in output message, etc.)."
"The plan is to implement new features (such as allowing a user to specify an epoch in MJD) based on this version."
Observation simulation
No news
User interface and infrastructure
No news.
Source Catalog
The catalog group met this week. Jean presented results on catalog analysis of the DC2 data using only the >100 MeV range; more sources than I would have expected were lost this way. We also discussed test files for source detection algorithms, and what would be a sensible agenda for the catalog session at the collaboration meeting.