Science Tools Working Group
We met last week (13 attendees) and probably will meet next week.
Since Jan. 11, the current version of the Science Tools is v9r4. It has the pointlike package, and omits gtorbsim, as have been previously described. It also includes for the first time the new gtptest pulsar tool. It also has a fix for simulations of binary pulsars. Other package updates since v9r3p4 are listed here . If you are a Science Tools user, you will want to upgrade. Note that you need to be careful to not mix livetime cubes from earlier versions with the v9r4 Science Tools (see below).
Data products: After a lot of work this past week, graciously and carefully updating the specifications of some of the Science Data Products to be consistent with current usage in the Science Tools, David has submitted the Science Data Products ICD and FFD for baselining. I'm not exactly sure what the process is. Several of the templates in the Science Tools for the headers of the data products needed updating in one way or another for consistency with HEASARC conventions and these are in progress or in some cases already complete. The detailed definition of LS-001 (the extended version of FT1) was left incomplete. This is understood to be an evolving data product in any case. David expects to rebaseline the documents before launch - this is a way of saying that many updates are expected, but they will be done in a formal way.
Other news: The GLAST Users Group will 'beta test' the Science Tools in early March. For this test the GUG members will be at their home institutions, i.e., no hands-on help will be available. Eric Winter and Brian Irby (GSFC) have succeeded in porting the Science Tools to the HEASARC hmake build system, which is how they will be distributed by HEASARC. The port is for v9r3 and currently the plan is to stick with it for the test. The Big Run data will not be ready by the time of the test; most likely the GUG members will be directed to the LEO 55-day data set (with some guidance about time ranges, etc.). In anticipation of the beta test, Analia Cillis has been assembling extended, current versions of our reference pages for the Science Tools in the User Workbook; Masa has been working on the pulsar tools versions of these pages. They anticipate circulating these more widely for comment soon, and delivering them to Chuck for incorporation in the Workbook.
Databases and related utilities
No news
Likelihood analysis
In v9r4 of the Science Tools the default coordinate system for the livetime cubes (HEALPix) was changed from equatorial to Galactic. In principle would not have to be a problem, since the coordinate system is listed in the headers of the livetime cubes. However, it turned out that the coordinate specification was not read in. Jean noticed the strong effect on exposure maps right away and Jim tracked down the problem. Toby has changed the default back to equatorial and I think has also made the code interpret the coordinate specification from the header.
GRB tools
No news.
Pulsar tools
No development news; last week Masa circulated a proposed updated format for the pulsar ephemeris database file. This incorporates parameters for additional binary pulsar timing models that David Smith pointed out we will want.
Observation simulation
Jim and Toby worked out a scheme to assign unique IDs to SourcePopulation sources. Until now, the faint blazars in the sky model have been specified as steady sources using SourcePopulation; this is a compact, convenient, and fast way to define theme but had the side effect of assigning only one ID to potentially thousands of different sources. This matters if you are studying source detection and the catalog analysis. The scheme that they worked out seems to work; Nicola's test run of the Ops Sim 2 sky model is evidence. The one-week gtobssim run has gamma rays from >18,000 unique sources.
As mentioned above, Max has implemented a fix to the simulation of binary pulsars.
User interface and infrastructure (& utilities)
Jim reported that the behavioral issues with APE in the hoops package are fixed in a more recent version of APE. James will import this - it may have already been done. APE is not yet in a release version of the Science Tools; waiting on a Windows build of the APE library.
Source Catalog
Did not meet last week.
Science Tools Working Group
We did not meet this week but probably will next Wednesday.
The current version of the Science Tools remains v9r3p4. The updates described below are not in an actual release yet.
Data products: Not much news. Julie has examples of LS-001 files - these are FT1++. They have the columns of FT1 plus about 200 more, representing all of the Merit variables that are used in the classification of events. The definitions of these columns will necessarily evolve some. The idea is that these will be the 'full event' data product that is provided to the GSSC. Julie wants to make sure the files are what she intended; when they are ok, and Warren gets the pipeline to make one of these files, Mali will be able to sign off on a lingering requirement. David needs to finish the specifications in the FFD soon - and I owe him some replies.
Databases and related utilities
No news
Likelihood analysis
No news.
GRB tools
No news.
Pulsar tools
Masa reports there's no news
Observation simulation
From Jim:
- fixed the memory leak in the Primack05 EBL attenuation model (celestialSources/eblAtten v0r5)
- fixed the 214 microsec sampling issue in SpectralTransient for the big run sky model (celestialSources/genericSources v1r9p5) [see comments on Nicola's posting on a test of the Big Run sky model; other artifacts from the limited precision of the random number generator can be found - e.g., the minimum non-zero time between photons in the diffuse model turns out to be 2 -25 seconds, for what it is worth.]
- disabled gtorbsim build (observationSim v7r2p1) [The interface to the flux package changed with Toby's rewrite - specifically how faint sources get disabled - and this broke gtorbsim. It probably could be reworked to work again but gttakosim is still our future.]
User interface and infrastructure (& utilities)
From Jim: "fixed DATE-OBS, TSTART, etc. keyword bug for empty FT1 files generated by makeFT1 (fitsGen v3r8p2)"
James has imported APE to replace PIL in the hoops package. It is in the LATEST builds of the Science Tools. Jim has reported a couple of bugs to James; one of them relates to why the unit test in the sane package has started failing.
Source Catalog
Met this week and covered a number of interesting topics. Jean gave an update on the comparison of the source detection algorithms, incorporating results from an improved PGwave algorithm. He also discussed his investigation of sources that the source detection algorithms find but the catalog pipeline discarded - evidence that another iteration step is needed in the catalog analysis. Toby gave a detailed update on what is new with pointfind and on development directions in the near future. He is also starting to generate the true source list from the sky model for the Big Run. This is a thankless task - so thank him.