GammaLib

The ontime in GCOMObservation::load(GFilename&, GFilename&, std::vector<GFilename>&) is now extracted from the ontime of the associated TIM datasets. Before it was computed from the start and stop time, yet this neglects the cut-out of time intervals in the TIM due to the SAA passage.

The default usage when loading TIM files is now set to YES and the default mode to NORMAL. Before, both strings were blank by default.

In GCOMEventList::read(GFits&) the Good Time Intervals are now set on the basis of the times of the first and last events. The reason is that the VISDAY, VISTIM, VIEDAY and VIETIM keywords that were used before do not seem to related to the GTI of the events but rather relate to the full time period of the observation.

In GCOMOads::print(GChatter&) only print the number of superpackets for higher than default chatter levels.

Put COMPTEL support functions in gammalib namespace.

I added remove_response_cache methods to the classes GObservations, GObservation, GCOMObservation, GCTAObservation and GCTAResponse to allow the controlled removal of one or all models from an observation container. For this purpose, the GCOMResponseCache::remove() method was also implemented.

The verification of parameter changes was removed in the COMPTEL response caching, and a call to GObservations ::remove_response_cache was added in cttsmap so that cached values for the test source are removed before computation of the TS for the test source.

I added a GCOMObservation::drm(GModels&) method that computes a model DRM cube for all applicable models in a model container.

I also added a GCOMDri::cone_content() method to compute the DRI content within an event cube.

Moved issue to next release.

The response caching is now dealt with at the level of the GResponse base class, I therefore removed the COMPTEL specific response caching. Specifically the GCOMObservation::drm(GModelSky&) and GCOMResponse::irf_spatial() methods were removed.

In finally implemented support for extended model in the COMPTEL response, completing response handling.

Here is a test case of a SRCLIX analysis of the Crab v1.0 in the 4 standard energy bands. The run gives the following result for a point source

2021-10-13T14:07:00: +============================================+
2021-10-13T14:07:00: | Iterative maximum likelihood model fitting |
2021-10-13T14:07:00: +============================================+
2021-10-13T14:07:05:  logL after iteration 1 ....: -67027.48666
2021-10-13T14:07:10:  logL after iteration 2 ....: -67041.94875 (14.46210)
2021-10-13T14:07:15:  logL after iteration 3 ....: -67044.96521 (3.01646)
2021-10-13T14:07:19:  logL after iteration 4 ....: -67045.57367 (0.60846)
2021-10-13T14:07:25:  logL after iteration 5 ....: -67045.63165 (0.05797)
2021-10-13T14:07:29:  logL after iteration 6 ....: -67045.59895 (-0.03270)
2021-10-13T14:07:37:  logL after final iteration : -67045.59894 (-0.00001)
2021-10-13T14:07:37: === GModelSky ===
2021-10-13T14:07:37:  Name ......................: Crab
2021-10-13T14:07:37:  Instruments ...............: all
2021-10-13T14:07:37:  Test Statistic ............: 1004.16178071871
2021-10-13T14:07:37:  Observation identifiers ...: all
2021-10-13T14:07:37:  Model type ................: PointSource
2021-10-13T14:07:37:  Model components ..........: "PointSource" * "PowerLaw" * "Constant" 
2021-10-13T14:07:37:  Number of parameters ......: 6
2021-10-13T14:07:37:  Number of spatial par's ...: 2
2021-10-13T14:07:37:   RA .......................: 83.7860667134836 +/- 0.111312108719039 deg (free,scale=1)
2021-10-13T14:07:37:   DEC ......................: 21.6291010912198 +/- 0.101021895111379 deg (free,scale=1)
2021-10-13T14:07:37:  Number of spectral par's ..: 3
2021-10-13T14:07:37:   Prefactor ................: 0.0016763606983258 +/- 8.56368715168633e-05 [1e-25,infty[ ph/cm2/s/MeV (free,scale=0.002,gradient)
2021-10-13T14:07:37:   Index ....................: -2.14913178754043 +/- 0.0382371945279881 [-10,10]  (free,scale=-2,gradient)
2021-10-13T14:07:37:   PivotEnergy ..............: 1 MeV (fixed,scale=1,gradient)
2021-10-13T14:07:37:  Number of temporal par's ..: 1
2021-10-13T14:07:37:   Normalization ............: 1 (relative value) (fixed,scale=1,gradient)
2021-10-13T14:07:37:  Number of scale par's .....: 0

2021-10-14T12:50:58: +============================================+
2021-10-14T12:50:58: | Iterative maximum likelihood model fitting |
2021-10-14T12:50:58: +============================================+
2021-10-14T12:53:01:  logL after iteration 1 ....: -67037.18725
2021-10-14T12:54:10:  logL after iteration 2 ....: -67043.74861 (6.56136)
2021-10-14T12:55:46:  logL after iteration 3 ....: -67046.41870 (2.67009)
2021-10-14T12:56:50:  logL after iteration 4 ....: -67047.15837 (0.73967)
2021-10-14T12:57:38:  logL after iteration 5 ....: -67047.24089 (0.08252)
2021-10-14T12:59:12:  logL after iteration 6 ....: -67047.11076 (-0.13013)
2021-10-14T13:00:00:  logL after final iteration : -67047.11374 (0.00298)
2021-10-14T13:00:00: === GModelSky ===
2021-10-14T13:00:00:  Name ......................: Crab
2021-10-14T13:00:00:  Instruments ...............: all
2021-10-14T13:00:00:  Test Statistic ............: 1101.97154647962
2021-10-14T13:00:00:  Observation identifiers ...: all
2021-10-14T13:00:00:  Model type ................: ExtendedSource
2021-10-14T13:00:00:  Model components ..........: "RadialDisk" * "PowerLaw" * "Constant" 
2021-10-14T13:00:00:  Number of parameters ......: 7
2021-10-14T13:00:00:  Number of spatial par's ...: 3
2021-10-14T13:00:00:   RA .......................: 83.7858226805196 +/- 0.116796678591813 deg (free,scale=1)
2021-10-14T13:00:00:   DEC ......................: 21.6527349414119 +/- 0.105916502256037 deg (free,scale=1)
2021-10-14T13:00:00:   Radius ...................: 1.31224708888018 +/- 0.415734840183593 [0.05,10] deg (free,scale=1)
2021-10-14T13:00:00:  Number of spectral par's ..: 3
2021-10-14T13:00:00:   Prefactor ................: 0.00170719506442216 +/- 8.76326119024385e-05 [1e-25,infty[ ph/cm2/s/MeV (free,scale=0.002,gradient)
2021-10-14T13:00:00:   Index ....................: -2.15088249332866 +/- 0.0366685007302425 [-10,10]  (free,scale=-2,gradient)
2021-10-14T13:00:00:   PivotEnergy ..............: 1 MeV (fixed,scale=1,gradient)
2021-10-14T13:00:00:  Number of temporal par's ..: 1
2021-10-14T13:00:00:   Normalization ............: 1 (relative value) (fixed,scale=1,gradient)
2021-10-14T13:00:00:  Number of scale par's .....: 0

Here the summary for various extended models:

While the disk and gauss model seem to give reasonable results, the elliptical disk model does not seem to adjust the spatial parameters, the elliptical gaussian model looks kind of ok.

The COMPTEL interface is now settled.