-
The mc() method went in endless loops as the exit condition of the while loop was inversed (the method continued looping when the energy was within the valid range and only exited when it was outside). The eval_gradients() method computed the gradients with respect to the parameter values and not the parameter factors. All gradients have be multiplied by the scales. Also, the gradients are now set to 0 if the parameters are fixed. Finally, the parameter "Prefactor" was renamed to "Normalization" because the parameter gives the integral flux under the Gaussian, not the prefactor.
-
…ralGauss mc method as suggested
-
Dates were appended in the middle of a text block if the buffer was flushed before a \n character was reached. A private member m_linestart has been added that tracks whether the buffer is actually at a line start or the middle of a text block. Dates are only set at the beginning of the buffer is the buffer is at a line start.
-
The code has been reformated after changing the class names from GPar to GApplicationPar and GPars to GApplicationPars. The renaming has been logged in the ChangeLog file.
-
(I chose `centre` because there were more `centre` than `center` around.)
-
The coding and design conventions (CDC) document was also updated.
-
The curvature matrix, which is equal to one-half time the Hessian matrix, is evaluated and not the covariance matrix. To make this clean, all covar terms have been renamed to curvature.
-
Conflicts: doc/source/user_manual/modules/model.rst
-
I ran this build: CXXFLAGS="-Wall -Wno-c++11-long-long -pedantic -std=c++98" ./configure make -j 1 This commit fixes the few compiler warnings that showed up (all unimportant). There is no way to fix the issue that `long long` is not standard C++98, but it's an extension that has been supported by most compilers for a decade, so it's OK to ignore: http://stackoverflow.com/questions/4017621/how-to-do-portable-64-bit-arithmetic-without-compiler-warnings
-
The likelihood computation has been removed from GObservations::likelihood and has been put in the virtual GObservation::likelihood method. This method implements standard likelihood computations for binned and unbinned analysis. It may be overloaded to implement observation specific likelihood computations.
-
Initially, there was a unlogic relation between GModels and GOptimizerPars, where GModels derives from GOptimizerPars, but both classes contained different types of objects. GModels is a model container, while GOptimizerPars is a parameter container. Now, GModels derives directly from GContainer and no longer from GOptimizerPars. The GModels::pars() method has been added to extract a flat array of model parameter pointers into a GOptimizerPars object. GOptimizerPars, which was abstract before, is now a real container class that also derives from GContainer.