ctools

Implemented. ctbkgcube now creates a model container where the original background models used for the computation have been replaced by a new background model that can be used for stacked analysis. The container can be retrieved using the models() method, or it can be saved as an Xml file using the save() method.

Note that the current code will strip any CTA, HESS, VERITAS or MAGIC background models from the input model container. This will pose problems when using the code for a joint CTA, HESS, VERITAS or MAGIC analysis. To do this properly we would need to track which background components have been used for background modeling, and remove only those components from the container.

Thinking more about this, I’m not sure as this is really an issue. The current stacked analysis just stacks everything together in a single cube, hence it would also combine any joint instrument data into a single cube. This sounds odd, but that’s exactly what would happen: HESS data would be lumped together with MAGIC data, and the instrument response functions would be averaged. The same is true for the background, just everything would be lumped together. Hence the implementation is okay.