ctools

I recently remove the CTA,HESS,MAGIC,VERITAS instruments from the XML file, but apparently this has a side effect. I’ll try to fix the code that it works without that string.

It appears that ctmodel needs either an instrument of CTA or no instrument attribute to work correctly.

This happens probably because ctmodel creates an observation with the instrument set to CTA in ctmodel::get_obs.

The issue comes in fact from a change in ctbkgcube::run where before the tool wrote always the attribute "CTA,HESS,MAGIC,VERITAS" into the XML file while now it collects the instruments for all input observations in the instrument attribute using the following code:

            // Collect instrument identifiers
            if (instruments.length() > 0) {
                instruments += ",";
            }
            instruments += m_bkgmdl[i]->instruments();

Now the issue still is that ctmodel does not know that a HESS observation is handled. ctmodel should probably extract the instrument from the counts cube TELESCOPE keyword.

I manually changed the TELESCOP keyword in an event list to HESS and ctbin then created a counts cube with the TELESCOP keyword set to HESS. So it appears that the issue is that ctobssim does create an event list with the TELESCOP keyword set to CTA.

So far ctobssim always creates a CTA observation. In principle, the instrument type can be inferred from the response function FITS file TELESCOP keywords, and the code should in fact check that the instrument for an observation and the response function is consistent. This needs to add an instrument attribute to the IRFs, which implies an interface change, hence I move the issue to version 1.4.0.

I decided to put back the setting of the CTA,HESS,MAGIC,VERITAS attribute in ctbkgcube, and do this in the bugfix release 1.3.1.

I created issue #2152 to request a proper taking into account of non-CTA instruments.