Do not assume a beam if one is not given

[MatthewWhiting]

If the beam information is not in the FITS file, the alternative behaviour is to use the beamFWHM etc parameters. However, if these are not provided, it uses a default size.

Tobias has pointed out that this default behaviour is actually to produce incorrect integrated fluxes. A better behaviour is to not calculate the integrated fluxes (perhaps just the total flux).

Method:

• If no beam info, check beamFWHM etc
• If these are the default (eg. negative values), set a flag to false
• Only calculate integrated fluxes if this flag is true

[MatthewWhiting]

Some observations:

• Best to do this in the readBeam function
• In calcIntegFlux(), we don't need the check at the end against isWCS() - the wcs status doesn't matter here
• Clean up a bit by having a FitsHeader::correctForBeam() function that calculates these things, with a bool member flag that records whether the input pars have been set. This could be defined in the readBeam function
• Make use of correctForBeam() in the units calculation as well as the calcIntegFlux()

[MatthewWhiting]

This is just about done, using the new Beam and DuchampBeam? classes that we put in FitsHeader? & Param.

It works quite nicely. The only thing left to look at is whether we should not bother trying to read the beam info from the FITS file if the units are not eg. Jy/beam (ie. if they are, say, Jy/pix, then we don't need the beam, and we don't want to complain unnecessarily if the beam info isn't present.

This used to be the case with needBeamSize(), but I changed it and removed it from the reading-in bit, so revisit that...

[MatthewWhiting]

Looking again at this - I think the degree of warning that we have in 1.1.11 is fine. The only time anything is printed is if there is no beam in the FITS file *and* the user has requested one of beamFWHM or beamArea (in which case it just notifies the user that that parameter is being used instead).

This ticket can be closed.