Reading the SubarrayDesciption

[kosack]

That info is stored in the TelescopeConfig_N branches which are of type Sash::TelescopeConfig. However, it seems they have custom streamers, or at least the DSTs do not contain the information to decode them.

>>> run["TelescopeConfig_5"].show(interpretation_width=60)

name                 | typename                 | interpretation                                              
---------------------+--------------------------+-------------------------------------------------------------
TelescopeConfig_5    | Sash::TelescopeConfig    | AsObjects(AsDynamic(model=Model_Sash_3a3a_TelescopeConfig)) 

That seems to mean we cannot read them, as re-implementing the streamer for this would be a lot of work (it has a lot of sub-classes that describe e.g. the telescope ground position, the pixel positions, and other aspects of the telescope and camera).

[kosack]

In fact, Sash::TelescopeConfig does indeed have a fully custom streamer implementation, so we could reverse-engineer that if we want, however it's 212 lines long, so maybe not worth the effort.

[kosack]

I also remembered that the FITS::EventListWriter writes out the full instrument description in to the FITS file, I think including the camera pixel info (at least I had implemented that in the past), so we could just dump one run, and then just put the resulting FITS HDUs into this module. They probably are even pretty close to the ctapipe format already. That would still make the assumption that nothing changes between runs, but I think even between the HESS-1 and HESS-1U switch, the pixels are the same, just the electronics changes.