Adapt HARP and OSKAR to handle rigidly rotated stations and complete EM simulations

The purpose of this feature is twofold, first is to work on HARP making sure that it can form station beams when using rigidly rotated stations. Currently all simulations are done with the assumption that stations and antennas are oriented towards cardinal points (N-S and E-W). However, the ECP-210113 informs that for SKA Low all the stations will be using rigidly rotated layouts. This poses a challenge in calibration on the station and array level. In order to test these, an ability to correctly simulate these cases is necessary. Currently OSKAR cannot simulate station or array with rigid rotation when using HARP data.

This work will ensure that HARP will be able to correctly form station beams and point them at required directions when using rigidly rotated layout. This requires modification to HARP beam library that is used by OSKAR as well as integration of corrected versions into the OSKAR.

Additionally, OSKAR can benefit largely from improving the way it handles input HARP data. Specifically, when loading HDF5 files holding Macro Basis Functions generated by HARP, OSKAR loads entire dataset to memory and copies it to number of stations listed in telescope model. That can be very burdensome on RAM and storage. The goal is to implement lazy loading that would speed up simulations.

 The last target is to finalise EM simulations of SKA Low station EEPs. Currently the simulations cover 50 - 252 MHz and 280 - 314 MHz out of 50 - 350 MHz necessary. During this PI the remaining frequencies will be simulated. The idea is to continue simulations going from 314 to 350 MHz range. Once done, simulations in 240 - 280 range will be done. This region is affected by RFI and can be done as last. Finally, during PI21 certain frequencies have been identified as problematic, i.e. the EEPs simulated are exhibiting unexpected shapes when compared to Curtin/INAF results. This also will be addressed during this PI.