CSET_*_CYCLE_POINT variables clash with general definition of cycle point

[jfrost-mo]

Describe the bug

By convention a forecast cycle refers to its initiation time (as has always been the way in the RES) thus this is what many will instinctively think of when setting any of the data with CYCLE in the name.

Currently in CSET the CYCLE variables refer to the variable being CYCLED over in Cylc. This is not clear, either the naming convention in CSET needs to change or the documentation should be improved to provide this clarity. I would suggest the former.

Note

This may become more obvious when multiple forecast cycles can be added into CSET, but will be important for when running RAL trials, and is vital for me [@daflack] switching between Services and RMED to aid the transitions each month.

FURTHER COMMENT: cycling via datatime could fix this and reduce this confusion.

[jfrost-mo]

Fixed in #765.