[RTA/DPA BW tests]: Don't run sprucing test over a compressed file, or handle it better

Last remaining point from #35 (closed).

read_event_numbers has a horrid hack for reading the number of events in a compressed .RAW file. We should probably never run this over compressed files: https://gitlab.cern.ch/lhcb/Moore/-/blob/master/Hlt/Hlt2Conf/python/Hlt2Conf/tests/bandwidth/read_event_numbers.py#L53-63

def _count_evts(ifiles, file_type, evt_max):
    """
    Count all events in a list of MDF/ROOT files

    HACK
    Dont know how to easily read the number of events in a .raw file,
    Hack it for now - use evt_max (will only currently be .raw for spruce
    nominal, which has a capped EVTMAX).
    Once we can actually count these, remove evt_max as arg here.
    """
    if file_type == "RAW":

When the input file for Moore_spruce_bandwidth is next updated, we should use the uncompressed HLT2 output so that this can be avoided.

Alternatively, we find a safer way to count the number of events from a compressed file, but that sounds like a bad idea to me.