Unpredictable behaviour - Time.exposure harvest via casatools

[ebrassfieldbourke]

After updating main to harvest from splits directory measurement sets for separate sources and ingest them as planes, the casatools code harvests different exposure time values for primary target source from splits target ms and the average ms, which is not true based on the observation logs.
This needs investigating to see what is actually in the splits target measurement set or if our code is causing this. PROPER UNIT TESTS ! :)