Delay calibration + unified flagging

[jmoldon]

• I have combined the two run_aoflagger functions into one. There are fundamentally the same, the only change is using -fields or using finding the /SUBMSS structure. Now the single function works for any aoflagger version. Also, this simplifies the main script considerably.
• I have separated the ms2mms part from aoflagger part to simplify main script. There are still two ms2mms functions.
• I have added a delay calibration executed on all calibrators and a plot of the delay solutions that are plotted to a png image. @jradcliffe5 please add the parameter do_delay to the interactive input list.
• I already noticed some time ago that aoflagger was doing strange flagging on MMS when the separation axis was 'auto', and more reasonable results when separation axis is 'baseline'. I will open an issue so we can investigate this in more detail and with more data sets.
• For some reason I don't get exactly the same results on the flagging on ms2mms data and ms2mms_fields data. There are some small differences, but I don't see what could be the cause. As we discussed we should maintain only the direct ms2mms approach (with aoflagger 2.9), but in any case we need understand why this is happening.

[jradcliffe5]

Nice work Javier. Done the GUI incorporation and checked that this works on mac architecture (I cannot sleep atm...). Think we can merge this, as you have set an issue #36 for comments raised here