read_ms function would benefit from more flexible column name options

[rlbyrne]

Some OVRO-LWA pipelines write model visibilities to an ms column called "MODEL_DATA". pyuvdata currently only supports a column called "MODEL." It would be helpful to have more flexibility in column naming.

[kartographer]

@rlbyrne -- poking around w/ casacore a bit more (and a few other data sets), it actually looks as though MODEL_DATA is what we are supposed to be writing to, and that this is actually a long-standing error on our side. I did see that you've got a branch open w/ some changes, but I'm happy to integrate this into #1391 since I'm doing a lot of updates to various MS-handling tools, since I think in the long term we'd likely want to switch to using MODEL_DATA instead of MODEL.

[rlbyrne]

@kartographer Yes, please do integrate with #1391. The branch I created is currently just a hack-y solution to make sure I could run some OVRO-LWA data.

[kartographer]

@rlbyrne done!