convert quantities before they might be strings in to_table

[iximeow]

trying to u.Quantity with a string, for example, u.Quantity([m42.ra, '']), yields a TypeError:

    print(u.Quantity([m42.ra, '']))
  File "~/.local/lib/python3.8/site-packages/astropy/units/quantity.py", line 382, in __new__
    raise TypeError("The value must be a valid Python or "
TypeError: The value must be a valid Python or Numpy numeric type.

this is exactly what happens in to_table() with transitions or unused slots shown, because of the append append('') in scheduling.py.

so instead, do the u.Quantity conversion earlier where we know the type is (should be!) a Numpy numeric (or other appropriate type).

this is, in a way, a rewrite of #370: really just moves the change there up a few lines. i tested that it works by hand on a table with TransitionBlock; what would yield the aforementioned TypeError on master now produces a table that can be printed:

     target         start time (UTC)         end time (UTC)      duration (minutes)      ra         dec      configuration  
--------------- ----------------------- ----------------------- ------------------- ----------- ----------- ----------------
            M42 2021-02-14 01:30:00.000 2021-02-14 02:23:20.000  53.333333333333414 83.82208333 -5.39111111 {'filter': 'Ha'}
TransitionBlock 2021-02-14 02:23:20.000 2021-02-14 02:23:29.136 0.15226966667478692                                       []
             M1 2021-02-14 02:23:29.136 2021-02-14 03:43:29.136   80.00000000000003 83.63308333     22.0145 {'filter': 'Ha'}
TransitionBlock 2021-02-14 05:53:29.136 2021-02-14 05:53:59.688   0.509194929995882                                       []
            M51 2021-02-14 05:53:59.688 2021-02-14 07:13:59.688   80.00000000000003  202.469575 47.19525833               {}

i'm not sure how to run the tests here, or i'd go ahead and add a test that to_table should work with examples from the docs :D

[bmorris3]

This looks great, thanks @iximeow!