Read multiple datasets with one call to `cfdm.read`

[davidhassell]

It would be nice to be able to read multiple datasets in one cfdm.read call, eith as a list, a list implied by globbing, or anything in a directory, possibly including its subdirectories:

>>> f = cfdm.read(['file1.nc', 'file2.nc'])  # list
>>> f
[<Field: specific_humidity(latitude(5), longitude(8)) 1>,
 <Field: air_temperature(height(1), grid_latitude(10), grid_longitude(9)) K>]

>>> f = cfdm.read('file[12].nc')  # glob
>>> f
[<Field: specific_humidity(latitude(5), longitude(8)) 1>,
 <Field: air_temperature(height(1), grid_latitude(10), grid_longitude(9)) K>]

>>> f = cfdm.read('~/directory/', recursive=True, followlinks=True)  # directories
>>> f
[<Field: specific_humidity(latitude(5), longitude(8)) 1>,
 <Field: air_temperature(height(1), grid_latitude(10), grid_longitude(9)) K>]

This is exactly how cf-python does things.