• Adds UnitsDataArray.open class method
• Changes create_data_array to create_dataarray to be consistent with xarray naming convention in methods.
• Deprecated create_dataarray in favor of UnitsDataArray.create

This feature adds the ability to symmetrically go between UnitsDataArray -> netcdf -> UnitsDataArray. The goal is move away from pickle in the lambda function interface and more generally, provide users a way to export/import binary outputs from/to PODPAC.

Both creare_dataarray and open_dataarray could probably be classmethods, like we have done in the Coordinates module. It may also be useful to export and import methods on UnitsDataArray that uses the to_netcdf and open_datarray in the background.
Done in e9abbc4

Example

# Eval with simple example
from podpac.data import Array
from podpac import Coordinates, UnitsDataArray
import numpy as np

# mock data
data = np.random.rand(5, 5)

# create native coordinates for data
lat = np.linspace(-10, 10, 5)
lon = np.linspace(-10, 10, 5)
native_coords = Coordinates([lat, lon], ['lat', 'lon'])

# create Array DataSource node
node= Array(source=data, native_coordinates=native_coords)

# evaluate
output = node.eval(node_array.native_coordinates)

# export
output_netcdf = output.to_netcdf()

# import
output2 = UnitsDataArray.open(output_netcdf)

# goal:
# output2 == output