add ability to retrieve text from CTS URN to use as the body of the annotation

[balmas]

For the characterizations, I think we are going to need the ability for the user to enter the text along with the CTS URN/URI of the text in which it occurs. This is because we don't have a nice easy way for them to select that from the source text right now and get it automatically calculated for them as a text selector. (i.e. they'll want an annotation body which is "some descriptive text" from http://data.perseus.org/citations/urn:cts:greekLit:tlg0012.tlg001.perseus-grc1:1.1 ) where "some descriptive text" is expressed as a TextSelector.

[balmas]

revisiting this, what we would like is the ability to have a typeahead (as in the sparrow code used for the treebank input form in perseids-clinet-apps) where user find the primary source by typing it, and then the cts api used to retrieve the text, display for the user to select, and then the urn + selector created by Plokamos in the annotation body

[balmas]

but of course we need a way to still create these annotations manually (as described initially) if the primary source isn't available on our cts endpoint