Working selector and defaults for InputLiteral using common reducers

[jermnelson]

On the way to finishing ticket #412, this PR uses Redux to set defaults based on Resource Template ID and URI of property using more generic reducers and a single getProperty selector to map the redux state to the InputLiteral's props.

Fixes #390
Fixes #391
Fixes #428

[coveralls]

Coverage decreased (-0.7%) to 76.537% when pulling 0d660f3 on inputliteral-selector into b36d952 on master.

[ndushay]

again, I have eyeballed, and don't feel qualified to approve or not.

I notice we seem to have tests with vague descriptions that have multiple expect statements. I think it's a better practice mostly to avoid tests with multiple expects, unless they are multiple examples showing the same result ("returns true for valid JSON"), or multiple aspects of the same thing ("has exactly the expected RDF contents"). In the cases I call out below, the description of the tests is vague, and the specific expect statements do lend themselves to better descriptions of specific behavior.

Clearly not something we're going to comb over our existing tests to fix ... but it can be a really helpful practice when someone needs to edit the code 6 months from now and is wondering what the tests are supposed to be proving.

[jgreben]

LGTM

[ndushay]

thanks for changing those test descriptions - I much prefer the new ones; hope it wasn't too time consuming. My question about that one test still stands -- is it testing what it says it's testing?

I'll leave it up to osh to merge; I'm not going to the mat over my remaining question.