Text Annotator: Tweak user-facing language provide user feedback mechanism on recognized entities

[jmcmurry]

• [ ] Short term fix: Once you get the annotated text, provide some text indication that the list of phenotypes can be tweaked afterward rather than just "analyze phenotypes"
• [ ] Longer-term fix: Provide thumbs up thumbs down feedback for each recognized term in situ.

[jmcmurry]

Here's how it should look when completed; this was done for Hippo publications browser which is no longer; however the basic user design applies and worked well at the time.

Right now, there is no way to tweak anything other than the list of recognized phenotypes. If you could use color codes for the kinds of entities recognized, that would help as well. Ideally, we should capture which recognized concepts the user validates as this can improve the algorithm, however, that is more of a long-range goal.

Any concepts which are recognized but not one of the above 5 types (or anatomy) should be ignored I think