Is it possible to modify the next example that the annotator sees according to its previous annotations?
For example, if the annotator has already annotated Elon Musk as a person in the first example, and this name appears again, say in the third example, is it possible to modify the example in a way that he won't need to annotate it twice?
Hi! So just to make sure I understand the question correctly: if the annotator highlights an entity span in one example, you want your annotation workflow to assume that all mentions of that span (in different examples) are also mentions of the same entity, and highlight them automatically?
I guess you could set
"instant_submit": True in your recipe so that each annotated example is sent back to the server immediately (instead of in batches). In your
update callback, you could then check the
"spans" highlighted in the example and add their text and label to a list. And for each incoming example in your stream generator, you could then check whether the text contains one of the previously added known entities and if so, prepopulate the
"spans" with the start/end/label of the known entity that occurs in the text.