So after I index a document, is it my job to parse the contents of a field, split it in words/tokens and implement logic to specify autocompletion options for each word and index them? Again, I don't think it can be this way: ES has the power to auto-correct words, but it doesn't have the power to auto-complete them?!
I would expect ES to be able to do that once I indexed the document (probably I would have to specify some analyzier or whatever)...