Actually, it is supposed to be [1, 1]. The actual value of the element corresponds to the score it has, and the index of it refers to the index in the data.
The article doesn't exactly simulate what Wade returns, but uses a simpler version. I added a note about it now.
[1, 1]
The item at index 0 is 1. This means that the item in index 0 of the data ("Hey") has a score of 1. The item at index 1 is 1, meaning that the item at index 1 of the data ("Hello") has a score of 1.
Edit: Your example with Wade isn't working because it ignores the first item in the query ("he") as it is a stop word. Searching for "h" will return both.
I'm still curious about using it. The GitHub example [1] works as described, but I can't get a result from the 'he' example. Is it saying no match was found by returning an empty array?
EDIT: Okay, saw your edit. Might be nice to use an example that's not a stopword. ;) Nice work on the library and the blog post.
Cool, glad you understood. The library returns an array of results, and will return an empty array if there are none. Try searching for "h" or "hey" for an example.