[Xapian-discuss] Statistical query completion

[Olly Betts]

On Mon, May 04, 2009 at 04:30:16PM +0200, Alexander Staubo wrote:
> Given the sequence of terms ["I", "like"], I would to discover the
> possible completions ["I", "like", "icecream"] and ["I", "like",
> "chicken"] ranked appropriately.
> 
> Does Xapian support something like this, or will I have to build my
> own Markov chain model?

We don't currently store the information in a suitable way for computing
this at all efficiently - you might as well start from the original
text.

This may change in the future (but not imminently) as such information
would probably help produce better spelling correction suggestions (the
most likely correction may vary depending on the other words in the
query).

Cheers,
    Olly