Funny you're answering my question as I just saw your webcast on
TheServerSide earlier today and am, at this very moment, reading the preview
of your "Tamming Text" book early release from manning (Chapter 5 seems to
be what I'm really looking for...) .

Basically I really am a Lucene newbie and moreover an Information Retrieval
profane so I got a lot of things to learn before I can pretend to touch or
document anything from the spatial contrib ...

Basically for now, the few things that are really important for my project

1) Query parsing --> identifying geographic place names (the name of an
administrative division or a country) , POI/location names (the name of a
store for example), location categories, adjectives concerning a place,
etc ...
2 ) When needed, limiting the results to only those available in a given
bounding box or radius around a central location. This could be done when
the result set is too large.
3) Sorting of results based on a mix of relevance and distance from an
identified point that was identified in step 1).

I know that the spatial contrib has something that will let me do distance
sorting and BBox filtering ( 2 and 3 above )and that's what I'd really LOVE
to see in action.

As for query parsing I'm still reading chapter 5 so I hope It'll give me
some ideas for my project. OpenNLP seems to something I could use.

Anyway if anybody is interested in putting up a simple demo of how the
saptial contrib can be used or even any documentation that could lead me
(and others) on the tracks that would be fantastic ! :)

Thanks for your help!



Search Discussions

Discussion Posts


Related Discussions

Discussion Navigation
viewthread | post
posts ‹ prev | 20 of 20 | next ›
Discussion Overview
groupdev @
postedSep 17, '09 at 5:38p
activeNov 22, '09 at 4:07a



site design / logo © 2021 Grokbase