- From: Gregory Williams <greg@evilfunhouse.com>
- Date: Sun, 16 Apr 2006 11:48:23 -0400
- To: Christopher Schmidt <crschmidt@crschmidt.net>
- Cc: Luk Vloemans <luk.vloemans@student.uhasselt.be>, semantic-web@w3.org
On Apr 16, 2006, at 8:53 AM, Christopher Schmidt wrote:
>> Another interesting query would be "Give me all pictures I took in
>> city X".
>> In order to solve this puzzle, one would have to know the exact
>> spatial data
>> of city X (a bounding rectangle would not be enough) and be able
>> to pinpoint
>> Lat/Lon within that range, using a RDF-querylanguage... I think
>> that's a
>> hard one to answer for.
>
> In this case, again, you'd take your post-processing solution.
> However,
> I think that conflating semweb/GIS databases is a bad idea. Instead,
> what I would personally do is take each geo:Point in my crawled
> database, store it in a PostGIS database with either the URL or bnode
> identifier. You can then submit queries against the PostGIS
> database for
> lat/long queries, and then look up the data in the RDF datastore
> for the
> actual content that you want.
Strangely enough, I've been working on exactly this during the past
week. I've been working on a patch to RDF::Query (perl sparql
implementation) that would allow spatial operators to be used in a
SPARQL FILTER, and have those queries be directly translated into
queries for a PostGIS database containing both the RDF data (using
Redland's PostgreSQL schema) and geographic data (maintained by
triggers on the RDF data). I haven't got much working at this stage,
but I agree that this is quite an interesting area to be looking into.
As for querying photos based on city, I've found it easiest to spend
the time pre-processing the photo data during import to add the city
names ahead of time. In this way, I'm able to query photos with
something like this:
SELECT ?place ?img
WHERE {
?img a foaf:Image ; dcterms:spatial [ a geo:Point ; cyc:inRegion
[ foaf:name "Boston, MA" ] ] .
}
This pre-processing is very easy using something like PostGIS for
areas where good spatial data is available (Tiger/LINE and the
USCensus ZIP Code Tabulation Area data sets come to mind), but I've
had to resort to a "name of the closest known point" strategy for
photos outside of these areas.
hope that helps,
.greg
--
"Writing code on one line is like playing
the trumpet without breathing!"
- Adam Pisoni
Received on Sunday, 16 April 2006 15:48:34 UTC