#POSTGIS IRC Log - 2008-08-25

For logs after Feb 3, 2007, all times are GMT-8. Prior logs are GMT-9.
Back to Logs
01:07:37 magnush: Anybody have a source tarball for 1.3.1 around? Seems the DL site is down - or at least unreachable from here...
05:04:53 unlord: hi guys
05:05:15 unlord: does postgis use a hierarchical set of boxes when indexing a point?
06:16:37 grobda24: unlord ... yes.
06:28:51 unlord: cool that is what I thoguht
06:28:58 unlord: i'm looking at locallucene and I'm not sure that it does
06:50:00 grobda24: unlord, you mean http://en.wikipedia.org/wiki/R-tree ?
06:50:01 sigq: Title: R-tree - Wikipedia, the free encyclopedia ( at en.wikipedia.org )
06:54:47 unlord: grobda24: yep
06:54:53 unlord: thats exactly what I am wondering about
06:57:12 unlord: grobda24: as far as I can tell, it does not
07:20:15 darkblue_B: you guys trying to figure out the internals of the PostGIS spatial indexing?
07:20:33 darkblue_B: .. a little light reading before lunch? :- )
07:20:36 mloskot: then better read GiST docs
08:02:33 springmeyer: mloskot: are you around?
08:03:27 mloskot: a bit
08:03:45 springmeyer: cool, can you dpaste the output of this...
08:03:55 springmeyer: ls /usr/lib/libboost* -l sort
08:04:11 mloskot: ah, I got it :- )
08:04:39 mloskot: perhaps I should join osm channel?
08:04:53 springmeyer: #mapnik would be good
08:05:07 : * springmeyer does not know about #osm, goes to look...
08:05:52 mloskot: springmeyer: I just known Artem was lurking on #osm channel but it's on different IRC server. I didn't know about #mapnik
08:05:54 mloskot: anyway, I'm there
08:06:28 springmeyer: ya, I started #mapnik last week with help from crshmidt
08:06:40 : * springmeyer has never bumping into artem on irc
08:06:51 mloskot: I have on #osm but months ago
08:08:26 melter: is svn.refractions.net down?
08:14:27 cgs_bob: melter: I was about to ask the same thing
08:21:47 mloskot: melter: works for me
08:21:48 mloskot: http://svn.refractions.net/fdopostgis/
08:21:49 sigq: Title: Revision 425: / ( at svn.refractions.net )
08:23:56 melter: when i go to http://postgis.refractions.net/documentation/ and click the "Release Change Log" link, it just hangs waiting for svn.refractions.net
08:23:58 sigq: Title: PostGIS : Documentation ( at postgis.refractions.net )
08:25:55 melter: same for README.postgis
08:27:27 mloskot: melter: you are right, it doesn't work
08:37:03 darkblue_B: postgis svn was down 12 hours ago for me
08:51:10 bcrosby: how would I return all records that are within a radius from a lat/long?
08:53:17 iggy: I think there are actually a couple of different ways to do it
08:53:17 bcrosby: yeah
08:53:17 bcrosby: just trying to think of the easiets
08:53:17 darkblue_B: thats a FAQ
08:53:18 bcrosby: really?
08:53:18 bcrosby: let me check the site
08:53:37 darkblue_B: the new manual is coming along
08:53:55 iggy: st_within( addr_point, st_expand( ST_SetSRID( ST_MakePoint( $lat, $lon ), 4326 ), $radDeg ) )
08:54:04 iggy: I use something like that
08:54:08 bcrosby: yeah
08:54:19 bcrosby: my problem is figuring out $radDeg
08:54:33 darkblue_B: expand... thats not a distance in meters...
08:54:48 darkblue_B: or miles.. kind of .. hmm not intuitive for most situations
08:54:52 bcrosby: darkblue_B: correct, that is degrees
08:54:57 iggy: I use a fairly simplistic method right now because it seems to work okay: $radDeg = ( float )$_REQUEST['mileradin'][0] / 69.05;
08:55:13 bcrosby: luckily, im working with nautical miles.. so 1nm = 1 minute of degrees
08:55:29 bcrosby: ( approx )
08:56:44 bcrosby: one thing you COULD do
08:57:06 bcrosby: is convert it to a projection that uses meters as the measurement
08:57:12 darkblue_B: the new manual .. tha nswer is section 3.7
08:57:14 bcrosby: then $radDeg is just your radius in meters
08:57:23 darkblue_B: it uses meters
08:57:44 iggy: yeah, that option was mentioned when I was putting this together
08:58:16 darkblue_B: yeah.. same in 1.3.2 manual
08:58:45 darkblue_B: ohh but the content is different
08:59:08 bcrosby: uhmm
08:59:19 bcrosby: that example is in meters because it assumes the co-ordinate system is in meters
08:59:30 bcrosby: but the doc page for ST_DWithin says "units"
08:59:41 darkblue_B: yeah.. units..
08:59:49 bcrosby: thats cool, I got the answer I was looking for
09:00:26 darkblue_B: oh so the new manual example is not to good actually, because they dont explicitly mention that assumption
09:14:42 jlivni: note that expand builds a box around your pt. you may want to use buffer instead if you want more like a radius search
09:19:29 bcrosby: thanks
09:30:09 bcrosby: I dont suppose Paul Ramsey is in this channel
09:33:22 bitner: nope he's not
09:33:52 bcrosby: hmm ok
09:34:09 bcrosby: I'm having a hard time trying to represent a plygon in my db, and he previously helped me
09:36:13 bitner: well -- as in most irc channels -- ask away
09:36:21 bitner: worst that'll happen is silence
09:36:21 bcrosby: the geometry in question can be found at www.blakecrosby.com/postgis.txt
09:36:53 bitner: heh, have fun with that :- )
09:37:23 bcrosby: I know
09:37:51 bcrosby: I can get the arc/circle part down
09:37:52 bcrosby: and the rest
09:38:00 bcrosby: but when I try to use ST_Union, I get an error
09:44:42 darkblue_B: bcrosby: is it *1* geometry.. or a method in general ?
09:45:27 bcrosby: it is one geometry
09:45:33 bcrosby: let me show you how its represented on a map
09:45:34 bcrosby: stand by
09:46:05 darkblue_B: well heck, just plot out on graph paper and synthesize ( make-up ) the points on the curve
09:46:16 bcrosby: lol
09:46:21 bcrosby: I have over 500 of these
09:46:57 bcrosby: www.blakecrosby.com/map.jpg
09:47:48 darkblue_B: I think you need a new approach
09:48:53 darkblue_B: you said you have the arc part working? but you cant add it to a non arc part?
09:49:24 darkblue_B: .. that seems fishy.. not being able to add together
09:49:51 bcrosby: yeah
09:49:53 bcrosby: this is what I get
09:49:54 bcrosby: NOTICE: TopologyException: found non-noded intersection between -80.2853 42.85, -80.3083 42.8458 and -80.3239 42.8647, -80.275 42.8172 -80.3051 42.8464
09:50:07 bcrosby: so i'm still trying to figure it out
09:50:24 darkblue_B: oh TopologyException
09:50:33 darkblue_B: I've gotten stuck on that
09:50:48 darkblue_B: its GEOS balking in its innards
09:51:34 bcrosby: yeah
09:51:35 darkblue_B: I have the JTS sources here, and run the test case GUI program to test these thigns out sometimes
09:51:50 darkblue_B: you have the source to your UNION handy?
09:52:28 bcrosby: its a mess
09:52:30 bcrosby: ST_Union( ST_Transform( ST_Buffer( ST_Transform( ST_GeomFromText( 'POINT( -80.285278 42.85 )', 4269 ),26910 ),4*1609 ),4269 ), ST_GeomFromText( 'POLYGON( ( -80.275 42.817222,-80.285278 42.85,-80.308333 42.845833,-80.323889 42.864722,-80.275 42.817222 ) )',4269 ) );
09:53:31 bcrosby: oh hmm
09:53:34 bcrosby: maybe this is my problem
09:53:45 bcrosby: ST_ISValid says "false" on my polygon
09:54:06 bcrosby: ahhhh
09:54:28 bcrosby: its because the polygon intersects it self on one of its sides
09:57:47 darkblue_B: ah hm
09:59:14 darkblue_B: you want to buffer in 26910 units.. but union in 4269 / Geodetic?
10:02:41 darkblue_B: ugh.. no copy and paste in the JTS java app
10:15:41 darkblue_B: bcrosby: yeah.. Self-intersection.. IsValid( ) = false
11:25:09 mloskot: hobu: perhaps you could also copy the post-commit hook for buildbot into geos repo?