00:01:29 <jfreeman> G'day, any TileMill expertise online?
00:01:52 *** jfreeman has quit (Quit: Konversation terminated!)
00:02:11 *** jfreeman (~jfreeman@mail.agileware.net) has joined #mapnik
00:03:05 <jfreeman>  G'day, any TileMill expertise online?
00:07:28 *** ajturner has quit (Quit: ajturner)
00:15:13 *** ajashton has quit (Remote host closed the connection)
00:15:40 <springmeyer> jfreeman: go for it
00:15:42 *** chad_burt has quit (Quit: Leaving...)
00:17:00 <jfreeman> g'day springmeyer. The problem I have is with Tilemill, submitting a job. I've been unable to set the key pair, using the -k parameter. Like: tilemill.py -k /home/jfreeman/tilemill/jfreeman.pem start
00:17:41 <jfreeman> I receive a xyz.pem does not exist error
00:17:48 <jfreeman> but it's there
00:18:26 <jfreeman> any ideas?
00:19:32 <springmeyer> jfreeman: ya, I'm not the developer on Tilemill but I've run it and hit that same thing
00:19:41 <springmeyer> http://dpaste.com the exact whole error?
00:19:52 <jfreeman> ok
00:21:09 <jfreeman> http://dpaste.com/154326/
00:22:06 <springmeyer> my sense is that the keypair is supposed to go into the .boto
00:22:12 <springmeyer> and that the -k may not work
00:22:22 <springmeyer> have you filled in your ~/.boto creds?
00:22:37 <jfreeman> yep
00:23:02 <jfreeman> I can execute the tilemill.py start
00:23:13 <jfreeman> and that starts the AWS EC2 instance
00:23:40 <jfreeman> problem is I have little feedback as to what the instance is doing
00:24:00 <jfreeman> so I wanted to start the job again with my private key so I can ssh into the box and check the logs
00:24:08 <jfreeman> jfreeman@aspire:~/tilemill-ec2> tilemill.py status
00:24:10 <jfreeman> The input_queue (/893824549119/agileware) contains approximately 1 messages
00:24:11 <jfreeman> The output_bucket (outputbucketmapbox) contains 0 keys
00:24:17 <springmeyer> okay, right I've now refreshed my memory...
00:24:28 <jfreeman> That's all I get now when querying the current running AWS EC2 instance, not much info.
00:24:38 <springmeyer> jfreeman: yes, thats as far as I have gotten
00:25:11 <springmeyer> my instance has booted but there is no sense of anything else happening or any way to figure out whether the rendering jobs have started, errored out, or finished
00:25:12 <jfreeman> so I've got this large EC2 instance, running, doing something supposedly
00:25:21 <jfreeman> yes, correct
00:25:26 <springmeyer> yep
00:25:26 <springmeyer> so, sorry I can't help you
00:25:42 <springmeyer> other than to say that I ended in exactly the same place...
00:25:52 <springmeyer> when the lead developer was helping me
00:25:55 <jfreeman> bugger, thanks for listening anyway. At least I know I'm following the same trail as you
00:26:03 <springmeyer> ya
00:26:10 <springmeyer> I hope I am wrong, but...
00:26:26 <springmeyer> my sense is that few have traveled here aside from those that created it
00:26:36 <jfreeman> right
00:26:41 <jfreeman> bugger
00:26:43 <springmeyer> "here" being as far at `tilemill.py start`
00:27:21 <tcarobruce> oh hey -- just catching up  on tilemill action
00:27:46 <tcarobruce> I contacted the developer -- had some similar issues
00:27:53 * jfreeman listens
00:27:56 <springmeyer> jfreeman: that said, perhaps you can just do the rendering on your machine like 99% of us...
00:28:11 *** jfreeman has parted #mapnik ("Konversation terminated!")
00:28:20 <tcarobruce> if you start a machine up with a key set, take a look at /var/log/boto.log
00:28:23 *** jfreeman (~jfreeman@mail.agileware.net) has joined #mapnik
00:28:36 <jfreeman> sorry hit wrong button
00:28:38 <jfreeman> damnit
00:28:51 <jfreeman> springmeyer: how do you render on a local pc?
00:29:06 <tcarobruce> it looks like it tries to use pyami (part of boto) to download a few scripts from an S3 bucket that's not world-readable
00:29:09 <jfreeman> springmeyer: I'm running Linux, if that helps
00:29:26 <tcarobruce> and that's where it hangs
00:29:28 <springmeyer> jfreeman: lots of ways, depends on your goal
00:29:40 <jfreeman> tcarobruce: any idea how to fix?
00:29:50 <springmeyer> tcarobruce: yes, permissions are certainly partly to blame
00:30:06 <tcarobruce> nope -- the code isn't published anywhere I could find
00:30:06 <springmeyer> that is what I tried to trac down with tom previously
00:30:18 <springmeyer> and which we have no control over
00:30:30 <springmeyer> so until those scripts are publishes, mapbox == blackbox
00:30:36 <tcarobruce> what we've done -- will check on permission to make public --
00:30:46 <tcarobruce> is fuse mapbox and tiledrawer
00:31:16 <tcarobruce> tiledrawer kindly included their build script
00:31:24 <jfreeman> tcarobruce: ok
00:31:47 <tcarobruce> so, we're using that with the MapBox EBS snapshot (of OSM db)
00:32:13 <tcarobruce> (and then got into mod_tile...)
00:32:46 <jfreeman> springmeyer: I am just trying to produce a custom set of tiles for Google Maps
00:32:56 <jfreeman> springmeyer: any recommendations?
00:33:13 <springmeyer> jfreeman: what is your extent? do you need a global tileset?
00:33:28 <jfreeman> springmeyer: it's just for a single country
00:33:55 <tcarobruce> small country? => try tiledrawer
00:34:24 * jfreeman googles  tiledrawer
00:34:40 <springmeyer> and jfreeman do you have mapnik installed locally?
00:35:49 <jfreeman> springmeyer: I've got the output from a mapnik project
00:36:21 <jfreeman> springmeyer: so I have the xml file and a TIFF
00:36:25 <springmeyer> not sure what you mean?
00:36:30 <springmeyer> oh, okay
00:36:37 <springmeyer> and how was that xml + tiff created?
00:36:48 <jfreeman> springmeyer: using mapnik I understand
00:37:22 <springmeyer> hmm, we should make sure, especially before trying to throw it at TillMill :)
00:37:39 *** ajturner (~ajturner@pool-72-66-109-70.washdc.fios.verizon.net) has joined #mapnik
00:37:51 <springmeyer> http://dpaste.com a part or all of the XML ?
00:39:05 <jfreeman> http://dpaste.com/154329/
00:40:32 <springmeyer> okay, yep that is a Mapnik XML mapfile
00:40:47 <jfreeman> cool
00:41:03 <springmeyer> so, first off jfreeman you are going to need to make a few changes to that XML to get it to work with Gmaps
00:41:31 <springmeyer> how big is that tiff ?
00:41:37 <jfreeman> 500kb
00:41:39 <jfreeman> tiny
00:41:43 <springmeyer> ah, tiny
00:41:44 <springmeyer> yep
00:42:02 <springmeyer> so this map is just a demo of sorts?
00:42:16 <jfreeman> correct, just trying to get it working first
00:42:24 <springmeyer> okay, cool, I'm following
00:43:34 <jfreeman> :)
00:43:45 <springmeyer> so, that is kinda what I was getting at - TileMill (even if it worked) is extreme overkill for your purposes
00:43:45 <springmeyer> so here's what I recommend:
00:44:49 <springmeyer> 1) Install Mapnik, 2) install nik2img.py, 3) test rending a single map image, 4) then go from there...
00:44:51 * jfreeman listens
00:45:25 <jfreeman> springmeyer: so what will then server up the tiles?
00:45:32 <jfreeman> serve
00:46:01 * tcarobruce agrees with, defers to, springmeyer
00:46:39 <jfreeman> hey, I figured out my bloody InvalidKeyPair.NotFound error
00:46:56 <jfreeman> you have to specify the key pair name, not the actual private key file
00:46:59 * jfreeman slaps head
00:47:02 <jfreeman> DOH!
00:47:26 <springmeyer> cool, thats nice.
00:47:55 <springmeyer> jfreeman: my advice is to not worry about what will serve the tiles. that is the easy part
00:48:11 <springmeyer> make sure your data and map are working, and looks like what you want
00:48:22 <jfreeman> springmeyer: right. so I'll just google nik2img.py
00:49:10 <springmeyer> jfreeman: sure, but you need Mapnik installed first
00:49:51 <jfreeman> springmeyer: what for? I've got the mapnik output already, can't I just use that? Or does nik2img.py require it?
00:51:06 <springmeyer> tcarobruce: cool you got the two working together
00:51:41 <tcarobruce> was frustrated I couldn't use the TileDrawer AMI with TileMill EBS volume
00:51:53 <springmeyer> jfreeman: what for? to render a map, to see if your XML works.
00:51:55 <tcarobruce> but it was good to get it working!
00:52:08 <jfreeman> ok
00:52:24 <springmeyer> otherwise if you prefer you go back to sending TileMill jobs with a bogus XML file
00:53:46 <jfreeman> springmeyer: I'll take your advise :)
00:53:52 * jfreeman starts building mapnik
01:00:06 <springmeyer> jfreeman: what linux flavor?
01:00:17 <jfreeman> springmeyer: OpenSUSE 11.2
01:00:44 <jfreeman> springmeyer: is it simple to install on latest Ubuntu?
01:00:54 <jfreeman> springmeyer: I might just fire up a Ubuntu VM
01:00:59 <springmeyer> yep, very easy on Karmic
01:01:23 <jfreeman> springmeyer: ok will do that. does it need a Gnome desktop?
01:01:31 <springmeyer> no
01:01:56 <springmeyer> on Karmic you can even use the mapnik package
01:02:27 <jfreeman> cool. is 9.10 karmic?
01:02:36 <springmeyer> yes
01:02:43 <springmeyer> `UbuntuInstallation
01:02:43 <nikq> http://trac.mapnik.org/wiki/UbuntuInstallation
01:02:52 <springmeyer> ^^ detailed notes, with just a few gochas
01:04:25 <jfreeman> ta
01:18:29 * jfreeman ubuntu installed, just doing the other bits now
01:20:08 <springmeyer> ubuntu installed or mapnik installed on ubuntu ? :)
01:24:39 <jfreeman> springmeyer: ubuntu and updates, it takes a while
01:25:16 <springmeyer> ah, gocha
01:25:56 <springmeyer> ya, I guess you could just use your running instance too and install things there... might be quicker
01:28:46 <jfreeman> springmeyer: this is OK, do EC2 later :)
01:28:46 *** cgs_bob has quit (Read error: Connection reset by peer)
01:29:02 <StormTide> springmeyer, how do you spec a layer in 900913... proj_init_error:failed to initialize projection with:+init=EPSG:900913
01:29:02 <StormTide>  comes from srs="+init=EPSG:900913"
01:29:28 <springmeyer> StormTide: you add 900913 to /usr/share/proj/epsg
01:29:51 <jfreeman> springmeyer: OK all done
01:30:05 <jfreeman> springmeyer: mapnik and nik2img installed
01:30:18 <springmeyer> StormTide: every last gory detail: http://docs.openlayers.org/library/spherical_mercator.html
01:30:30 <springmeyer> jfreeman: good stuff. are you able to
01:30:36 <springmeyer> python -c "import mapnik"
01:30:37 <springmeyer> ?
01:30:55 <jfreeman> python -c "import mapnik" -> no errors
01:31:32 <springmeyer> then run nik2img on your xml file and check the output
01:32:43 <jfreeman> springmeyer: I get this error: .\9kmImage.tif does not exist in layer 'Australia'
01:33:07 <springmeyer> yep, that was error #1 I noticed
01:33:07 <nikq> Ticket #1: no such ticket. (list index out of range)
01:33:17 <springmeyer> try removing the '.\'
01:33:29 <jfreeman> yup
01:33:33 <springmeyer> its as if that file came from windows...
01:33:34 <StormTide> springmeyer, do i need to recompile something to get it to pickup the change?
01:33:50 <springmeyer> StormTide:  no
01:34:09 <jfreeman>  nik2img.py map.xml output.png -> works
01:34:13 <jfreeman> :)
01:34:14 <springmeyer> StormTide: but you might need to ensure you don't have two libproj.so libraries around....
01:34:27 <springmeyer> jfreeman: good
01:34:41 <springmeyer> jfreeman: so the next problem is that you need your tiff in a different projection
01:34:48 <springmeyer> if you want to overlay it on google maps
01:35:12 <jfreeman> springmeyer: right. We only want to have one zoom level at this stage (if that makes sense)
01:35:25 <springmeyer> oh, interesting
01:35:28 <springmeyer> sure thats fine
01:35:59 <StormTide> springmeyer, weird... i just have the one... made the change ... restarted renderd still no go... and quantumnik wont load it
01:36:23 <StormTide> nik2img crap out too
01:36:26 <springmeyer> StormTide: odd
01:36:35 <StormTide>     <Layer name="US5WA41M_soundg_layer" srs="+init=EPSG:900913" clear_label_cache="1">
01:36:40 <StormTide> is my layer tag
01:36:46 <springmeyer> StormTide: what does this give....
01:36:57 <springmeyer> $ ldd /usr/local/lib/libmapnik.so
01:37:06 <springmeyer> e.g what proj does it point to...
01:37:12 <StormTide> <900913> +proj=merc +a=6378137 +b=6378137 +lat_ts=0.0 +lon_0=0.0 +x_0=0.0 +y_0=0 +k=1.0 +units=m +nadgrids=@null +no_defs <>
01:37:12 <StormTide>  added to proj... (tried with and without final <>)
01:37:28 <StormTide> 	libproj.so.0 => /usr/lib/libproj.so.0 (0xb7525000)
01:37:47 <jfreeman> springmeyer: I think we're at the 4) go from there stage... so what next?
01:37:49 <springmeyer> hmm, so that should look to /usr/share/proj/epsg
01:38:13 <springmeyer> jfreeman: #4 is reproject you raster
01:38:13 <nikq> Ticket #4: database connection id, http://trac.mapnik.org/ticket/4
01:38:24 <springmeyer> sorry, you/your
01:39:32 <springmeyer> jfreeman: $ gdal_translate -t_srs EPSG:900913 9kmImage.tif merc.tif
01:39:34 <StormTide> ah, figured it out
01:39:39 <StormTide> epsg has to be lowercase
01:39:53 <springmeyer> ah, thats true
01:40:10 <springmeyer> StormTide: file a ticket and I'll fix that
01:40:24 <springmeyer> thats a common mistake...
01:41:04 <StormTide> wow, cool, this actually worked... sorta
01:41:09 <StormTide> but i have broken tiles...
01:41:14 <jfreeman> linux@ubuntu:~/windlab$ gdal_translate -t_srs EPSG:900913 9kmImage.tif merc.tif
01:41:16 <jfreeman> Option -t_srs incomplete, or not recognised.
01:41:35 <springmeyer> jfreeman: my bad
01:41:43 * springmeyer tired
01:41:55 * jfreeman is very thankful
01:41:58 <StormTide> springmeyer, http://dev.stormtide.ca/test.htm ... ever seen broken tiles like that before?
01:42:02 <springmeyer> gdalwarp is the command...
01:42:21 <jfreeman> springmeyer: cool. that works
01:42:34 <jfreeman> springmeyer: now have a merc.tif
01:43:06 <StormTide> http://dev.stormtide.ca/test.png <-- nik2img seems to be working
01:43:12 <StormTide> is there a timeout or something going on/?
01:43:13 <springmeyer> StormTide: do you need to clear your cache?
01:43:35 <springmeyer> ew, thats pretty :)
01:43:38 <StormTide> nope, there just did that again
01:43:44 <StormTide> and still broken tiles all over
01:43:56 <springmeyer> what do you mean by broken?
01:44:03 <StormTide> zoom around
01:44:10 <StormTide> you'll see pink tiles with a broken image
01:44:12 <StormTide> icon
01:44:25 <jfreeman> springmeyer: so whats' next buddy?
01:44:31 <StormTide> oh
01:44:32 <StormTide> hrm
01:44:36 <StormTide> maybe this is my browsers cache
01:45:01 <StormTide> nope
01:45:07 <StormTide> its like theres a time limit for tile download
01:45:14 <StormTide> and if its not there, it pink image's me
01:45:22 <StormTide> if i go  view-image the image works
01:45:38 <StormTide> might be clientside that this is going on...
01:46:04 <springmeyer> StormTide: your OL may not be initializing correctly
01:46:39 <StormTide> hrmph... the code for that page is basically yanked from your tilelite thing
01:46:40 <springmeyer> StormTide: I'd recommend trying to add a global OSM layer as well, so you can compare to your layer
01:46:51 <springmeyer> seems a bit to yanked...
01:46:54 <springmeyer> to/too
01:46:59 <jfreeman> springmeyer: so now that I have this merc.tif generated, what's next?
01:47:01 <StormTide> (just testing)
01:47:05 <StormTide> wont stay that way
01:47:06 <springmeyer> jfreeman: "buddy" ? ;)
01:47:49 <jfreeman> springmeyer: well, you've been very helpful so it's an appropriate title
01:48:12 <springmeyer> jfreeman: #5 would be to test serving tiles, for which I recommend you try "TileLite"
01:48:12 <nikq> Ticket #5: Accented characters not working properly, http://trac.mapnik.org/ticket/5
01:48:28 <StormTide> springmeyer, what happens when you add two maps for the same extent?
01:48:33 <jfreeman> right
01:48:38 <StormTide> i think that may be what is going on here
01:49:10 <springmeyer> oh, jfreeman wait, gotta edit the 'srs' value in your XML...
01:49:43 <springmeyer> change the 'srs' in your XML for the map and the layer to:
01:49:45 <springmeyer> srs="+proj=merc +a=6378137 +b=6378137 +lat_ts=0.0 +lon_0=0.0 +x_0=0.0 +y_0=0 +k=1.0 +units=m +nadgrids=@null +no_defs"
01:50:02 <springmeyer> StormTide: not sure what you are asking....
01:50:20 <StormTide> springmeyer, like, im pretty sure this script i wrote has added two layers that cover the same area
01:50:31 <StormTide> might that be confusing mod_tile/
01:50:56 <springmeyer> hmm, not sure
01:50:57 <jfreeman> springmeyer: ok done
01:51:30 <StormTide> the pink tile issue is definitely some sort of timeout in the OL lib
01:51:54 <springmeyer> StormTide: right click on the pink tile, view it and check if the tile loads...
01:51:59 <StormTide> it does
01:52:14 <StormTide> and the more i reload the page, and the more it picks up cached tiles
01:52:19 <StormTide> they work
01:52:34 <jfreeman> springmeyer: so for TileLite I do a complete Apache install with Mod_Tile ?
01:53:03 <springmeyer> nope, it has a development server that has no dependencies
01:53:08 <springmeyer> other than python and mapnik
01:53:49 <springmeyer> and it has a sample HTML file to view the tiles, look in the /utils folder
01:54:17 <springmeyer> once you see that your tiles line up with google maps, then I'd say you are ready to try pushing your XML to mapbox.com again....
01:54:24 <jfreeman> springmeyer: so to run it, I just do: liteserv.py map.xml
01:54:28 <springmeyer> er, tilemille I mean
01:54:34 <springmeyer> jfreeman: what do you think?
02:00:02 <jfreeman> springmeyer: mate, I'm lost. How do I configure this tilelite app?
02:00:24 <springmeyer> $ liteserv.py map.xml
02:01:29 <jfreeman> springmeyer: ok, done that. then how do I view the map?
02:02:06 *** cgs_bob (~bob@122.135-78-65.ftth.swbr.surewest.net) has joined #mapnik
02:02:28 <springmeyer> open the /utils/openlayers.html file in a browser from the filesystem
02:03:20 <jfreeman> springmeyer: ok done that
02:04:53 <jfreeman> springmeyer: am I supposed to be able to see the custom map tiles?
02:05:24 <springmeyer> yep, but needs to be connected to the net...
02:05:36 <springmeyer> what do you see?
02:07:37 <jfreeman> springmeyer: I see the Tilelist Mapnik Server page, with Australia
02:08:10 <jfreeman> springmeyer: on the right hand side I have some options, for other layers. Open Street Map, Tile Lite, Google
02:10:09 <springmeyer> so seeing australia is a good sign. are you concerned that it does not look like your .tif file?
02:10:31 <StormTide> springmeyer, gah, figured it out... so it is that there's multiple layers covering the same area (least for the looking like crap part)... will have do some sort of sort to emit the xml
02:10:47 <StormTide> im _floored_ at how well mapnik is handling this data tho
02:11:04 <springmeyer> ya, it will render a lot faster if it does not have to reproject on the fly
02:11:13 <jfreeman> springmeyer: correct. does not look like the tif file has been applied at all, just a standard Australia map
02:11:45 <springmeyer> "standard" like google maps tiles or something?
02:12:03 <jfreeman> springmeyer: correct
02:12:05 <springmeyer> if your data is not showing up then
02:12:08 <springmeyer> okay
02:12:17 <springmeyer> stop the "liteserv.py' server
02:12:28 <springmeyer> and try starting it again with the flag
02:12:36 <jfreeman> springmeyer: ok
02:12:38 <springmeyer> -b 0
02:12:41 <jfreeman> springmeyer: stopped
02:12:53 <jfreeman> springmeyer: ctrl-c to kill
02:12:57 <springmeyer> yep
02:13:18 <StormTide> is the first <layer> in the file the first to render or the last...
02:13:28 <springmeyer> first
02:13:38 <springmeyer> so it will get covered up by others
02:13:42 <StormTide> k kew
02:13:46 <StormTide> i can fix this
02:17:23 <jfreeman> springmeyer: you can view my Tilelist server at http://203.217.61.228:8000
02:17:40 <jfreeman> springmeyer: if that helps
02:17:51 <springmeyer> nope, can't view
02:18:01 <springmeyer> jfreeman: can you post your .tif ?
02:19:11 <jfreeman> will try
02:19:22 *** tcarobruce has quit (Quit: tcarobruce)
02:25:48 <jfreeman> springmeyer: direct DCC ok?
02:26:36 <springmeyer> I've never got it to work in the past
02:26:43 <springmeyer> dbsgeo@gmail.com
02:26:53 <jfreeman> springmeyer: ok ta
02:29:34 <jfreeman> springmeyer: sent both the 9km image and the merc.tif
02:29:51 <springmeyer> http://dpaste.com your xml real quick?
02:30:42 <jfreeman> springmeyer: http://dpaste.com/154345/
02:31:41 <springmeyer> ah, thats your problem
02:31:56 <springmeyer> springmeyer: change the 'srs' in your XML for the map and the layer to:
02:31:56 <springmeyer> [5:49pm] springmeyer: srs="+proj=merc +a=6378137 +b=6378137 +lat_ts=0.0 +lon_0=0.0 +x_0=0.0 +y_0=0 +k=1.0 +units=m +nadgrids=@null +no_defs"
02:32:07 <springmeyer> BOTH map and layer 'srs'
02:32:34 <springmeyer> an you need to change the hix,lox, etc...
02:34:47 <jfreeman> springmeyer: right, changed the layer srs. What's the hix, lox etc. need to be changed too?
02:35:11 <springmeyer> try removing them
02:35:21 <springmeyer> and change the type="gdal"
02:36:17 <jfreeman> does it matter that the image being served is not the merc.tif?
02:36:29 <springmeyer> yep
02:36:37 <springmeyer> you need to point it at merc.tif
02:36:47 <jfreeman> ok
02:37:23 <jfreeman> shit
02:37:25 <jfreeman> that works
02:37:28 <jfreeman> fucking hell
02:37:29 <jfreeman> :)
02:37:35 <StormTide> is there an openlayers.js channel?
02:37:42 <springmeyer> yep, this works for me as well: http://dpaste.com/154347/
02:37:45 <StormTide> need to figure out how to alter this tile timeout
02:38:07 <springmeyer> StormTide: #openlayers yes
02:38:37 <jfreeman> springmeyer: your version is quite different
02:39:04 <jfreeman> xml I mean
02:39:08 <jfreeman> will give it a go
02:39:21 <springmeyer> yep, the srs I read using gdalinfo
02:39:30 <springmeyer> seems your reprojection did not work quite right
02:39:40 <springmeyer> but your data is so low-res it won't matter
02:40:11 <jfreeman> ok
02:40:12 <springmeyer> jfreeman: try also for fun:  <CssParameter name="scaling">bilinear</CssParameter>
02:40:30 <springmeyer> later guys, I'm headed out
02:40:32 *** springmeyer has quit (Quit: springmeyer)
03:10:43 *** jfxberns (~jfxberns@ppp-61-90-16-242.revip.asianet.co.th) has joined #mapnik
03:42:41 *** mperry_ (~mperry@c-67-164-175-65.hsd1.co.comcast.net) has joined #mapnik
03:42:41 *** mperry has quit (Read error: Connection reset by peer)
03:42:42 *** mperry_ is now known as mperry
04:14:48 *** jfreeman has quit (Remote host closed the connection)
04:17:14 *** jfreeman (~jfreeman@mail.agileware.net) has joined #mapnik
04:27:08 <StormTide> anyone know the solution to generate_tiles.py dying saying 'MemoryError'
04:51:20 *** jfxberns has quit (Read error: Connection reset by peer)
04:52:55 *** jfxberns (~jfxberns@ppp-58-8-123-233.revip2.asianet.co.th) has joined #mapnik
04:54:07 *** StormTide has quit (Remote host closed the connection)
05:03:35 *** racicot has quit (Ping timeout: 256 seconds)
05:04:30 *** racicot (~chatzilla@dsl-209-166-85-189.whidbey.net) has joined #mapnik
05:04:48 *** cgs_bob has quit (Ping timeout: 252 seconds)
05:09:20 *** StormTide (~Kevin@2002:6036:ec1:0:21d:60ff:fe5e:cf66) has joined #mapnik
05:10:13 *** cgs_bob (~bob@122.135-78-65.ftth.swbr.surewest.net) has joined #mapnik
05:27:01 <StormTide> ah, think i figured it out... generate_times.py and generate_image.py use mapnik.Envelope... since im running head... thats wrong. Should be mapnik.Box2d ... might be a good idea to add a deprecated alias for the renamed stuff?
05:59:56 *** HounD (~HounD@unics1.grfc.ru) has joined #mapnik
06:08:54 *** ajturner_ (~ajturner@pool-72-66-109-70.washdc.fios.verizon.net) has joined #mapnik
06:08:54 *** ajturner has quit (Read error: Connection reset by peer)
06:08:54 *** ajturner_ is now known as ajturner
06:14:42 *** ajturner has quit (Ping timeout: 272 seconds)
06:41:26 <jfreeman> What's the preferred Map tile server on Ubuntu?
06:50:33 *** springmeyer (~springmey@c-76-28-156-154.hsd1.wa.comcast.net) has joined #mapnik
06:50:42 <StormTide> wb springmeyer  ;)
06:55:54 <springmeyer> wb?
06:56:01 <StormTide> (welcome back)
06:56:07 * springmeyer :)
06:56:24 <springmeyer> actually crashing out now, take it easy StormTide
06:56:31 <StormTide> you too ;)
06:56:33 <springmeyer> jfreeman: mod_tile
06:56:39 <StormTide> i got the map workin
06:56:41 <springmeyer> yep, g'night
06:56:41 <StormTide> its awesome
06:56:46 <springmeyer> cheers
06:57:03 <jfreeman> springmeyer: welcome back. that was a quick sleep!
06:57:10 <springmeyer> go find yerself a cold one StormTide
06:57:23 <StormTide> i intend to
06:57:29 <StormTide> fixed a bug in generate_image.py too
06:57:43 <StormTide> needs a Box2d instead of Envelope when running against trunk
06:57:49 <springmeyer> ya, those scripts are not trunk ready yet
06:58:11 <springmeyer> file a ticket at trac.openstreetmap.org if you have a moment tomorrow...
06:58:19 <StormTide> k
06:58:45 <StormTide> nite nite
06:58:47 * springmeyer fades
06:59:28 <jfreeman> springmeyer: so will mod_tile work with the xml and merc.tif file
07:00:16 <jfreeman> anyway, I'll check it
07:20:10 *** mperry_ (~mperry@c-67-164-175-65.hsd1.co.comcast.net) has joined #mapnik
07:23:24 *** mperry has quit (Ping timeout: 252 seconds)
07:23:24 *** mperry_ is now known as mperry
07:28:21 *** jfreeman has quit (Remote host closed the connection)
08:08:41 <Ldp__> springmeyer, StormTide: http://trac.openstreetmap.org/changeset/19853
08:09:26 <StormTide> Ldp__, upon some more testing i fear that it may take more
08:09:35 <StormTide> whenever i try an extent in latlong
08:09:42 <StormTide> it seems to render a large white spot on top
08:09:46 <Ldp__> generate_tiles.py was rendering for me, but I didn't check the output
08:10:03 <StormTide> it renders, but the output file appears to not match the defined extent
08:10:06 <StormTide> but that could be me
08:10:11 <Ldp__> anything else than checking for the version and doing Box2d or Envelope would point to another issue
08:10:26 <StormTide> my map/layers are in 900913... and the box in latlong... so im not 100% sure
08:10:33 <StormTide> might be a projection issue too
08:11:27 <Ldp__> generate_image.py does a prj.forward to 900913
08:11:51 <StormTide> ya... which is why i think there might be an issue
08:12:25 <StormTide> ll = (-121,42,-153,61)
08:12:28 <StormTide> im trying that extent
08:13:38 <StormTide> http://www.stormtide.ca/image.png
08:14:16 <StormTide> notice the white at the top
08:14:18 <StormTide> i dont believe that matches the extent
08:14:57 <StormTide> nor the bottom really
08:15:02 *** jfxberns has quit (Quit: Going to hide someplace dark and quiet...)
08:15:05 <Ldp__> what happens if you take an extent that matches an area from the bottom of that image?
08:15:20 <StormTide> come again?
08:15:55 <Ldp__> part of your image renders something, right?
08:16:10 <StormTide> yes... can you not see it?
08:16:20 <Ldp__> set your ll to an area within that area, see how that renders
08:17:19 <Ldp__> also, does it matter if you swap -121 with -153
08:18:47 <StormTide> lemme try a few things
08:18:47 <StormTide> sec
08:18:51 <Ldp__> you're doing (east,south,west,north) instead of (west,south,east,north). I don't know if that matters to Box2d
08:20:11 <StormTide> hrm
08:20:47 <StormTide> i thought i was doing toplet latlong to botright latlong
08:20:51 <StormTide> lemme try switching
08:21:16 <Ldp__> -121 is east of -153
08:21:46 <Ldp__> 42 is south of 61, too ;)
08:22:02 <StormTide>     ll = (-121,42,-153,61)
08:22:29 <StormTide> er
08:22:30 <StormTide>     ll = (-128.380737,47.838971,-121.992187,50.746015)
08:22:35 <StormTide> that behaves the same as
08:23:18 <StormTide>     ll = (-121.992187,47.838971,-128.380737,50.746015)
08:23:26 <Ldp__> okay, bummer
08:23:35 <StormTide> as for your smaller subset idea
08:23:36 <StormTide> it works
08:23:38 <StormTide> however
08:23:44 <StormTide> it pads the top with more data than it should
08:24:04 <StormTide> http://maps.google.ca/?ie=UTF8&ll=50.746015,-128.380737&spn=0.649111,1.783905&z=10
08:24:09 <StormTide> for example, should be the tip of the island
08:24:23 <Ldp__> last time I tried generate_image.py, which was actually a few days ago and with mapnik trunk, it seemed to work fine
08:24:35 <StormTide> http://maps.google.ca/?ie=UTF8&ll=47.838971,-121.992187&spn=2.754226,7.13562&z=8 <-- bottom right
08:24:48 <StormTide> now lemme upload this img
08:24:49 <StormTide> sec
08:26:17 <StormTide> http://www.stormtide.ca/image.png
08:26:17 <StormTide> there
08:26:22 <StormTide> see how much above and outside that bounding box it renders
08:26:28 <StormTide> same with below
08:27:02 <StormTide> the image seems to be terribly skewed in the vertical directions
08:27:33 <Ldp__> hmm, now that you mention it, it seemed I got the wrong extent rendered a few days ago. I was thinking about the padding you showed earlier, which I didn't encounter
08:28:03 <StormTide> im wondering if it could be the projection conversion failing right
08:28:15 <Ldp__> but I was trying to render Port-au-Prince, but got an image of the area to the north of it. Didn't investigate further
08:28:37 <Ldp__> would be interesting to go back to 0.7.0 and run the same generate_image.py to see what that gives me
08:28:58 <StormTide> ;)
08:35:05 <Ldp__> compiling will take a while
08:46:25 <Ldp__> hm, looking at the extent I was given and what it actually renders, seems to be correct. I might not get the same as you're getting
08:49:45 *** StormTide has quit (Ping timeout: 260 seconds)
09:03:13 *** StormTide (~Kevin@2002:186c:64c0:0:21d:60ff:fe5e:cf66) has joined #mapnik
10:12:20 *** jfreeman (~jfreeman@60-242-98-237.static.tpgi.com.au) has joined #mapnik
10:25:45 *** mperry_ (~mperry@c-67-164-175-65.hsd1.co.comcast.net) has joined #mapnik
10:25:45 *** mperry has quit (Read error: Connection reset by peer)
10:25:46 *** mperry_ is now known as mperry
10:58:34 *** mishok13 has quit (Read error: Operation timed out)
11:05:56 *** mishok13 (~gdmfsob@smtp-kyiv.cogniance.com) has joined #mapnik
11:28:42 *** jfreeman has quit (Ping timeout: 240 seconds)
13:17:33 *** artem (~artem@i-83-67-73-6.freedom2surf.net) has joined #mapnik
13:31:15 *** HounD has parted #mapnik (None)
14:24:38 *** tomhughes has quit (Read error: Operation timed out)
14:28:28 *** tomhughes (~tom@gosford.compton.nu) has joined #mapnik
14:30:11 *** HounD (~HounD@unics1.grfc.ru) has joined #mapnik
14:50:40 *** ajashton (~aj@c-69-136-229-112.hsd1.dc.comcast.net) has joined #mapnik
15:04:58 *** ajturner (~ajturner@209.155.228.129) has joined #mapnik
15:05:16 *** HounD has parted #mapnik (None)
15:33:41 *** artem has quit (Quit: artem)
15:35:08 *** artem (~artem@i-83-67-73-6.freedom2surf.net) has joined #mapnik
16:00:56 *** chad_burt (~chad_burt@mm-01.msi.ucsb.edu) has joined #mapnik
16:46:15 *** cgs_bob has quit (Ping timeout: 256 seconds)
16:47:57 <CIA-31> mapnik-utils: dane.springmeyer * r907 /trunk/serverside/cascadenik/tests: remove duplicate tests
16:53:30 <CIA-31> mapnik-utils: dane.springmeyer * r908 /trunk/serverside/cascadenik/setup.py: don't really need setuptools (only benefit of it autodownload of cssutils), so ditch it and avoid the various pitfalls of setuptools black magic
16:53:30 <CIA-31> mapnik-utils: dane.springmeyer * r909 /trunk/serverside/cascadenik/ (cascadenik/compile.py test.py cascadenik/style.py): apply patch with tests added from tmcw for supporting the polygon-gamma attribute available in Mapnik 0.7.0
16:54:02 <StormTide> morning springmeyer ;)
16:55:09 <springmeyer> morning
16:55:46 <StormTide> did you catch the convo with Ldp__ last night about the extent bug in generate_*.py with trunk?
16:55:57 <StormTide> even with Box2d it seems to render the wrong extent
16:56:00 <springmeyer> just skimmed
16:56:07 <springmeyer> for one your patch is wrong
16:56:10 <springmeyer> if hasattr(mapnik,'mapnik_version') and mapnik.mapnik_version() >= 700:
16:56:12 <springmeyer> needs to be
16:56:15 <springmeyer> if hasattr(mapnik,'mapnik_version') and mapnik.mapnik_version() >= 800:
16:56:30 <artem> morning
16:56:41 <springmeyer> afternoon artem
16:56:42 <Ldp__> springmeyer: hmm, I lifted it from your patch for generate_xml.py
16:56:45 <StormTide> (that was Ldp__ ) ... i _just_ changed to box2d on my side
16:57:13 <springmeyer> oh, ya the other method was added in 0.7.0, aka 700
16:57:23 <springmeyer> but the box2d is currently only in trunk...
16:57:35 <Ldp__> ok, it was early, I wasn't very awake yet. I'll fix it
16:58:13 <springmeyer> okay, thanks
16:58:48 <springmeyer> StormTide: you might want to fiddle with this map level setting:
16:58:49 <springmeyer> http://media.mapnik.org/api_docs/python/mapnik._mapnik.aspect_fix_mode-class.html
16:59:44 <StormTide> hrm, why would there be an aspect ratio at work in generate_image.py...
16:59:50 <springmeyer> or use nik2img
17:00:13 <springmeyer> because the height and width of the image is hardcoded and is a different aspect ratio than the bbox
17:00:19 <springmeyer> ... in generate_tiles.py
17:00:34 <StormTide> oooh
17:00:38 <StormTide> bahaha that makes sense
17:00:53 <StormTide> imgx and imgy
17:01:19 <StormTide> really, i wonder if its possible to glean those values from the bbox instead of hardcode
17:01:23 <springmeyer> generate_tiles should likely use GROW_CANVAS so that the size ratio is changed by the bbox
17:01:29 <springmeyer> yes, that would be possible too
17:01:46 <springmeyer> generate_tiles.py is 49 lines long :)
17:02:05 <StormTide> yep, been hacking on it ;)
17:02:10 <StormTide> _image that is
17:02:18 <springmeyer> can't expect it to be perfect, its really just used to make sure you data is rendering at all
17:03:06 <StormTide> anyway, i was building with gen tiles... but it was resulting in mostly white (blank) tiles... so ive a hunch that its using the same issue
17:03:46 <springmeyer> StormTide: again, what I think is going on is that you OL instance is fragile
17:03:52 <springmeyer> and sending odd bbox's
17:04:06 <springmeyer> did you do what I said and put another reliable layer in that OL file?
17:04:09 <StormTide> nono the openlayers side works flawly
17:04:13 <StormTide> flawlessly
17:04:17 <StormTide> ya, i added osm to it
17:04:19 <StormTide> and it all lined up nice
17:04:25 <StormTide> the mod_tile approach is flawless
17:04:30 <springmeyer> okay
17:04:37 <springmeyer> then what is wrong?
17:04:44 <StormTide> generate_tiles.py
17:05:09 <StormTide> when given a ll extent is rendering a totally different one... but it makes sense if its changing to an aspect
17:05:37 <springmeyer> ah, sorry thought you were talking about generate_image.py...
17:05:41 <StormTide> both
17:05:58 <StormTide> image and tiles.py seem to have the same issue
17:06:05 <springmeyer> hmm
17:08:08 <StormTide> ll = (-153,61,-121,42) try that for example (have also tried swapping the 121 and 153 )
17:08:31 <springmeyer> that second coordinate is bogus
17:08:50 <springmeyer> oh sorry, sleepy
17:08:57 * springmeyer blinks
17:09:47 <StormTide> anyway, 61 should be about inline with anchorage alaska
17:09:59 <StormTide> but it actually renders like 100 miles north of that point
17:10:11 <StormTide> actually, like 1000
17:10:44 <StormTide> i'll upload another img
17:11:06 <springmeyer> StormTide: issues are going to be different between generate_image.py and generate_tiles.py
17:11:36 <springmeyer> what is wrong with the generate_tiles.py output?
17:12:01 <StormTide> it appears to be the same blank tiles gen image gives for the map
17:12:16 <springmeyer> in places where there should be data?
17:12:44 <StormTide> it just renders the extent well north of where i told it too
17:12:49 <StormTide> one sec my ip changed last night
17:12:51 <StormTide> *pita*
17:14:45 <StormTide> httpwww.stormtide.ca/image.png
17:14:48 <StormTide> http://www.stormtide.ca/image.png
17:15:11 <StormTide> i get directories full of blank tiles for that extent
17:15:52 <StormTide> so ive a hunch they're working the same, adding whitespace to the image, maybe to maintain the ratio
17:20:53 *** jctull (~jctull@ppp-71-142-138-235.dsl.renocs.pacbell.net) has joined #mapnik
17:23:45 <springmeyer> ratio does not apply with generate_tiles.py because the aspect ratio is at the map level the "map" that generate_tiles.py creates are tiles 256x256
17:24:20 <springmeyer> is generate_tiles.py rendering tiles outside of the bbox you are submitting, sure that could be the case
17:25:01 <StormTide> i can try the tiles thing again, but it seemed bogus last time...
17:25:08 <StormTide> the _image.py produces that image.png
17:25:11 <springmeyer> you can monkey with your map background color to see if that is the case
17:25:20 <StormTide> ooo
17:25:22 <StormTide> *tries*
17:25:54 <springmeyer> bogus is different that just creating a few tiles outside your area of interest but still rendering the right tiles
17:26:49 <StormTide> ok, so it renders blue above with map bgcolor changed to blue with generate_image.py
17:27:04 <StormTide> i'll try the tiles
17:28:10 *** artem has quit (Quit: artem)
17:33:20 <StormTide> appears ive borked something
17:33:25 <StormTide> gen tiles wont gen tiles anymore heh
17:37:41 <StormTide> ah, ok with gen tiles its just the difference in the box2d co-ord order
17:38:12 <StormTide> (west,south,east,north)
17:38:16 <StormTide> is the format it needs to be
17:40:53 <StormTide> its still rendering a lot outside that bbox tho
17:40:57 <StormTide> like... a lot of blue
17:41:07 <StormTide> im not exactly sure how its supposed to work on the tiles tho
17:42:28 <StormTide> http://www.stormtide.ca/0.png
17:42:51 <StormTide> as the zoom level increases, i seem to get more and more blue (blank) tiles
17:43:03 <StormTide> and few actual tiles
17:43:36 <springmeyer> umm, that looks correct to me
17:44:22 <springmeyer> 0.png is the left side of the globe and your data coverage only is for a certain part, so of course you would get blue
17:44:42 <springmeyer> seems that you think generate_tiles.py is being limited to just rendering tiles for you data extent
17:44:48 <springmeyer> but that does not appear to be the case
17:44:58 <StormTide> thats what i thought it did, yes
17:44:59 <springmeyer> so I think you are likely not using it correctly
17:45:02 <StormTide> but i guess not
17:45:51 <StormTide> seems off that the result of my generation at higher zoom levels would have blue tiles somewhere in russia, but i can try a full build and see what it looks like when plugged in
17:46:02 <StormTide> its just 99% blank tiles
17:46:35 <springmeyer> because it is generating tiles at the global extent
17:46:49 <StormTide> so what does the bbox() extent do then?
17:47:10 <springmeyer> paste your hacked copy of generate_tiles.py ?
17:47:14 <StormTide> sec
17:50:02 <StormTide> view-source:http://dev.stormtide.ca/generate_tiles.py
17:50:20 <StormTide> have played with the zoom level a bunch too
17:50:37 <StormTide> set to 3 now, but 0.png was from 1
17:50:50 *** tcarobruce (~tcarobruc@adsl-75-10-247-30.dsl.pltn13.sbcglobal.net) has joined #mapnik
17:50:58 <StormTide>         bbox = mapnik.Box2d(c0.x,c0.y, c1.x,c1.y) <-- is the only change to the actual script
17:51:51 <springmeyer> I'd guess it is the zoom level that is tripping you up
17:52:34 <springmeyer> level 3 is likely forcing the bounding box to be bigger than you want
17:52:41 <springmeyer> try launching tilelite
17:52:56 <springmeyer> and that OL example which displays the zoom level in the lower left
17:53:18 <springmeyer> and see what the zoom level is once you get down to your data
17:53:24 * springmeyer heads out for a meeting
17:53:26 <springmeyer> later StormTide
17:53:28 <StormTide> later
17:53:31 <StormTide> thanks for the help
18:00:53 *** springmeyer has parted #mapnik (None)
18:30:30 *** Phurl has quit (Quit: Ex-Chat)
19:05:15 *** cgs_bob (~bob@71.sub-75-210-241.myvzw.com) has joined #mapnik
19:09:10 *** artem (~artem@i-83-67-73-6.freedom2surf.net) has joined #mapnik
19:10:52 *** Phurl (~mdupont@ip-81-210-228-126.unitymediagroup.de) has joined #mapnik
19:26:02 *** mperry_ (~mperry@c-67-164-175-65.hsd1.co.comcast.net) has joined #mapnik
19:26:32 *** mperry has quit (Ping timeout: 265 seconds)
19:26:32 *** mperry_ is now known as mperry
19:48:10 *** jctull has quit (Quit: jctull)
20:06:12 *** springmeyer (~springmey@c-76-28-156-154.hsd1.wa.comcast.net) has joined #mapnik
20:08:30 *** jfreeman (~jfreeman@60-242-98-237.static.tpgi.com.au) has joined #mapnik
20:08:56 *** artem has quit (Quit: artem)
20:10:07 *** jbronn (~jbronn@70-138-113-15.lightspeed.hstntx.sbcglobal.net) has joined #mapnik
20:10:40 *** artem (~artem@i-83-67-73-6.freedom2surf.net) has joined #mapnik
20:24:21 *** Pepper (~Pepper@91-114-227-13.adsl.highway.telekom.at) has joined #mapnik
20:45:59 <StormTide> what's the usual bottleneck on generate_tiles.py speed. I cant see hdd, memory or cpu maxing out and ive got it set to 4 threads....
20:48:31 *** jfreeman has quit (Read error: Operation timed out)
20:48:38 <StormTide> http://www.phpriot.com/2673
20:50:06 <Ldp__> what's the difference between 'from mapnik import *' and 'import mapnik' ?
20:50:25 <StormTide> one puts the contents in the local scope i think
20:50:36 <StormTide> dont quote me on that tho
20:50:46 <Ldp__> generate_image.py has 'from mapnik import *' and it fails on the version test
20:50:53 <Ldp__> with 'import mapnik' it's fine
20:50:57 <StormTide> eg import mapnik, you cant just use Box2d... you need mapnik.Box2d ... i think
20:51:19 <StormTide> Ldp__, did you try dropping the mapnik.
20:51:24 <StormTide> in the one that imports *
20:52:38 <Ldp__> crap, why is it so hard to break multithreaded python with ctrl-c ?
20:52:52 <StormTide> is generate_tiles confined to a certain memory/cpu quota?
20:53:01 <StormTide> Ldp__, you can control z it then killalll
20:53:11 <StormTide> been doing that all morning ;)
20:53:33 <Ldp__> screen ctrl-a c works too
20:53:51 <CIA-31> mapnik-utils: dane.springmeyer * r910 /trunk/serverside/cascadenik/ (cascadenik/compile.py test.py cascadenik/style.py): apply patch from macwright extending support for Mapnik 0.7.0 character_spacing on Text and Shield Symbolizers
20:53:55 <Ldp__> anyway, I read that generate_images.py doesn't currently work
20:54:31 <StormTide> Ldp__, springmeyer seems to think it has to do with the fact the width/height are hardcoded and are out of ratio with my bbox
20:54:48 <StormTide> which is the explanation for the aspect ratio issues
20:54:55 <Ldp__> I read that part, but that's not the cause. This is different, but I'll add mapnik. to the right calls
20:56:45 <StormTide> i'd be happy if gen tiles would actually use the resources that are available to it
21:01:57 <CIA-31> mapnik-utils: dane.springmeyer * r911 /trunk/serverside/cascadenik/AUTHORS.txt: Add Tom to authors list with other patch contributers - did I miss anyone?
21:02:43 <springmeyer> Ldp__: what do you mean "fails on the version test" ?
21:02:55 <springmeyer> oh, i know
21:03:05 <Ldp__> springmeyer: fixed it now
21:03:12 <springmeyer> ya, you need to be able to have 'mapnik' in the namespace for that to work
21:03:14 <springmeyer> k
21:03:20 <springmeyer> ya, need `import mapnik`
21:03:24 <Ldp__> http://trac.openstreetmap.org/changeset/19869
21:03:27 <Ldp__> http://trac.openstreetmap.org/changeset/19871
21:03:45 <Ldp__> that'll teach me to actually run it before committing
21:03:55 <springmeyer> great
21:04:17 <Ldp__> now let's see if the Germans pick up on it
21:04:27 <Ldp__> http://lists.openstreetmap.org/pipermail/talk-de/2010-February/062571.html
21:06:58 <dodobas> ahh finaly...boost 1.42...compiled, damn thats big
21:07:11 <dodobas> over 2gb of memory used during build
21:07:26 <dodobas> mapnik 0.7 builded, tested
21:07:37 <dodobas> all systems nominal
21:07:38 <dodobas> :)
21:08:19 <springmeyer> ya, boost is getting large :)
21:08:31 <dodobas> over 2h of compiling
21:08:36 <artem> dodobas: moved to boost_1_42 over here - all is fine
21:08:58 <artem> dodobas : what hardware are you on?
21:09:06 <dodobas> actually a could build mapnik trunk :)
21:09:22 <dodobas> laptop Linux darch64 2.6.32-ARCH #1 SMP PREEMPT Fri Jan 29 09:10:49 CET 2010 x86_64 Intel(R) Core(TM)2 Duo CPU T7500 @ 2.20GHz GenuineIntel GNU/Linux
21:09:23 <nikq> Ticket #7500: no such ticket. (list index out of range)
21:09:45 <artem> ram ?
21:09:48 <dodobas> 4gb of ram
21:10:11 <artem> should be fine :D
21:10:50 <dodobas> mapnik svn build is memory hog?
21:11:21 <artem> dodobas : you can streamline boost build by only enabling required features - by default it builds release/debug static/shared
21:12:30 <dodobas> i hope that 1.42 will hit repos soon, but anyway i've got a running system
21:13:52 <artem> has anyone experienced gdal creating mosaics with gdal ? I'm getting alignment issues which seems to be originated in gdal. I guess my question is : how to create perfect image mosaic with gdal tools ?
21:14:07 <dodobas> image mosaic?
21:14:11 <artem> yes
21:15:48 <dodobas> ahaa, merging
21:16:13 <dodobas> mostly tiff
21:16:36 <dodobas> did not experinece problems, or i just did not see them
21:16:45 <artem> yep, exactly : how to merge and avoid black lines popping out randomly when reading vrt using gdal?
21:17:09 <dodobas> do you nees vrt?
21:17:13 <dodobas> *need
21:18:18 <artem> well, gdal_merge.py is not capable to merge all rasters (on my system) - they're quite large
21:18:40 <dodobas> you could use bigtiff
21:19:00 <dodobas> but that needs libtiff4
21:19:02 <artem> sure I'm using big tiffs > 10gb
21:19:41 <dodobas> anyway, i never experienced black lines
21:19:42 <artem> I gave up after a leaving gdal_merge for a couple of hours
21:20:06 <dodobas> i was using gdal trunk, before it was released as gdal 1.7
21:20:09 <springmeyer> artem: yesterday on #gdal even was discussing subtle issues that may be related...
21:20:29 <springmeyer> the nearblack tool might help (though I've not used it)
21:20:39 <springmeyer> and he just modified it yesterday: http://trac.osgeo.org/gdal/changeset/18712
21:21:50 <artem> ok, tnx
21:23:46 <springmeyer> and maybe also vaguely helpful - when using vrt with mapserver, I've depended on the OFFSITE param to nuke black collars
21:24:06 <springmeyer> OFFSITE in http://mapserver.org/tutorial/example1-5.html
21:24:25 <artem> it might be because I'm warping from wgs84 to spherical mercator .. OFFSITE
21:29:15 <artem> http://media.mapnik.org/black_line.jpg
21:30:05 <springmeyer> hmm, ya that looks like different issue that I was thinking
21:30:24 <springmeyer> but certainly one I've seen before
21:30:30 <artem> me too :)
21:31:06 <springmeyer> it is the map background color shining through? or something else?
21:31:06 *** Pepper has quit (Quit: Verlassend)
21:31:29 <artem> nope bg=lightblue  black comes from gdal;
21:31:47 <springmeyer> hmm
21:31:51 <artem> but perhaps I'm not using it correctly in the first place
21:32:10 <artem> I'll think how to improve my pipeline
21:33:27 <springmeyer> hmm, okay, so again slighty different that what I've seen then
21:33:46 <springmeyer> I've seen shine-through from rounding, like graphics in this thread: http://groups.google.com/group/cugos/browse_thread/thread/888c30b6791953cf
21:34:18 <springmeyer> do you get the same thing with raster.input?
21:35:07 <artem> haven't tried it with raster.input yet
21:35:54 <springmeyer> k, likely a bit big for that
21:36:38 <artem> yep
21:36:40 <springmeyer> hmm, looks like you may want to try
21:36:51 <springmeyer> SOURCE_EXTRA... arg to gdalwarp
21:36:56 * springmeyer digging up email...
21:37:20 <springmeyer> http://lists.osgeo.org/pipermail/gdal-dev/2009-January/019432.html
21:37:31 <artem> only 12gb (91 tiffs)
21:38:25 <artem> springmeyer: yep, sounds like it  ^^
21:39:01 <springmeyer> hope that helps
21:39:58 <artem> but I can't warp the whole earth on my machine as suggested , anyway thanks for pointers I'll think something out ..
21:41:05 <springmeyer> warping sub sets and then combining as vrt should work...
21:41:45 <artem> looks I need to do some extra cutting and chopping to achieve perfect fit
21:42:40 <springmeyer> k
21:44:42 <CIA-31> mapnik-utils: dane.springmeyer * r912 /trunk/serverside/cascadenik/ (cascadenik/compile.py test.py cascadenik/style.py): apply patch from macwright for line_spacing support
21:57:24 <Ldp__> springmeyer: did you get a chance to mention a 0.7.1 to artem?
21:57:41 <springmeyer> Ldp__: no, thanks for reminder
21:57:47 <springmeyer> milestone 0.7.0
21:57:48 <nikq> No Milestone for that release number
21:57:53 <springmeyer> milestone 0.7.1
21:57:53 <nikq> No Milestone for that release number
21:58:00 <springmeyer> hrm
21:58:07 <artem> Ldp__: it doesn't exists :)
21:58:40 <springmeyer> artem: I'm keeping tickets here that are nagging me:
21:58:41 <springmeyer> http://trac.mapnik.org/milestone/0.7.1
21:59:26 <artem> springmeyer: sure, looks like good candidates for a point release
21:59:28 <springmeyer> I need to test more to confirm #508, but I'm thinking a quick release could be good
21:59:29 <nikq> Ticket #508: Image type uneeded in Mapnik 0.7.0 but still enforced in load_map, http://trac.mapnik.org/ticket/508
21:59:52 <artem> springmeyer: ok
21:59:58 <springmeyer> cool
22:00:44 <springmeyer> if we do in a no more than a few weeks then that will give it a good change for debian sid, ubuntu lucid...
22:01:27 <springmeyer> #484 in particular I don't have a fix yet for but has been driving many folks mad
22:01:28 <nikq> Ticket #484: Scons: ability to ensure -L/usr/local/lib comes before -L/usr/lib, http://trac.mapnik.org/ticket/484
22:01:52 <springmeyer> who failed to build trunk, then also could not get linking right with source installed icu+boost with 0.7.0
22:02:57 *** jctull (~jctull@75.0.13.129) has joined #mapnik
22:04:03 <artem> k
22:04:32 <springmeyer> I think they are all mine to fix, unless you have ideas about #502
22:04:32 <nikq> Ticket #502: Map buffer_size not accounted for within GDAL plugin scaling logic, http://trac.mapnik.org/ticket/502
22:04:51 <springmeyer> which is my error, but need to think through solution
22:06:05 <artem> no probs I can have a look too
22:06:39 <springmeyer> cool. seems to relate a bit to other raster stuff you fixed in trunk
22:10:25 <artem> exactly
22:10:30 *** jfreeman (~jfreeman@mail.agileware.net) has joined #mapnik
22:10:35 *** jfreeman has quit (Remote host closed the connection)
22:10:43 <artem> springmeyer: can qgis read vrt raster ?
22:10:57 * artem thinks it should 
22:11:00 <springmeyer> yes
22:12:27 <artem> I just merged original bluemarble A1,A2,B1,B2 .. parts - all in wgs84 lonlat
22:12:27 <artem>  
22:12:48 <artem> and gdal gives me black seams
22:12:55 *** jfreeman (~jfreeman@mail.agileware.net) has joined #mapnik
22:13:48 <springmeyer> hmm, bummer
22:16:30 <artem> http://media.mapnik.org/bm.jpg
22:17:19 <artem> I'll try qgis
22:19:14 <artem> both single merged tiff using gdal_merge.py and vrt created with gdalbuildvrt have these lines ..
22:20:06 <artem> which lead to conclusion that it's gdal's rather than mapnik's bug
22:21:41 <springmeyer> what gdal version?
22:21:57 <artem> 1.7 i think let me check ..
22:22:17 <artem> opt/gdal/bin/gdal-config --version
22:22:18 <artem> 1.7.0
22:22:30 <springmeyer> hmm
22:23:04 <springmeyer> I'm still running 1.6 if you want me to try the blue marble test
22:23:30 <racicot> artem: are you reprojecting before mosaicing?
22:23:52 <artem> nope in this case
22:24:18 <artem> but I resample (scaled down) original rasters to be small for testing
22:24:26 <artem> using gdal
22:24:50 <racicot> is the data someplace do download?
22:25:03 <racicot> your copy of the original and the processed data
22:25:08 <artem> let me upload it just a sec ..
22:28:40 <artem> racicot: originals are quite large : http://visibleearth.nasa.gov/view_detail.php?id=7101
22:29:34 * racicot reads back up to see what the end goal is...
22:29:54 <artem> processed data + vrt is here : http://media.mapnik.org/bluemarble_test.tar.gz  (still uploading...)
22:30:24 <racicot> artem: in the end... what are you after?
22:31:14 <artem> racicot: the end goal is to be able to display/tile  big rasters using gdal.input
22:31:40 <artem> raster.input works fine
22:31:56 * artem downloading qgis for confirm ..
22:32:05 <racicot> so you just want a world wide raster that is "big"?
22:32:15 <racicot> for testing?  how big are you after?
22:32:37 * racicot might have some bluemarble stuff already processed laying around
22:32:47 <springmeyer> racicot is getting the big guns out
22:32:53 <artem> he he
22:33:41 <artem> well, I also would like to feel confident I can use gdal to create a gigantic raster of the whole planet
22:34:24 <racicot> I have created single bigtiffs using gdal that where over 600gb... never a problem.  just takes a long time
22:35:06 <artem> racicot: ok, sounds big enough :) how long did it take ?
22:35:42 <artem> and was it projected into something ?
22:35:55 <racicot> that case was for orthos for all of Puerto Rico... took like 4-5 days...
22:36:05 <artem> ok
22:36:23 <racicot> and it was reprojected into 900913
22:38:43 *** springmeyer_ (~springmey@c-76-28-156-154.hsd1.wa.comcast.net) has joined #mapnik
22:38:50 *** springmeyer_ has quit (Remote host closed the connection)
22:38:50 <racicot> artem: let me know if I can help... happy to try to dig up and process some test data for you... would just need to know what you want... both extent and target size/format/projection etc.
22:38:52 *** springmeyer has quit (Read error: Connection reset by peer)
22:39:00 *** springmeyer (~springmey@c-76-28-156-154.hsd1.wa.comcast.net) has joined #mapnik
22:40:37 <artem> racicot: thanks! I might well need some test data, but first I'd like to verify couple things ..I haven't been involved in gdal.input impl there can be some gotchas there ..
22:40:54 * springmeyer ducks
22:41:23 <artem> should I go for qgis 1.4 ?
22:41:35 <springmeyer> yes, certainly
22:41:54 <artem> 'all in one' ?
22:42:08 <racicot> springmeyer knows how to find me if you guys end up needing some help... ;-)
22:42:24 <artem> cool, cheers
22:42:35 <springmeyer> artem: on mac I recommend grabbing the Frameworks separate
22:42:53 <artem> hmm.. to late :)
22:43:09 <springmeyer> no worry, either should work
22:43:22 <springmeyer> just he updates the frameworks more often
22:47:02 <artem> right, the same story on qgis
22:50:25 <artem> ok, time to get some sleep thanks everyone !
22:50:34 *** artem has quit (Quit: artem)
23:22:06 *** StormTide has quit (Quit: Leaving)
23:38:00 *** ajturner has quit (Quit: ajturner)