Nabble - Mapnik - Users

Re: Problems with "estimate-extent"

2009-12-22T07:22:49Z

Tobias Wendorff wrote:

> Yes, we discussed this yesterday in IRC ;-))

All these nicks and other nicks, and real names, hard to keep track of
who's who these days! :)

Since more people may read this then be on irc, it won't hurt to
reiterate some points.

--
Lennard
_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: Problems with "estimate-extent"

2009-12-22T07:21:08Z

Hi,

sorry, but I'm not able to debug this by myself.

I also found this ticket #402 There was a problem with reprojecting coordinates, but when I render a map in the native projection of the data, then I also get the error.

cheers,
Jürgen

Tobias Wendorff-2 wrote:

Hi there,

spillocaster schrieb:
> Does anyone know, where the problem is and how to solve this?

I think, that's a problem of estimation in PostGIS...

AFAIK Mapnik runs this query:
select xmin(ext),ymin(ext),xmax(ext),ymax(ext) from (select
estimated_extent('<your table>','<your selection>') as ext) as tmp

Perhaps this helps for debugging?

Best regards,
Tobias
_______________________________________________
Mapnik-users mailing list
Mapnik-users@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: Problems with "estimate-extent"

2009-12-22T07:04:53Z

Lennard schrieb:
> Tobias Wendorff wrote:
>
> From a trace I saw yesterday, mapnik performs that exact same query for
> each layer, even if the queried table is the same as it encountered
> before. Even though the query usually executes in little time, there is
> opportunity to cache the results, as an enhancement to mapnik.

Yes, we discussed this yesterday in IRC ;-))
_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: Problems with "estimate-extent"

2009-12-22T06:47:57Z

Tobias Wendorff wrote:

> AFAIK Mapnik runs this query:
> select xmin(ext),ymin(ext),xmax(ext),ymax(ext) from (select
> estimated_extent('<your table>','<your selection>') as ext) as tmp
>
> Perhaps this helps for debugging?

From a trace I saw yesterday, mapnik performs that exact same query for
each layer, even if the queried table is the same as it encountered
before. Even though the query usually executes in little time, there is
opportunity to cache the results, as an enhancement to mapnik.

--
Lennard
_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: Problems with "estimate-extent"

2009-12-22T06:37:36Z

Hi there,

spillocaster schrieb:
> Does anyone know, where the problem is and how to solve this?

I think, that's a problem of estimation in PostGIS...

AFAIK Mapnik runs this query:
select xmin(ext),ymin(ext),xmax(ext),ymax(ext) from (select
estimated_extent('<your table>','<your selection>') as ext) as tmp

Perhaps this helps for debugging?

Best regards,
Tobias
_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Problems with "estimate-extent"

2009-12-22T06:20:55Z

Hi all,

I have a huge problem with rendering maps with mapnik. In some regions, mainly at the border of the data extent, some data is not fully rendered. For example, the forest is rendered properly, but when I go to the border of the country which is also the border of the data extent, the forest disappears.
This happens also with other layers like built-up areas and streets. Here is an example layer from my XML:

...
<Layer name="forest" status="on" srs="+proj=latlong +datum=WGS84">
<StyleName>forest</StyleName>
<Datasource>
<Parameter name="type">postgis</Parameter>
<Parameter name="host">localhost</Parameter>
<Parameter name="dbname">default</Parameter>
<Parameter name="user">user</Parameter>
<Parameter name="password">user</Parameter>
<Parameter name="table">(select the_geom,class from polygons) as forest</Parameter>
<Parameter name="estimate_extent">true</Parameter>
</Datasource>
</Layer>
...

This is a general problem, but I think only with huge tables. Here are some of my tables:

Table "polygons":
mapping=# SELECT count(*) from polygons;
count
---------
6485787
(1 row)

Table with streets:
mapping=# SELECT count(*) from streets;
count
---------
6386694
(1 row)

Table with paths:
mapping=# SELECT count(*) from paths;
count
---------
9684396
(1 row)

The only solution to get a correct map ist to set the parameter ="estimate_extent" to false, but then the whole table will be scanned and the rendering time is much too long. I also tried to reduce the extent of the layer for rendering smaller areas:

<Parameter name="estimate_extent">false</Parameter>
<Parameter name="extent">9.2,47.0,13.7,48.8</Parameter>

But when I want to render the whole country, then this is no solution. See also my examples with activated and deactivated estimate_extent parameter:
map_estimate_true.png
map_estimate_false.png

Does anyone know, where the problem is and how to solve this?

cheers
Jürgen

Re: Map of fictional galaxy: What's the best way to create shapefile(s) from x, y, x?

2009-12-21T15:19:34Z

On Mon, 2009-12-21 at 06:56 -0800, owntheweb wrote:

> Jon,
> Thanks for your great feedback!
>
> > The labels would end up in a field (column) in the DBF file
>
> Awesome.
>
> > If might be easiest to put all the data into a csv file: x,y,z,name
> > http://www.gdal.org/ogr/drv_csv.html
> > Then you can either convert this to a shapefile or mapnik may be able to
> > read this directly via the ogr driver.
>
> I'll check out the csv option as that hugely simplifies things.

Once you have a the csv file then ogr2ogr should do the work of
converting to a shapefile.

> > The scale is arbitrary and can be whatever you want. Often it is in
> > degress (lattitude/longitude) or metres or feet. In your case it could
> > be millons of km, light-years or whatever make most sense to you.
>
> The scale is clear as crystal now.
>
> > Are you using the 'z' values for 3D data?
> > If so then Mapnik may not work for you. It currently only does 2D maps
> > and typically ignores the z value of a 3D geometry.
>
> 'z' is out as the in-game 3D map tends to confuse people anyway. A
> consistent 2D map that doesn't rotate every which way will be helpful.
>
> http://old.nabble.com/file/p26874951/mapDesign1A.gif

Ignoring Z does make things much easier.

> Do you see any rendering challenges with the style plan? Right now the guess
> is that I'll need to make multiple layers for the points as they have double
> borders. (I'll dive into this area more soon)

It looks like you will need one layer for the plents and another for the
lines between them.

Mapnik currently has fairly limited ability to draw geometric shapes
itself. The best way to draw the points is probably to create icons in
the 4 different colours, each with the spot and the surrounding ring.
Then you'll need to use the ShieldSymbolizer[1] with rules which choose
the right colour icon and specify the text label.

Alternatively you could use a PointSymbolizer + TextSymbolizer to draw
the icon and text separately but the ShieldSymbolizer is probably better
since it does both at the same time.

For the lines between the planets you will need to create a table with
which lines you want to be shown. Then you can render the dashed
appearance using the LineSymbolizer using the stroke-dasharray
feature[2].

Jon

1: http://trac.mapnik.org/wiki/ShieldSymbolizer
2: http://trac.mapnik.org/wiki/LineSymbolizer

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Stumbled on bug #13, Mapnik does not display TextSymbolizer labels longer than the supporting line

2009-12-21T07:59:18Z

Hello,

For the record, in MapOSMatic, maps for cities with index
(http://maposmatic.org/), we stumbled on bug #13
(http://trac.mapnik.org/ticket/13): TextSymbolizer does not display
the label when it is longer than the supporting line.

I read recent threads on the subject but unfortunately there is no
easy solution. :-(
https://lists.berlios.de/pipermail/mapnik-users/2009-December/thread.html#2584
http://lists.berlios.de/pipermail/mapnik-users/2009-October/thread.html#2384

Our own maposmatic bug: https://savannah.nongnu.org/bugs/index.php?27467

Best regards,
david
_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: Map of fictional galaxy: What's the best way to create shapefile(s) from x, y, x?

2009-12-21T06:56:11Z

Jon,
Thanks for your great feedback!

> The labels would end up in a field (column) in the DBF file

Awesome.

> If might be easiest to put all the data into a csv file: x,y,z,name
> http://www.gdal.org/ogr/drv_csv.html
> Then you can either convert this to a shapefile or mapnik may be able to
> read this directly via the ogr driver.

I'll check out the csv option as that hugely simplifies things.

> The scale is arbitrary and can be whatever you want. Often it is in
> degress (lattitude/longitude) or metres or feet. In your case it could
> be millons of km, light-years or whatever make most sense to you.

The scale is clear as crystal now.

> Are you using the 'z' values for 3D data?
> If so then Mapnik may not work for you. It currently only does 2D maps
> and typically ignores the z value of a 3D geometry.

'z' is out as the in-game 3D map tends to confuse people anyway. A consistent 2D map that doesn't rotate every which way will be helpful.

Do you see any rendering challenges with the style plan? Right now the guess is that I'll need to make multiple layers for the points as they have double borders. (I'll dive into this area more soon)

Best regards,

Chris

Jon Burgess-2 wrote:

On Sun, 2009-12-20 at 21:16 -0800, owntheweb wrote:
> http://old.nabble.com/file/p26869227/mapDesign1A.gif
>
> Hello gurus,
> I'm hoping to make a galactic map of the fictional galaxy from the
> EVE-Online game. I have the data (x,y,z coordinates). I set a style game
> plan (seen above, created in Illustrator). I've done some research and
> created a couple "Hello world" images. Now I'm guessing I need to create a
> shapefile?

Are you using the 'z' values for 3D data?
If so then Mapnik may not work for you. It currently only does 2D maps
and typically ignores the z value of a 3D geometry.

> One tool that may be helpful is Shapelib. There's some documentation at:
> http://shapelib.maptools.org/shapelib-tools.html
>
> Using shapelib though, I'm not sure what kind of numbers/scale I should be
> using. The polygon example from the link above uses numbers like 100000 to
> 7000000. Is there an ideal scale to use? Is there another shapefile
> generator you would recommend? Where would the labels go?

The scale is arbitrary and can be whatever you want. Often it is in
degress (lattitude/longitude) or metres or feet. In your case it could
be millons of km, light-years or whatever make most sense to you.

The labels would end up in a field (column) in the DBF file

If might be easiest to put all the data into a csv file: x,y,z,name
http://www.gdal.org/ogr/drv_csv.html

Then you can either convert this to a shapefile or mapnik may be able to
read this directly via the ogr driver.

> http://old.nabble.com/file/p26869227/Picture2.png
>
> Here's an image of the entire galaxy rendered in Flash. It would be awesome
> to get this thing in a slidy map instead (better usability for a fun in-game
> project).
>
> I'd appriciate any feedback you may have. Please be kind as I'm new to
> mapnik. If anything though I should have a map of my hometown running soon.
> Mapnik rocks my world.

Handling 3D will probably be the biggest issue with Mapnik. If you know
what you are doing then you could transform the data to let you see the
map from an arbitrary 3D viewpoint before sending the data as 2D to
Mapnik.

If you really want a flexible 3D rendering then you might need to look
at something which is designed specifically for this, like
http://www.blender.org/

Jon

_______________________________________________
Mapnik-users mailing list
Mapnik-users@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: Update on Trunk and 0.7.0 release

2009-12-21T04:32:09Z

Am Mo, 21.12.2009, 09:37 schrieb Tom Hughes:

> The problem with going all bleeding edge like that is that it makes it
> very hard for people to work with the code. I'm running a full up to
> date F12, an operating system released only a few weeks ago, and yet I
> only have version 1.39.0 of boost so I'm unlikely to be able to do any
> work on mapnik now for at least six months.

I had the same problem using Debian Lenny...

My workaround:
I've compiled Mapnik2 in a chroot-environment with self-build Boost v1.41
in there.

Then I copied the app "out" to my normal system.

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: dirtying adjacent tiles

2009-12-21T04:18:09Z

Hi!

Am 15.12.2009 20:39, schrieb Aryeh Leib Taurog:
> Does the mapnik API provide an easy way to get the maximum
> PointSymbolizer icon size for a given zoom, so I can determine myself
> whether or not the icon will run off the edge onto the next tile? Is
> there an easier way to do this?

Usually, this problem is circumvented by always rendering with a larger
canvas with an additional border around the tiles, which is discarded.
This way, icons and texts overlapping an edge are drawn twice and part
of them is on one tile and part on the other.

Mapnik has a feature to create such a buffer border.

bye
Nop
_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: Map of fictional galaxy: What's the best way to create shapefile(s) from x, y, x?

2009-12-21T04:13:54Z

On Sun, 2009-12-20 at 21:16 -0800, owntheweb wrote:
> http://old.nabble.com/file/p26869227/mapDesign1A.gif
>
> Hello gurus,
> I'm hoping to make a galactic map of the fictional galaxy from the
> EVE-Online game. I have the data (x,y,z coordinates). I set a style game
> plan (seen above, created in Illustrator). I've done some research and
> created a couple "Hello world" images. Now I'm guessing I need to create a
> shapefile?

Are you using the 'z' values for 3D data?
If so then Mapnik may not work for you. It currently only does 2D maps
and typically ignores the z value of a 3D geometry.

> One tool that may be helpful is Shapelib. There's some documentation at:
> http://shapelib.maptools.org/shapelib-tools.html
>
> Using shapelib though, I'm not sure what kind of numbers/scale I should be
> using. The polygon example from the link above uses numbers like 100000 to
> 7000000. Is there an ideal scale to use? Is there another shapefile
> generator you would recommend? Where would the labels go?

The scale is arbitrary and can be whatever you want. Often it is in
degress (lattitude/longitude) or metres or feet. In your case it could
be millons of km, light-years or whatever make most sense to you.

The labels would end up in a field (column) in the DBF file

If might be easiest to put all the data into a csv file: x,y,z,name
http://www.gdal.org/ogr/drv_csv.html

Then you can either convert this to a shapefile or mapnik may be able to
read this directly via the ogr driver.

> http://old.nabble.com/file/p26869227/Picture2.png
>
> Here's an image of the entire galaxy rendered in Flash. It would be awesome
> to get this thing in a slidy map instead (better usability for a fun in-game
> project).
>
> I'd appriciate any feedback you may have. Please be kind as I'm new to
> mapnik. If anything though I should have a map of my hometown running soon.
> Mapnik rocks my world.

Handling 3D will probably be the biggest issue with Mapnik. If you know
what you are doing then you could transform the data to let you see the
map from an arbitrary 3D viewpoint before sending the data as 2D to
Mapnik.

If you really want a flexible 3D rendering then you might need to look
at something which is designed specifically for this, like
http://www.blender.org/

Jon

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: [Mapnik-devel] Update on Trunk and 0.7.0 release

2009-12-21T03:28:28Z

You can always link yourself to the 0.7.0 branch and continue working with that.

Lucio

On Mon, Dec 21, 2009 at 9:37 AM, Tom Hughes <tom@...> wrote:

> On 21/12/09 06:34, Dane Springmeyer wrote:
>
>> As of http://trac.mapnik.org/changeset/1495 Artem has merged the
>> Mapnik2 branch intro trunk. This is really excited, but we warned that
>> now current trunk requires Boost 1.41 (latest release) and introduces
>> a few subtle backward incompatible changes (they are needed) to
>> support Expressions, and we've just barely started documenting here: http://trac.mapnik.org/wiki/Mapnik2
>
> The problem with going all bleeding edge like that is that it makes it
> very hard for people to work with the code. I'm running a full up to
> date F12, an operating system released only a few weeks ago, and yet I
> only have version 1.39.0 of boost so I'm unlikely to be able to do any
> work on mapnik now for at least six months.
>
> Not that I've done much of late anyway, but I just thought I'd point out
> that you limit the community you can involve when you require bleeding
> edge dependencies.
>
> Tom
>
> --
> Tom Hughes (tom@...)
> http://www.compton.nu/
> _______________________________________________
> Mapnik-devel mailing list
> Mapnik-devel@...
> https://lists.berlios.de/mailman/listinfo/mapnik-devel
>

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: Update on Trunk and 0.7.0 release

2009-12-21T00:37:25Z

On 21/12/09 06:34, Dane Springmeyer wrote:

> As of http://trac.mapnik.org/changeset/1495 Artem has merged the
> Mapnik2 branch intro trunk. This is really excited, but we warned that
> now current trunk requires Boost 1.41 (latest release) and introduces
> a few subtle backward incompatible changes (they are needed) to
> support Expressions, and we've just barely started documenting here: http://trac.mapnik.org/wiki/Mapnik2

The problem with going all bleeding edge like that is that it makes it
very hard for people to work with the code. I'm running a full up to
date F12, an operating system released only a few weeks ago, and yet I
only have version 1.39.0 of boost so I'm unlikely to be able to do any
work on mapnik now for at least six months.

Not that I've done much of late anyway, but I just thought I'd point out
that you limit the community you can involve when you require bleeding
edge dependencies.

Tom

--
Tom Hughes (tom@...)
http://www.compton.nu/
_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Update on Trunk and 0.7.0 release

2009-12-20T22:34:16Z

Mapnik Users/Dev,

As of http://trac.mapnik.org/changeset/1495 Artem has merged the
Mapnik2 branch intro trunk. This is really excited, but we warned that
now current trunk requires Boost 1.41 (latest release) and introduces
a few subtle backward incompatible changes (they are needed) to
support Expressions, and we've just barely started documenting here: http://trac.mapnik.org/wiki/Mapnik2
.

Users that were previously running trunk and wish to stay on that
codebase it is easy to switch to the new 0.7 branch:

$ cd mapnik-trunk-checkout
$ svn switch http://svn.mapnik.org/branches/0.7/ . #note the 'dot'

Also, we're going to shift the 0.7.0 release date forward to just
after the New Year to give a bit more time for testing and attention
to the last issues.

If you have a few moments over the holidays I encourage you to pull
freshly from the 0.7 branch and test building mapnik and running in
your existing applications.

http://trac.mapnik.org/milestone/0.7.0

Please report any problems or issues to Trac tickets.

Thanks!

Dane

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Map of fictional galaxy: What's the best way to create shapefile(s) from x,y,x?

2009-12-20T21:16:49Z

Hello gurus,
I'm hoping to make a galactic map of the fictional galaxy from the EVE-Online game. I have the data (x,y,z coordinates). I set a style game plan (seen above, created in Illustrator). I've done some research and created a couple "Hello world" images. Now I'm guessing I need to create a shapefile?

One tool that may be helpful is Shapelib. There's some documentation at: http://shapelib.maptools.org/shapelib-tools.html

Using shapelib though, I'm not sure what kind of numbers/scale I should be using. The polygon example from the link above uses numbers like 100000 to 7000000. Is there an ideal scale to use? Is there another shapefile generator you would recommend? Where would the labels go?

Here's an image of the entire galaxy rendered in Flash. It would be awesome to get this thing in a slidy map instead (better usability for a fun in-game project).

I'd appriciate any feedback you may have. Please be kind as I'm new to mapnik. If anything though I should have a map of my hometown running soon. Mapnik rocks my world.

Thank you,

Chris

Re: [OSM-talk] Projection for processed_p.shp?

2009-12-18T04:02:08Z

Am Do, 17.12.2009, 23:32 schrieb Frederik Ramm:

> processed_p.shp is spherical mercator (properly EPSG:3857, formerly
> EPSG:900913, and erroneously EPSG:3857).

3785 ;-)

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: [OSM-talk] Projection for processed_p.shp?

2009-12-18T01:55:47Z

Frederik Ramm schrieb am 17.12.2009 23:43:
> Hi,
>
> Frederik Ramm wrote:
>> processed_p.shp is spherical mercator (properly EPSG:3857, formerly
>> EPSG:900913, and erroneously EPSG:3857).
>

Thanks, that was the problem. With http://spatialreference.org/ref/epsg/3785/ it
works.

Alex

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

creating a database and selecting data through c++

2009-12-18T00:50:47Z

hi,

I have seen some mapnik reply mails :

http://lists.berlios.de/pipermail/mapnik-users/2007-July/000339.html

https://lists.berlios.de/pipermail/mapnik-users/2007-June/000300.html

and the page : http://trac.mapnik.org/wiki/OptimizeRenderingWithPostGIS

My question is as follows:

I have added the following code to rundemo.cpp !

{

parameters p;

p["type"]="postgis";

p["host"]= "abc";

p["port"]=5432;

p["dbname"]="gis";

p["user"]="anisha";

p["password"]="";

p["table"]="tree";

Layer lyr("Roads");

lyr.set_datasource(datasource_cache::instance()->create(p));

lyr.add_style("roads");

}

I have installed the postgis plugin through this command :

Que 1: Now how i am supposed to check whether a table tree has been created or not ?

Que 2: Where should write the query "select * from tree", in rundemo.cpp ? This query is supposed to be independent of table creation !!

Thanks ,

_anisha



-------------------------------------
Hi-Tech Gears Limited, Gurgaon, India

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Semi-transparent background?

2009-12-17T23:50:49Z

Hi!

I am creating my tiles with a semi-transparent background and polygons
(but with solid streets and pois) in order to superimpose them on a
hillshading layer. Currently I am achieving this by rendering with
bgcolor="transparent" and compositing the result on a semi-transparent
gray image in python PIL.

Now I am trying to switch to on-demand rendering on the server with
renderd, where the mapnik calls are hardcoded in renderd and only the
style file is easily changed.

Is there a way to get a semi-transparent background directly from mapnik?

Or is there a way to render with a transparent background but define a
layer that draws a semi-transparent object over everything without
having geometry for that in the DB?

thanks
Nop

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: [OSM-talk] Projection for processed_p.shp?

2009-12-17T15:03:39Z

EPSG, too, seems to have problems with tagging. My understanding is that that officially it is nowadays EPSG:3785, see
http://spatialreference.org/ref/epsg/3785/

However, that code is not widely supported yet. I believe EPSG:900913 is still the most common in use. Best to keep the spatialrefererence page open and learn to give thy cryptic WKT or Proj4 strings if nothing else helps.

-Jukka Rahkonen-

-----Alkuperäinen viesti-----
Lähettäjä: mapnik-users-bounces@... puolesta: Frederik Ramm
Lähetetty: pe 18.12.2009 0:43
Vastaanottaja: Alexrk
Kopio: talk@...; mapnik-users@...
Aihe: Re: [Mapnik-users] [OSM-talk] Projection for processed_p.shp?

Hi,

Frederik Ramm wrote:
> processed_p.shp is spherical mercator (properly EPSG:3857, formerly
> EPSG:900913, and erroneously EPSG:3857).

The last bit should've been 3785. It was bound to happen.

Bye
Frederik

--
Frederik Ramm ## eMail frederik@... ## N49°00'09" E008°23'33"
_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: [OSM-talk] Projection for processed_p.shp?

2009-12-17T14:43:00Z

Hi,

Frederik Ramm wrote:
> processed_p.shp is spherical mercator (properly EPSG:3857, formerly
> EPSG:900913, and erroneously EPSG:3857).

The last bit should've been 3785. It was bound to happen.

Bye
Frederik

--
Frederik Ramm ## eMail frederik@... ## N49°00'09" E008°23'33"
_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: [OSM-talk] Projection for processed_p.shp?

2009-12-17T14:32:10Z

Hi,

Alexrk wrote:
>> The projection is spherical mercator (900913), you can download a .PRJ
>> file from the menu on
>>
>> http://spatialreference.org/ref/sr-org/6627/
>
> I've tried it in ArcMap - doesn't fit with a regular Mercator .prj file.

Yes; "regular Mercator" is something other than "spherical Mercator".

> I don't know, what projection this file actually is :(

processed_p.shp is spherical mercator (properly EPSG:3857, formerly
EPSG:900913, and erroneously EPSG:3857).

If you don't like it, you can use ogr2ogr to reproject the shape to
something else.

Bye
Frederik

--
Frederik Ramm ## eMail frederik@... ## N49°00'09" E008°23'33"
_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: Projection for processed_p.shp?

2009-12-17T14:20:52Z

Frederik Ramm schrieb am 09.11.2009 02:23:
> The projection is spherical mercator (900913), you can download a .PRJ
> file from the menu on
>
> http://spatialreference.org/ref/sr-org/6627/

I've tried it in ArcMap - doesn't fit with a regular Mercator .prj file.

I don't know, what projection this file actually is :(

Alex

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: mapnik.render causing segmentation fault

2009-12-17T14:09:49Z

Nop,

Did this issue get resolved?

Dane

On Dec 13, 2009, at 3:10 PM, Nop wrote:

>
> Hi!
>
> Am 13.12.2009 23:25, schrieb Dane Springmeyer:
>> install cairo's python bindings like:
>>
>> sudo apt-get install python-cairo
>>
>> and I think the problem will be fixed.
>
> Before I ask my admin, two questions:
>
> - Does this explain why it crashes on render(), but it does work
> with render_to_file() ?
>
> - Would this be fixed in 0.7?
>
>
> thanks
> Nop

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: Mercator output blank

2009-12-17T14:05:33Z

Andrew,

Glad you got it figured, this is indeed an issue with invalid
geometries that Mapnik's proj4 implementation can't cope with:

http://trac.mapnik.org/ticket/308

In the case of the world_borders file try using this one called
'tm_wgs84_sans_antarctica' here:

http://mapnik-utils.googlecode.com/svn/data/

Dane

On Dec 16, 2009, at 6:46 PM, Andrew Frieze wrote:

> Ok, I think I have this figured out.
>
> Apparently the world_borders file has some trouble projecting. I am
> able to successfully render a different shp file.
>
>
> On Wed, Dec 16, 2009 at 11:55 AM, Andrew Frieze <afrieze@...>
> wrote:
> Hi all
>
> I just discovered Mapnik and am very excited about the potential I
> see here. I have run into a problem that I am hoping is easily
> solved by this community.
>
> I followed the HelloWorld XML tutorial using the world_borders file
> successfully. I would now like to have my output image rendered to
> a Mercator projection(for google). Unfortunately, when I attempt
> this I just get a blank tile.
>
> Here is a summary of my code(some parts removed for readability):
>
> mapfile = 'world_styles.xml'
> m = Map(256, 256,"+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")
> load_map(m, mapfile)
> googleProj = Projection("+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")
>
> c0 = googleProj.forward(Coord(leftLong, botLat))
> c1 = googleProj.forward(Coord(rightLong, topLat))
>
> bbox = Envelope(c0.x,c0.y,c1.x,c1.y)
> m.zoom_to_box(bbox)
> map_output = 'H:/inetpub/wwwroot/MapDev/images/image.png'
> render_to_file(m, map_output)
>
>
> Thanks!
> Andrew
>
>
> _______________________________________________
> Mapnik-users mailing list
> Mapnik-users@...
> https://lists.berlios.de/mailman/listinfo/mapnik-users

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: installing mapnik on suse 10.3

2009-12-17T14:03:17Z

On Dec 15, 2009, at 9:14 PM, Anisha Kaul wrote:

>
> This is what i found on my system !
>
> cairomm-1.2.2-85
> python-cairo-1.2.2-91
> cairo-devel-1.4.10-25
> cairo-doc-1.4.10-25
> cairo-1.4.10-25
>
> I also found:
>
> proj4:
>
> /usr/share/doc/packages/proj4
> /usr/share/doc/packages/proj4/AUTHORS
> /usr/share/doc/packages/proj4/ChangeLog
> /usr/share/doc/packages/proj4/COPYING
> /usr/share/doc/packages/proj4/NEWS
> /usr/share/doc/packages/proj4/README
>

this looks like the source and not the installed proj4. How did you
install it?

> boost-filesystem:
>
> /home/anisha/Desktop/boost_1_39_0/libs/filesystem/test/msvc/
> boost_filesystem.sln

what is that? The source checkout or where you installed it? I doubt
that is going to work.

> /usr/lib/libboost_filesystem.a
> /usr/lib/libboost_filesystem.so
> /usr/lib/libboost_filesystem.so.1.33.1
>

These looks like properly installed libraries, but Mapnik requires
>1.34 so 1.33 will not work.

> Now how should i proceed with the installation ?

compile and install boost with the default settings and it should
install in /usr/local, then do:

python scons/scons.py BOOST_LIBS=/usr/local/lib BOOST_INCLUDES=/usr/
local/lnclude

If that does not work then send the terminal output and attach the
'config.log' file.

Dane
_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: Announcing Mapnik OS X installers

2009-12-17T02:04:57Z

Great work, Dane!
Cheers
Artem

2009/12/17 Dane Springmeyer <blake@...>:

> I've released the first public offering of the Mapnik OS X Framework
> installers:
>
> http://mapnik.org/news/2009/dec/16/osx_installers/
>
> They currently provide the latest from the 0.7 branch which will be
> soon promoted to the next release.
>
> My hope is that one side benefit of having development releases as
> easy-to-install binaries is that more users will help find problems
> and create tickets before each release.
>
> If you have suggestions for improving the installers please file Trac
> tickets against the "Mac OS X Binaries" component.
>
> Dane
>
> _______________________________________________
> Mapnik-users mailing list
> Mapnik-users@...
> https://lists.berlios.de/mailman/listinfo/mapnik-users
>

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Announcing Mapnik OS X installers

2009-12-16T23:30:59Z

I've released the first public offering of the Mapnik OS X Framework
installers:

http://mapnik.org/news/2009/dec/16/osx_installers/

They currently provide the latest from the 0.7 branch which will be
soon promoted to the next release.

My hope is that one side benefit of having development releases as
easy-to-install binaries is that more users will help find problems
and create tickets before each release.

If you have suggestions for improving the installers please file Trac
tickets against the "Mac OS X Binaries" component.

Dane

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: Mercator output blank

2009-12-16T18:46:20Z

Ok, I think I have this figured out.

Apparently the world_borders file has some trouble projecting. I am able to successfully render a different shp file.

On Wed, Dec 16, 2009 at 11:55 AM, Andrew Frieze <afrieze@...> wrote:

Hi all

I just discovered Mapnik and am very excited about the potential I see here. I have run into a problem that I am hoping is easily solved by this community.

I followed the HelloWorld XML tutorial using the world_borders file successfully. I would now like to have my output image rendered to a Mercator projection(for google). Unfortunately, when I attempt this I just get a blank tile.

Here is a summary of my code(some parts removed for readability):

mapfile = 'world_styles.xml'
m = Map(256, 256,"+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")
load_map(m, mapfile)
googleProj = Projection("+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")

c0 = googleProj.forward(Coord(leftLong, botLat))
c1 = googleProj.forward(Coord(rightLong, topLat))

bbox = Envelope(c0.x,c0.y,c1.x,c1.y)
m.zoom_to_box(bbox)
map_output = 'H:/inetpub/wwwroot/MapDev/images/image.png'
render_to_file(m, map_output)

Thanks!
Andrew

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Bottlenecks for performance when rendering open street map world wide data via Mapnik OGC server

2009-12-16T09:27:52Z

I would like to display the entire world wide open street map data via Mapnik which is pointing to this world wide data on postgres/postgis.

The web clients will be sending WMS requests to display the open street map data via Mapnik OGC server.

I have successfully prototyped the above for just the open street map data within Australia.

Now I would like to import the entire world into postgres/postgis and then eventually have open street map as our base layer for our GIS production application.

I would like to know what are generally the bottlenecks for performance when rendering open street map world wide data via Mapnik OGC server such as:

a.) Number of CPU's
b.) Amount of RAM
c.) Speed of RAM
d.) Hard drive speed/size
e.) CPU speed
f.) Type of server Windows or Linux
g.) anything else that I missed?

Which item(s) above would normally be the bottleneck?

Thanks,

John

--
John J. Mitchell

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Mercator output blank

2009-12-16T08:55:08Z

Hi all

I just discovered Mapnik and am very excited about the potential I see here. I have run into a problem that I am hoping is easily solved by this community.

I followed the HelloWorld XML tutorial using the world_borders file successfully. I would now like to have my output image rendered to a Mercator projection(for google). Unfortunately, when I attempt this I just get a blank tile.

Here is a summary of my code(some parts removed for readability):

mapfile = 'world_styles.xml'
m = Map(256, 256,"+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")
load_map(m, mapfile)
googleProj = Projection("+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")

c0 = googleProj.forward(Coord(leftLong, botLat))
c1 = googleProj.forward(Coord(rightLong, topLat))

bbox = Envelope(c0.x,c0.y,c1.x,c1.y)
m.zoom_to_box(bbox)
map_output = 'H:/inetpub/wwwroot/MapDev/images/image.png'
render_to_file(m, map_output)

Thanks!
Andrew

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Re: installing mapnik on suse 10.3

2009-12-15T21:14:40Z

>
> First, the INCLUDES/LIBS directives expect a directory like /usr/
> include or /usr/lib not the actual library name.
>
> Also, in the specific case of Cairo, pkg-config is actually required
> to locate it and Scons will not use the options you have sent.
>
> For Cairo you need to make sure pkg-config is found on your system and
> it is able to locate Cairo.
>
> On my system I can test the command:
>
> $ /opt/local/bin/pkg-config --libs cairo
> -L/opt/local/lib -lcairo -lpixman-1 -lfontconfig -lexpat -lfreetype -
> liconv -lpng12 -lz
>
> But, in general the Cairo dependency is optional, so I would focus on
> trying to satisfy the other requirements it appears you are
> missing.... proj4, icu, and boost filesystem.
>
> See also for more info:
>
> http://trac.mapnik.org/wiki/UsingScons
>
> Dane

Thanks for the reply !

This is what i found on my system !

cairomm-1.2.2-85
python-cairo-1.2.2-91
cairo-devel-1.4.10-25
cairo-doc-1.4.10-25
cairo-1.4.10-25

I also found:

proj4:

/usr/share/doc/packages/proj4
/usr/share/doc/packages/proj4/AUTHORS
/usr/share/doc/packages/proj4/ChangeLog
/usr/share/doc/packages/proj4/COPYING
/usr/share/doc/packages/proj4/NEWS
/usr/share/doc/packages/proj4/README

boost-filesystem:

/home/anisha/Desktop/boost_1_39_0/libs/filesystem/test/msvc/boost_filesystem.sln
/usr/lib/libboost_filesystem.a
/usr/lib/libboost_filesystem.so
/usr/lib/libboost_filesystem.so.1.33.1

Now how should i proceed with the installation ?

regards,
anisha

-------------------------------------
Hi-Tech Gears Limited, Gurgaon, India

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Non-standard maps: galaxies? hotels?

2009-12-15T19:28:12Z

Howdy,
Forgive my ignorance. I'm new to mapnik but am fascinated by the beautiful
maps I'm seeing.

While digging into tools offered at mapbox.com, I'm hope to to build a
custom galaxy map based on the fictitious galaxy of EVE-online. It's odd
yes, but fun projects like this help make things fun for future drier
projects like exhibitors at a hotel ballroom.

Any thoughts on where to get started? Should I start with some normal local
maps and work my way into outer space? Can I work with my own vector files?
I do have access to data somewhere where I can generate fake coordinates for
everything, and could draw stuff pragmatically. No problem there, just not
sure where to start digging.

Super ugly rough I created of what I'm talking about:

http://old.nabble.com/file/p26805511/Picture2.png

Each dot has a name. They connect to neighboringish dots (jump gates) using
other data I can look up. It would be cool to show this galactic highway,
with some landmarks and such. :)

I appreciate any feedback/ideas you may have. Thanks much in advance!

EDIT: I found some great resources here regarding shapefiles and such:
http://trac.mapnik.org/wiki/LearningMapnik

Maybe it's good to start from the beginning! Any feedback/pointers are still
welcome.

--
View this message in context: http://old.nabble.com/Non-standard-maps%3A-galaxies--hotels--tp26805511p26805511.html
Sent from the Mapnik - Users mailing list archive at Nabble.com.

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users

Non-standard maps: galaxies? hotels?

2009-12-15T18:59:40Z

Howdy,
Forgive my ignorance. I'm new to mapnik but am fascinated by the beautiful
maps I'm seeing.

While digging into tools offered at mapbox.com, I'm hope to to build a
custom galaxy map based on the fictitious galaxy of EVE-online. It's odd
yes, but fun projects like this help make things fun for future drier
projects like exhibitors at a hotel ballroom.

Any thoughts on where to get started? Should I start with some normal local
maps and work my way into outer space? Can I work with my own vector files?
I do have access to data somewhere where I can generate fake coordinates for
everything, and could draw stuff pragmatically. No problem there, just not
sure where to start digging.

Super ugly rough I created of what I'm talking about:

http://old.nabble.com/file/p26805511/Picture2.png

Each dot has a name. They connect to neighboringish dots (jump gates) using
other data I can look up. It would be cool to show this galactic highway,
with some landmarks and such. :)

I appreciate any feedback/ideas you may have. Thanks much in advance!

--
View this message in context: http://old.nabble.com/Non-standard-maps%3A-galaxies--hotels--tp26805511p26805511.html
Sent from the Mapnik - Users mailing list archive at Nabble.com.

_______________________________________________
Mapnik-users mailing list
Mapnik-users@...
https://lists.berlios.de/mailman/listinfo/mapnik-users