pymag-devel — pymagellan development

[Pymag-devel] Updated OSM Tags

[Ralf D. <dr...@vi...>]

Hi,

pymag is very useful for me, thanks for developing it.

I would like to contrubute updated rules for new OSM tags and a shell
script to automatically download and add DEM data. Using this script,
creating high quality topo maps for Magellan from OSM data gets really
simple.

Thanks,
Ralf

[Pymag-devel] Restarting OSM for Magellan

[Matthias M. <di...@ar...>]

Hi all


I'm from the OpenStreetmap project and always interested in community
building and promoting OSM.
http://wiki.openstreetmap.org/wiki/User:!i!

We would like to get a better support of Magellan maps and a open
process to create maps out of our data.
For that reason I try to bring the people together and I'm in contact
with Magellan to ask for support
http://wiki.openstreetmap.org/wiki/OSM_Map_On_Magellan

So to anybody out there, please help us to make the custom map support
even better than for Garmin :)


bye
Matthias

Re: [Pymag-devel] TypeError: unsupported operand

[BP <bp...@po...>]

After further investigation it seems that some cell elements koords are 
not discretized

When entering get_best_cell with good walues (all values are integer):

Get_Best_Cell max_cell_level: 3  bounds: 
Rec((1806231,-5393593),(2518935,-4680889))  r: 
Rec((2255666,-5357273),(2257702,-5355556))
Get_Best_Cell i: 8
Get_Best_Cell v: 89088  h: 89088
Get_Best_Cell n: 0  s: 0  w: 5  e: 5

When exception occures the input values are float:

Get_Best_Cell max_cell_level: 3  bounds: 
Rec((1795075.0,-5395331.0),(2540347.0,-4650059.0))  r: 
Rec((2154286.0,-5316756.0),(2154286.0,-5316756.0))
Get_Best_Cell i: 8
Get_Best_Cell v: 93159.0  h: 93159.0
Get_Best_Cell n: 0.843450444938  s: 0.843450444938  w: 3.85589154027  e: 
3.85589154027
Traceback (most recent call last): ...


Peter

[Pymag-devel] TypeError: unsupported operand

[BP <bp...@po...>]

Hi!

I get this error when osm file contains a lot of pois or points.


Regards
Peter

DEBUG Optimizing layer 00_Major_cities
DEBUG nlevels=0 for layer 00_Major_cities
DEBUG Optimizing layer 00_Large_cities
DEBUG nlevels=0 for layer 00_Large_cities
DEBUG Optimizing layer 00_Medium_cities
DEBUG Max cell elements exceeded for layer 00_Medium_cities. cellnum=1, 
# of cell elements=4127
DEBUG Max cell elements exceeded for layer 00_Medium_cities. cellnum=1, 
# of cell elements=4127
DEBUG nlevels=2 for layer 00_Medium_cities
WARNING bbox should be a multiple of minimum cell size, adjusting bbox 
borders
Traceback (most recent call last):
   File "/home/peter/pymag/osmmagellan/osmmag.py", line 260, in <module>
     main()
   File "/home/peter/pymag/osmmagellan/osmmag.py", line 256, in main
     db=options.db
   File "/home/peter/pymag/osmmagellan/osmmag.py", line 97, in osmmag
     m.close()
   File "/usr/local/lib/python2.6/dist-packages/magellan/Map.py", line 
369, in close
     group.optimizeLayers()
   File 
"/usr/local/lib/python2.6/dist-packages/magellan/SearchGroup.py", line 
116, in optimizeLayers
     cellrefremap[self.map.getLayerIndex(layer)] = layer.optimize()
   File "/usr/local/lib/python2.6/dist-packages/magellan/Layer.py", line 
480, in optimize
     newcellrefs = self.addCellElement(ce)
   File "/usr/local/lib/python2.6/dist-packages/magellan/Layer.py", line 
750, in addCellElement
     self.nlevels)
   File "/usr/local/lib/python2.6/dist-packages/magellan/Layer.py", line 
1086, in get_best_cell
     m = (n ^ s) | (w ^ e)
TypeError: unsupported operand type(s) for ^: 'numpy.float64' and 
'numpy.float64'

[Pymag-devel] Long ways displayed incorrectly in MapSend

[Aleksandr I. <vit...@gm...>]

osmmag.py -n Bug -r ./rules.xml -o ./bug.imi
--name-tag-list=name:ru,name --debug ./bug.osm

[Pymag-devel] New revision control system

[Henrik J. <he...@jo...>]

Hi all,

I have decided to change revision control system from subversion to git. The 
reason is that git offers some nice features that subversion doesn't that I 
got used to in other projects. 

The command to checkout the latest git version is:
git clone git://pymag.git.sourceforge.net/gitroot/pymag/pymag 

Hope this will not cause too much pain for you. One can always download a 
tarball from the SF git web gui. 
 
/Henrik

Re: [Pymag-devel] Points that cause MapSend to crash

[Mikael A. <mik...@gm...>]

Great news indeed! I'll give it a try as soon as I can!

/ Mikael

On Sat, Aug 14, 2010 at 19:24, Henrik Johansson <he...@jo...> wrote:
> On Wednesday 04 November 2009 11:04:41 Henrik Johansson wrote:
>> On Sunday 26 July 2009 20:08:11 Mikael Arvids wrote:
>> > Great project! Definitely in my top ten list of open source projects
>> > here on sourceforge!
>> >
>> >  I've stumbled on some points that causes MapSend to crash.
>> >
>> > The first node in the badPOI map is a shopping mall in Borlänge called
>> > "Kupolen". I've generated the map with the skimap.xml rules in the
>> > svn-tree. This makes the node a POI in the imi-file, and if this is
>> > present it will crash MapSend.
>> >
>> > The second node in the badPoint map is just a kindergarten in
>> > Linköping. The map is generated with the osmmagrules.xml rules. This
>> > will only add the node as a named point in the imi-file, but it seems
>> > to be enough to put MapSend out of its misery. It seems though that
>> > the problem only exist if I request that the names should be added
>> > (with --name-tag-list=int_name,name). If I omit this MapSend will
>> > display the map without any problem.
>>
>> Hi Mikael,
>>
>> Thanks for finding this bug and sorry about the very late reply.
>>
>> I have reproduced the error but I haven't dug any deeper yet. I looks like
>> it could be some problem with the name index. I have to dig deeper..
>>
> Hi Mikael,
>
> Finally I got some time to work a little with pymagellan...
>
> I found the problem. MapSend doesn't like layers with a bounding box with zero
> area. A single POI did cause exactly that to happen and by making the bbox a
> little bigger than zero the problem was solved it seems.
>
> /Henrik
>
>

Re: [Pymag-devel] Points that cause MapSend to crash

[Henrik J. <he...@jo...>]

On Wednesday 04 November 2009 11:04:41 Henrik Johansson wrote:
> On Sunday 26 July 2009 20:08:11 Mikael Arvids wrote:
> > Great project! Definitely in my top ten list of open source projects
> > here on sourceforge!
> > 
> >  I've stumbled on some points that causes MapSend to crash.
> > 
> > The first node in the badPOI map is a shopping mall in Borlänge called
> > "Kupolen". I've generated the map with the skimap.xml rules in the
> > svn-tree. This makes the node a POI in the imi-file, and if this is
> > present it will crash MapSend.
> > 
> > The second node in the badPoint map is just a kindergarten in
> > Linköping. The map is generated with the osmmagrules.xml rules. This
> > will only add the node as a named point in the imi-file, but it seems
> > to be enough to put MapSend out of its misery. It seems though that
> > the problem only exist if I request that the names should be added
> > (with --name-tag-list=int_name,name). If I omit this MapSend will
> > display the map without any problem.
> 
> Hi Mikael,
> 
> Thanks for finding this bug and sorry about the very late reply.
> 
> I have reproduced the error but I haven't dug any deeper yet. I looks like
> it could be some problem with the name index. I have to dig deeper..
> 
Hi Mikael,

Finally I got some time to work a little with pymagellan...

I found the problem. MapSend doesn't like layers with a bounding box with zero 
area. A single POI did cause exactly that to happen and by making the bbox a 
little bigger than zero the problem was solved it seems.

/Henrik

[Pymag-devel] ValueError: Absolute coordinates are always positive

[bp <bp...@po...>]

Hello again

Now an otherone. This exception is thrown only with specific datasets.
Does negative coo. mean that one of the elements vertex falls outside 
the cell?
Does it help, if I send you the dataset?

INFO Closing map
Traceback (most recent call last):
  File "pymagellan\osmmagellan\osmmag.py", line 260, in <module>
    main()
  File "pymagellan\osmmagellan\osmmag.py", line 256, in main
    db=options.db
  File "pymagellan\osmmagellan\osmmag.py", line 97, in osmmag
    m.close()
  File "d:\Python26\lib\site-packages\magellan\Map.py", line 419, in close
    self._poiconfig.close()
  File "d:\Python26\lib\site-packages\magellan\Layer.py", line 294, in close
    layer.close()
  File "d:\Python26\lib\site-packages\magellan\Layer.py", line 498, in close
    celldata = cell.serialize()
  File "d:\Python26\lib\site-packages\magellan\Cell.py", line 172, in 
serialize
    data += self._serialize_cellelement(s)
  File "d:\Python26\lib\site-packages\magellan\Cell.py", line 85, in 
_serialize_
cellelement
    cellelementdata = cellelement.serialize(self, self.layer.bigendian)
  File "d:\Python26\lib\site-packages\magellan\CellElement.py", line 
369, in ser
ialize
    data, bbox = self._serialize_bbox(cell, bigendian)
  File "d:\Python26\lib\site-packages\magellan\CellElement.py", line 
216, in _se
rialize_bbox
    tprec, tdata = self._encodecoord(bbox.c1[1], bigendian)
  File "d:\Python26\lib\site-packages\magellan\CellElement.py", line 
246, in _en
codecoord
    raise ValueError("Absolute coordinates are always positive, c = %f"%c)
ValueError: Absolute coordinates are always positive, c = -529.000000



Regards
Peter

Re: [Pymag-devel] IndexError when too many POIs

[bp <bp...@po...>]

Hi Henrik,

Thanks. It works. POI index problem seems to be solved.

Regards
Peter

Henrik Johansson wrote:
> Hi Peter,
>
> Thanks for taking the time to find the bug. I have checked in your fix but 
> instead of putting it in Map.py I created an optimze method to the POIGroup 
> and made the first part of the layer optimization code include POI groups in 
> addition to the normal groups.
>
> Please try it  out.
>
> /Henrik
>
>   

Re: [Pymag-devel] Points that cause MapSend to crash

[Henrik J. <hen...@gm...>]

On Sunday 26 July 2009 20:08:11 Mikael Arvids wrote:
> Great project! Definitely in my top ten list of open source projects
> here on sourceforge!
> 
>  I've stumbled on some points that causes MapSend to crash.
> 
> The first node in the badPOI map is a shopping mall in Borlänge called
> "Kupolen". I've generated the map with the skimap.xml rules in the
> svn-tree. This makes the node a POI in the imi-file, and if this is
> present it will crash MapSend.
> 
> The second node in the badPoint map is just a kindergarten in
> Linköping. The map is generated with the osmmagrules.xml rules. This
> will only add the node as a named point in the imi-file, but it seems
> to be enough to put MapSend out of its misery. It seems though that
> the problem only exist if I request that the names should be added
> (with --name-tag-list=int_name,name). If I omit this MapSend will
> display the map without any problem.
> 

Hi Mikael,

Thanks for finding this bug and sorry about the very late reply.

I have reproduced the error but I haven't dug any deeper yet. I looks like it 
could be some problem with the name index. I have to dig deeper..

Best regards,

Henrik

Re: [Pymag-devel] invalid maps generated on windows

[Henrik J. <hen...@gm...>]

On Wednesday 28 October 2009 11:23:22 bp wrote:
> Hi,
> 
> I'm a newbee user of pymagellan. I found it great.
> However I'v installed it both on XP and Ubuntu. The XP version produced
> invalid map files.
> MapSend says: "Can not open map" though the linux version was ok.
> Datasources and rule files were the same in both cases.
> Than after a little search I found this topic:
> http://sourceforge.net/tracker/?func=detail&aid=2350976&group_id=239832&ati
> d=1111141 and applied the patch in Layer.py changing
> tmplay = tempfile.NamedTemporaryFile('w')
> to
> tmplay = tempfile.NamedTemporaryFile('wb')
> after applying the modification, the maps created on windows were ok.
> I'v applied the modification on linux also.
> MapSend can open both the windows and the linux version of maps.
> So I think this patch sould be in svn.
> 
> Regards
> Peter

Hi Peter,

I have now fixed this problem in trunk. 

Thanks for reporting it.

/Henrik

Re: [Pymag-devel] IndexError when too many POIs

[Henrik J. <hen...@gm...>]

Hi Peter,

Thanks for taking the time to find the bug. I have checked in your fix but 
instead of putting it in Map.py I created an optimze method to the POIGroup 
and made the first part of the layer optimization code include POI groups in 
addition to the normal groups.

Please try it  out.

/Henrik

On Monday 02 November 2009 10:29:41 bp wrote:
> Ok, after some day of debugging and trying to figure out how the code
> works, I found that layer.optimize is called
> for optimizing layers before closing them. Here cellelements could be
> moved to new cells.
> This method returns a dictionary  with changes, however  this result was
> not used after optimizing POI layer.
> Than when closing POI layer, POI features were referenced  to the old
> cell numbers, causing index error.
> 
> To update POI features, here are the differencies in Map.py.
> Henrik could you revise it?
> 
> @@ -359,7 +359,7 @@
> 
>          ## Optimize layers in groups
>          logging.info('Optimizing cell structure of normal layers')
> -        remaininglayers = list(self.layers + self._poiconfig.layers)
> +        remaininglayers = list(self.layers)
>          if self.groups != None:
>              for group in self.groups:
>                  group.optimizeLayers()
> @@ -380,8 +380,16 @@
> 
> 
>          ## Optimize remaining layers
> +
> +        for layer in self._poiconfig.layers:
> +            remapdict = layer.optimize()
> +            for feature in self._poigroup._features:
> +                if len(remapdict) > 0:
> +                    feature.cellelementrefs = tuple([remapdict[ceref]
> for ceref in feature.cellelementrefs])
> +
>          if self.maptype == MapTypeStreetRoute:
>              logging.info('Optimizing cell structure of routing layers')
> +
>          for layer in remaininglayers:
>              layer.optimize()
> 
> 
> Regards
> Peter
> 
> > Hi,
> >
> > I have a datasource with a huge number of POIs
> > When I compile the map I get the following error:
> > IndexError: num_in_cell (1136) is outside the # of cellelements (0) in
> > cell 1, layer poi
> > If I limit the number of POIs by filtering out some types in rule file,
> > than map is compiled ok.
> >
> > Is this a limitation in the map format or a bug?
> >
> > ...
> > DEBUG Optimizing layer poi
> > DEBUG Max cell elements exceeded for layer poi. cellnum=1, # of cell
> > elements=18994
> > DEBUG Max cell elements exceeded for layer poi. cellnum=1, # of cell
> > elements=18994
> > DEBUG nlevels=2 for layer poi
> > INFO Closing map
> > Traceback (most recent call last):
> >   File "pymagellan\osmmagellan\osmmag.py", line 260, in <module>
> >     main()
> >   File "pymagellan\osmmagellan\osmmag.py", line 256, in main
> >     db=options.db
> >   File "pymagellan\osmmagellan\osmmag.py", line 97, in osmmag
> >     m.close()
> >   File "d:\Python26\lib\site-packages\magellan\Map.py", line 413, in
> > close self._poigroup.close()
> >   File "d:\Python26\lib\site-packages\magellan\POI.py", line 201, in
> > close for ref, e in zip(f.cellelementrefs, f.getCellElements(self.map)):
> > File "d:\Python26\lib\site-packages\magellan\POI.py", line 518, in
> > getCellElements
> >     return [layer.getCellElement(ceref) for ceref in
> > self.cellelementrefs] File
> > "d:\Python26\lib\site-packages\magellan\Layer.py", line 712, in
> > getCellElement
> >     raise IndexError,"num_in_cell (%d) is outside the # of cellelements
> > (%d) in cell %d, layer %s"%(num_in_cell,len(cell),cellnum, self.name)
> > IndexError: num_in_cell (1136) is outside the # of cellelements (0) in
> > cell 1, layer poi
> 

Re: [Pymag-devel] IndexError when too many POIs

[bp <bp...@po...>]

Ok, after some day of debugging and trying to figure out how the code 
works, I found that layer.optimize is called
for optimizing layers before closing them. Here cellelements could be 
moved to new cells.
This method returns a dictionary  with changes, however  this result was 
not used after optimizing POI layer.
Than when closing POI layer, POI features were referenced  to the old 
cell numbers, causing index error.

To update POI features, here are the differencies in Map.py.
Henrik could you revise it?

@@ -359,7 +359,7 @@
        
         ## Optimize layers in groups
         logging.info('Optimizing cell structure of normal layers')
-        remaininglayers = list(self.layers + self._poiconfig.layers)
+        remaininglayers = list(self.layers)
         if self.groups != None:
             for group in self.groups:
                 group.optimizeLayers()
@@ -380,8 +380,16 @@
 
 
         ## Optimize remaining layers
+
+        for layer in self._poiconfig.layers:
+            remapdict = layer.optimize()
+            for feature in self._poigroup._features:
+                if len(remapdict) > 0:
+                    feature.cellelementrefs = tuple([remapdict[ceref] 
for ceref in feature.cellelementrefs])
+
         if self.maptype == MapTypeStreetRoute:
             logging.info('Optimizing cell structure of routing layers')
+
         for layer in remaininglayers:
             layer.optimize()


Regards
Peter


> Hi,
>
> I have a datasource with a huge number of POIs
> When I compile the map I get the following error:
> IndexError: num_in_cell (1136) is outside the # of cellelements (0) in 
> cell 1, layer poi
> If I limit the number of POIs by filtering out some types in rule file, 
> than map is compiled ok.
>
> Is this a limitation in the map format or a bug?
>
> ...
> DEBUG Optimizing layer poi
> DEBUG Max cell elements exceeded for layer poi. cellnum=1, # of cell 
> elements=18994
> DEBUG Max cell elements exceeded for layer poi. cellnum=1, # of cell 
> elements=18994
> DEBUG nlevels=2 for layer poi
> INFO Closing map
> Traceback (most recent call last):
>   File "pymagellan\osmmagellan\osmmag.py", line 260, in <module>
>     main()
>   File "pymagellan\osmmagellan\osmmag.py", line 256, in main
>     db=options.db
>   File "pymagellan\osmmagellan\osmmag.py", line 97, in osmmag
>     m.close()
>   File "d:\Python26\lib\site-packages\magellan\Map.py", line 413, in close
>     self._poigroup.close()
>   File "d:\Python26\lib\site-packages\magellan\POI.py", line 201, in close
>     for ref, e in zip(f.cellelementrefs, f.getCellElements(self.map)):
>   File "d:\Python26\lib\site-packages\magellan\POI.py", line 518, in 
> getCellElements
>     return [layer.getCellElement(ceref) for ceref in self.cellelementrefs]
>   File "d:\Python26\lib\site-packages\magellan\Layer.py", line 712, in 
> getCellElement
>     raise IndexError,"num_in_cell (%d) is outside the # of cellelements 
> (%d) in cell %d, layer %s"%(num_in_cell,len(cell),cellnum, self.name)
> IndexError: num_in_cell (1136) is outside the # of cellelements (0) in 
> cell 1, layer poi
>   

[Pymag-devel] IndexError when too many POIs

[bp <bp...@po...>]

Hi,

I have a datasource with a huge number of POIs
When I compile the map I get the following error:
IndexError: num_in_cell (1136) is outside the # of cellelements (0) in 
cell 1, layer poi
If I limit the number of POIs by filtering out some types in rule file, 
than map is compiled ok.

Is this a limitation in the map format or a bug?

...
DEBUG Optimizing layer poi
DEBUG Max cell elements exceeded for layer poi. cellnum=1, # of cell 
elements=18994
DEBUG Max cell elements exceeded for layer poi. cellnum=1, # of cell 
elements=18994
DEBUG nlevels=2 for layer poi
INFO Closing map
Traceback (most recent call last):
  File "pymagellan\osmmagellan\osmmag.py", line 260, in <module>
    main()
  File "pymagellan\osmmagellan\osmmag.py", line 256, in main
    db=options.db
  File "pymagellan\osmmagellan\osmmag.py", line 97, in osmmag
    m.close()
  File "d:\Python26\lib\site-packages\magellan\Map.py", line 413, in close
    self._poigroup.close()
  File "d:\Python26\lib\site-packages\magellan\POI.py", line 201, in close
    for ref, e in zip(f.cellelementrefs, f.getCellElements(self.map)):
  File "d:\Python26\lib\site-packages\magellan\POI.py", line 518, in 
getCellElements
    return [layer.getCellElement(ceref) for ceref in self.cellelementrefs]
  File "d:\Python26\lib\site-packages\magellan\Layer.py", line 712, in 
getCellElement
    raise IndexError,"num_in_cell (%d) is outside the # of cellelements 
(%d) in cell %d, layer %s"%(num_in_cell,len(cell),cellnum, self.name)
IndexError: num_in_cell (1136) is outside the # of cellelements (0) in 
cell 1, layer poi



Thanks in advance
Regards
Peter

[Pymag-devel] invalid maps generated on windows

[bp <bp...@po...>]

Hi,

I'm a newbee user of pymagellan. I found it great.
However I'v installed it both on XP and Ubuntu. The XP version produced 
invalid map files.
MapSend says: "Can not open map" though the linux version was ok.
Datasources and rule files were the same in both cases.
Than after a little search I found this topic:
http://sourceforge.net/tracker/?func=detail&aid=2350976&group_id=239832&atid=1111141
and applied the patch in Layer.py changing
tmplay = tempfile.NamedTemporaryFile('w')
to
tmplay = tempfile.NamedTemporaryFile('wb')
after applying the modification, the maps created on windows were ok.
I'v applied the modification on linux also.
MapSend can open both the windows and the linux version of maps.
So I think this patch sould be in svn.

Regards
Peter

[Pymag-devel] Points that cause MapSend to crash

[Mikael A. <mik...@gm...>]

Great project! Definitely in my top ten list of open source projects
here on sourceforge!

 I've stumbled on some points that causes MapSend to crash.

The first node in the badPOI map is a shopping mall in Borlänge called
"Kupolen". I've generated the map with the skimap.xml rules in the
svn-tree. This makes the node a POI in the imi-file, and if this is
present it will crash MapSend.

The second node in the badPoint map is just a kindergarten in
Linköping. The map is generated with the osmmagrules.xml rules. This
will only add the node as a named point in the imi-file, but it seems
to be enough to put MapSend out of its misery. It seems though that
the problem only exist if I request that the names should be added
(with --name-tag-list=int_name,name). If I omit this MapSend will
display the map without any problem.

I hope that this helps,
Mikael Arvids

Re: [Pymag-devel] geofabrick error

[henrik j. <hen...@gm...>]

Hi,

Thanks for reporting the problem and sorry about the slow reply. The
problem is because of a restriction I added sometime ago for a certain
polygon type.
Unfortunately I don't remember why I added the restriction.

In the latest trunk I have removed it and instead of exiting it prints
a warning. Hopefully it will work but it could be that it won't work
on all devices.

Best regards,

Henrik

2009/6/20 ouɐɯnH <fre...@gm...>:
> hello i test to a map generete from geofabrick .osm
> http://download.geofabrik.de/osm/south-america/colombia.osm.bz2 but i
> have this error:
>
> debian:# osmmag.py -v -n "Colombia 20090620" -o colombia_20090620.imi
> -r osmmagellan/data/osmmagrules.xml colombia.osm
> INFO Loading rules: osmmagellan/data/osmmagrules.xml
> INFO Reading osm data and creating map
> INFO Optimizing cell structure
> Traceback (most recent call last):
> File "/usr/bin/osmmag.py", line 180, in <module>
> main()
> File "/usr/bin/osmmag.py", line 176, in main
> bigendian=options.bigendian
> File "/usr/bin/osmmag.py", line 61, in osmmag
> m.close()
> File "/usr/lib/python2.5/site-packages/magellan/Map.py", line 358, in close
> self._laycfg.close()
> File "/usr/lib/python2.5/site-packages/magellan/Layer.py", line 278, in close
> layer.close()
> File "/usr/lib/python2.5/site-packages/magellan/Layer.py", line 466, in close
> celldata = cell.serialize()
> File "/usr/lib/python2.5/site-packages/magellan/Cell.py", line 168, in
> serialize
> data += self._serialize_cellelement(s)
> File "/usr/lib/python2.5/site-packages/magellan/Cell.py", line 85, in
> _serialize_cellelement
> cellelementdata = cellelement.serialize(self, self.layer.bigendian)
> File "/usr/lib/python2.5/site-packages/magellan/CellElement.py", line
> 465, in serialize
> cdata, subpolytype, nvertices = self.encode_polygon(bbox, p, bigendian)
> File "/usr/lib/python2.5/site-packages/magellan/CellElement.py", line
> 624, in encode_polygon
> raise Exception("Cannot encode polygon")
> Exception: Cannot encode polygon
>
> but with the export from
> http://www.openstreetmap.org/export?lat=4.6924&lon=-74.1438&zoom=14&layers=B000FTFT
> the script works well
>
> the problem is ; i need all the country.
>
> thanks for the help
> humano
> --
> Lee un artículo sobre nosotros en el tiempo:
> http://www.eltiempo.com/colombia/bogota/empresas-de-tecnologia-basadas-en-software-libre-se-expanden-en-bogota_4708463-1
>
> ------------------------------------------------------------------------------
> Are you an open source citizen? Join us for the Open Source Bridge conference!
> Portland, OR, June 17-19. Two days of sessions, one day of unconference: $250.
> Need another reason to go? 24-hour hacker lounge. Register today!
> http://ad.doubleclick.net/clk;215844324;13503038;v?http://opensourcebridge.org
> _______________________________________________
> Pymag-devel mailing list
> Pym...@li...
> https://lists.sourceforge.net/lists/listinfo/pymag-devel
>

[Pymag-devel] geofabrick error

[ouɐɯnH <fre...@gm...>]

hello i test to a map generete from geofabrick .osm
http://download.geofabrik.de/osm/south-america/colombia.osm.bz2 but i
have this error:

debian:# osmmag.py -v -n "Colombia 20090620" -o colombia_20090620.imi
-r osmmagellan/data/osmmagrules.xml colombia.osm
INFO Loading rules: osmmagellan/data/osmmagrules.xml
INFO Reading osm data and creating map
INFO Optimizing cell structure
Traceback (most recent call last):
File "/usr/bin/osmmag.py", line 180, in <module>
main()
File "/usr/bin/osmmag.py", line 176, in main
bigendian=options.bigendian
File "/usr/bin/osmmag.py", line 61, in osmmag
m.close()
File "/usr/lib/python2.5/site-packages/magellan/Map.py", line 358, in close
self._laycfg.close()
File "/usr/lib/python2.5/site-packages/magellan/Layer.py", line 278, in close
layer.close()
File "/usr/lib/python2.5/site-packages/magellan/Layer.py", line 466, in close
celldata = cell.serialize()
File "/usr/lib/python2.5/site-packages/magellan/Cell.py", line 168, in
serialize
data += self._serialize_cellelement(s)
File "/usr/lib/python2.5/site-packages/magellan/Cell.py", line 85, in
_serialize_cellelement
cellelementdata = cellelement.serialize(self, self.layer.bigendian)
File "/usr/lib/python2.5/site-packages/magellan/CellElement.py", line
465, in serialize
cdata, subpolytype, nvertices = self.encode_polygon(bbox, p, bigendian)
File "/usr/lib/python2.5/site-packages/magellan/CellElement.py", line
624, in encode_polygon
raise Exception("Cannot encode polygon")
Exception: Cannot encode polygon

but with the export from
http://www.openstreetmap.org/export?lat=4.6924&lon=-74.1438&zoom=14&layers=B000FTFT
the script works well

the problem is ; i need all the country.

thanks for the help
humano
-- 
Lee un artículo sobre nosotros en el tiempo:
http://www.eltiempo.com/colombia/bogota/empresas-de-tecnologia-basadas-en-software-libre-se-expanden-en-bogota_4708463-1

Re: [Pymag-devel] Bounding box

[Josh <jo...@fe...>]

That looks a bit better...
The map built without error this time, but crashes mapsend so I can  
only guess it won't work too well on the GPSr. I'll have a muck around  
today and try to work out which layer is causing the problem.


Here's the output from this attempt, anything look bad? I assume the  
nlevels=2 means that it's worked out another level is needed?

INFO Create hydro polygons from coastline
INFO Optimizing cell structure of normal layers
DEBUG Optimizing layer 00_Freeways
DEBUG nlevels=0 for layer 00_Freeways
DEBUG Optimizing layer 00_Major_Highways
DEBUG nlevels=2 for layer 00_Major_Highways
DEBUG Optimizing layer 00_Minor_Highways
DEBUG Max cell elements exceeded for layer 00_Minor_Highways.  
cellnum=1, # of cell elements=3103
DEBUG Max cell elements exceeded for layer 00_Minor_Highways.  
cellnum=1, # of cell elements=3103
DEBUG nlevels=2 for layer 00_Minor_Highways
DEBUG Optimizing layer 00_Major_Roads
DEBUG Max cell elements exceeded for layer 00_Major_Roads. cellnum=1,  
# of cell elements=5700
DEBUG Max cell elements exceeded for layer 00_Major_Roads. cellnum=1,  
# of cell elements=5700
DEBUG nlevels=2 for layer 00_Major_Roads
DEBUG Optimizing layer 00_Streets
DEBUG Max cell elements exceeded for layer 00_Streets. cellnum=1, # of  
cell elements=81942
DEBUG Max cell elements exceeded for layer 00_Streets. cellnum=1, # of  
cell elements=81942
DEBUG nlevels=2 for layer 00_Streets
DEBUG Optimizing layer 00_Trails_4WD
DEBUG nlevels=0 for layer 00_Trails_4WD
DEBUG Optimizing layer 00_Railroads
DEBUG nlevels=0 for layer 00_Railroads
DEBUG Optimizing layer 00_Hydro_Polygons
DEBUG nlevels=0 for layer 00_Hydro_Polygons
DEBUG Optimizing layer 00_Hydro_Lines
DEBUG nlevels=2 for layer 00_Hydro_Lines
DEBUG Optimizing layer 00_Parks
DEBUG nlevels=0 for layer 00_Parks
DEBUG Optimizing layer 00_Trails
DEBUG Max cell elements exceeded for layer 00_Trails. cellnum=1, # of  
cell elements=5417
DEBUG Max cell elements exceeded for layer 00_Trails. cellnum=1, # of  
cell elements=5417
DEBUG nlevels=2 for layer 00_Trails
DEBUG Optimizing layer 00_Buildings
DEBUG nlevels=0 for layer 00_Buildings
DEBUG Optimizing layer 00_Farm
DEBUG nlevels=0 for layer 00_Farm
DEBUG Optimizing layer 00_Points
DEBUG nlevels=0 for layer 00_Points
DEBUG Optimizing layer poi
DEBUG nlevels=0 for layer poi
INFO Closing map


Cheers,
Josh


Quoting henrik johansson <hen...@gm...>:

> Thanks!
>
> I think I found the bug, try now.
>
> /Henrik

Re: [Pymag-devel] Bounding box

[henrik j. <hen...@gm...>]

Thanks!

I think I found the bug, try now.

/Henrik

2009/6/4 Josh <jo...@fe...>:
> Hi Henrik,
>
> Here's today's output.
>
> INFO Optimizing cell structure of normal layers
> DEBUG Optimizing layer 00_Freeways
> DEBUG nlevels=0 for layer 00_Freeways
> DEBUG Optimizing layer 00_Major_Highways
> DEBUG nlevels=1 for layer 00_Major_Highways
> DEBUG Optimizing layer 00_Minor_Highways
> DEBUG Max cell elements exceeded for layer 00_Minor_Highways. cellnum=1, #
> of cell elements=3103
> DEBUG nlevels=1 for layer 00_Minor_Highways
> DEBUG Optimizing layer 00_Major_Roads
> DEBUG Max cell elements exceeded for layer 00_Major_Roads. cellnum=1, # of
> cell elements=5700
> DEBUG nlevels=1 for layer 00_Major_Roads
> DEBUG Optimizing layer 00_Streets
> DEBUG Max cell elements exceeded for layer 00_Streets. cellnum=1, # of cell
> elements=81942
> DEBUG nlevels=1 for layer 00_Streets
> Traceback (most recent call last):
>  File "/usr/bin/osmmag.py", line 260, in <module>
>    main()
>  File "/usr/bin/osmmag.py", line 256, in main
>    db=options.db
>  File "/usr/bin/osmmag.py", line 97, in osmmag
>    m.close()
>  File "pymagellan/trunk/magellan/Map.py", line 365, in close
>    group.optimizeLayers()
>  File "pymagellan/trunk/magellan/SearchGroup.py", line 116, in
> optimizeLayers
>    cellrefremap[self.map.getLayerIndex(layer)] = layer.optimize()
>  File "pymagellan/trunk/magellan/Layer.py", line 473, in optimize
>    newcellrefs = self.addCellElement(ce)
>  File "pymagellan/trunk/magellan/Layer.py", line 745, in addCellElement
>    assert self.nlevels == 0 or nincell < 2**16
> AssertionError
>
>
> Cheers,
> Josh
>
>
> Quoting henrik johansson <hen...@gm...>:
>
>> Hi Josh,
>>
>> I found that this assertion was not correct. I added a new option that
>> gives more verbose output.
>>
>> Can you please try again with the --debug option?
>>
>> Regards,
>>
>> 2009/6/3 Josh <jo...@fe...>:
>>>
>>> Hi Henrik,
>>>
>>> It looks like your guess was right - here's the output from svn trunk
>>> 28 on the same map build as yesterday:
>>>
>>> Traceback (most recent call last):
>>>   File "/usr/bin/osmmag.py", line 244, in <module>
>>>     main()
>>>   File "/usr/bin/osmmag.py", line 240, in main
>>>     fromdb=options.fromdb
>>>   File "/usr/bin/osmmag.py", line 84, in osmmag
>>>     data.load()
>>>   File "osmmagellan/osm.py", line 199, in load
>>>     parser.parse(self.filename)
>>>   File "/usr/lib/python2.5/xml/sax/expatreader.py", line 107, in parse
>>>     xmlreader.IncrementalParser.parse(self, source)
>>>   File "/usr/lib/python2.5/xml/sax/xmlreader.py", line 123, in parse
>>>     self.feed(buffer)
>>>   File "/usr/lib/python2.5/xml/sax/expatreader.py", line 207, in feed
>>>     self._parser.Parse(data, isFinal)
>>>   File "/usr/lib/python2.5/xml/sax/expatreader.py", line 304, in
>>> end_element
>>>     self._cont_handler.endElement(name)
>>>   File "osmmagellan/osm.py", line 247, in endElement
>>>     self.add_element(statement, self.tags, coords)
>>>   File "osmmagellan/osm.py", line 105, in add_element
>>>     cellelementrefs = layer.addCellElement(cellelement)
>>>   File "magellan/Layer.py", line 745, in addCellElement
>>>     assert nincell < 2**16
>>> AssertionError
>>>
>>>
>>> Cheers,
>>> Josh
>>>
>>>
>>>
>>> Quoting henrik johansson <hen...@gm...>:
>>>
>>>> Hi Josh,
>>>>
>>>> Thanks for testing.
>>>>
>>>> I think I know what the problem is. There's a limit of 64k cell
>>>> elements in a cell. So it's clearly a pymagellan fault since it could
>>>> just increase the number of cell levels if it reaches this limit. The
>>>> threshold is set to max 2000 elements per cell but I think this might
>>>> not be respected for some reason.
>>>>
>>>> I added some more asserts that should catch this problem.
>>>>
>>>> /Henrik
>>>
>>>
>>>
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> OpenSolaris 2009.06 is a cutting edge operating system for enterprises
>>> looking to deploy the next generation of Solaris that includes the latest
>>> innovations from Sun and the OpenSource community. Download a copy and
>>> enjoy capabilities such as Networking, Storage and Virtualization.
>>> Go to: http://p.sf.net/sfu/opensolaris-get
>>> _______________________________________________
>>> Pymag-devel mailing list
>>> Pym...@li...
>>> https://lists.sourceforge.net/lists/listinfo/pymag-devel
>>>
>>
>>
>> ------------------------------------------------------------------------------
>> OpenSolaris 2009.06 is a cutting edge operating system for enterprises
>> looking to deploy the next generation of Solaris that includes the latest
>> innovations from Sun and the OpenSource community. Download a copy and
>> enjoy capabilities such as Networking, Storage and Virtualization.
>> Go to: http://p.sf.net/sfu/opensolaris-get
>> _______________________________________________
>> Pymag-devel mailing list
>> Pym...@li...
>> https://lists.sourceforge.net/lists/listinfo/pymag-devel
>>
>
>
>
>

Re: [Pymag-devel] Bounding box

[Josh <jo...@fe...>]

Hi Henrik,

Here's today's output.

INFO Optimizing cell structure of normal layers
DEBUG Optimizing layer 00_Freeways
DEBUG nlevels=0 for layer 00_Freeways
DEBUG Optimizing layer 00_Major_Highways
DEBUG nlevels=1 for layer 00_Major_Highways
DEBUG Optimizing layer 00_Minor_Highways
DEBUG Max cell elements exceeded for layer 00_Minor_Highways.  
cellnum=1, # of cell elements=3103
DEBUG nlevels=1 for layer 00_Minor_Highways
DEBUG Optimizing layer 00_Major_Roads
DEBUG Max cell elements exceeded for layer 00_Major_Roads. cellnum=1,  
# of cell elements=5700
DEBUG nlevels=1 for layer 00_Major_Roads
DEBUG Optimizing layer 00_Streets
DEBUG Max cell elements exceeded for layer 00_Streets. cellnum=1, # of  
cell elements=81942
DEBUG nlevels=1 for layer 00_Streets
Traceback (most recent call last):
   File "/usr/bin/osmmag.py", line 260, in <module>
     main()
   File "/usr/bin/osmmag.py", line 256, in main
     db=options.db
   File "/usr/bin/osmmag.py", line 97, in osmmag
     m.close()
   File "pymagellan/trunk/magellan/Map.py", line 365, in close
     group.optimizeLayers()
   File "pymagellan/trunk/magellan/SearchGroup.py", line 116, in optimizeLayers
     cellrefremap[self.map.getLayerIndex(layer)] = layer.optimize()
   File "pymagellan/trunk/magellan/Layer.py", line 473, in optimize
     newcellrefs = self.addCellElement(ce)
   File "pymagellan/trunk/magellan/Layer.py", line 745, in addCellElement
     assert self.nlevels == 0 or nincell < 2**16
AssertionError


Cheers,
Josh


Quoting henrik johansson <hen...@gm...>:

> Hi Josh,
>
> I found that this assertion was not correct. I added a new option that
> gives more verbose output.
>
> Can you please try again with the --debug option?
>
> Regards,
>
> 2009/6/3 Josh <jo...@fe...>:
>> Hi Henrik,
>>
>> It looks like your guess was right - here's the output from svn trunk
>> 28 on the same map build as yesterday:
>>
>> Traceback (most recent call last):
>>   File "/usr/bin/osmmag.py", line 244, in <module>
>>     main()
>>   File "/usr/bin/osmmag.py", line 240, in main
>>     fromdb=options.fromdb
>>   File "/usr/bin/osmmag.py", line 84, in osmmag
>>     data.load()
>>   File "osmmagellan/osm.py", line 199, in load
>>     parser.parse(self.filename)
>>   File "/usr/lib/python2.5/xml/sax/expatreader.py", line 107, in parse
>>     xmlreader.IncrementalParser.parse(self, source)
>>   File "/usr/lib/python2.5/xml/sax/xmlreader.py", line 123, in parse
>>     self.feed(buffer)
>>   File "/usr/lib/python2.5/xml/sax/expatreader.py", line 207, in feed
>>     self._parser.Parse(data, isFinal)
>>   File "/usr/lib/python2.5/xml/sax/expatreader.py", line 304, in end_element
>>     self._cont_handler.endElement(name)
>>   File "osmmagellan/osm.py", line 247, in endElement
>>     self.add_element(statement, self.tags, coords)
>>   File "osmmagellan/osm.py", line 105, in add_element
>>     cellelementrefs = layer.addCellElement(cellelement)
>>   File "magellan/Layer.py", line 745, in addCellElement
>>     assert nincell < 2**16
>> AssertionError
>>
>>
>> Cheers,
>> Josh
>>
>>
>>
>> Quoting henrik johansson <hen...@gm...>:
>>
>>> Hi Josh,
>>>
>>> Thanks for testing.
>>>
>>> I think I know what the problem is. There's a limit of 64k cell
>>> elements in a cell. So it's clearly a pymagellan fault since it could
>>> just increase the number of cell levels if it reaches this limit. The
>>> threshold is set to max 2000 elements per cell but I think this might
>>> not be respected for some reason.
>>>
>>> I added some more asserts that should catch this problem.
>>>
>>> /Henrik
>>
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> OpenSolaris 2009.06 is a cutting edge operating system for enterprises
>> looking to deploy the next generation of Solaris that includes the latest
>> innovations from Sun and the OpenSource community. Download a copy and
>> enjoy capabilities such as Networking, Storage and Virtualization.
>> Go to: http://p.sf.net/sfu/opensolaris-get
>> _______________________________________________
>> Pymag-devel mailing list
>> Pym...@li...
>> https://lists.sourceforge.net/lists/listinfo/pymag-devel
>>
>
> ------------------------------------------------------------------------------
> OpenSolaris 2009.06 is a cutting edge operating system for enterprises
> looking to deploy the next generation of Solaris that includes the latest
> innovations from Sun and the OpenSource community. Download a copy and
> enjoy capabilities such as Networking, Storage and Virtualization.
> Go to: http://p.sf.net/sfu/opensolaris-get
> _______________________________________________
> Pymag-devel mailing list
> Pym...@li...
> https://lists.sourceforge.net/lists/listinfo/pymag-devel
>

Re: [Pymag-devel] Bounding box

[henrik j. <hen...@gm...>]

Hi Josh,

I found that this assertion was not correct. I added a new option that
gives more verbose output.

Can you please try again with the --debug option?

Regards,

2009/6/3 Josh <jo...@fe...>:
> Hi Henrik,
>
> It looks like your guess was right - here's the output from svn trunk
> 28 on the same map build as yesterday:
>
> Traceback (most recent call last):
>   File "/usr/bin/osmmag.py", line 244, in <module>
>     main()
>   File "/usr/bin/osmmag.py", line 240, in main
>     fromdb=options.fromdb
>   File "/usr/bin/osmmag.py", line 84, in osmmag
>     data.load()
>   File "osmmagellan/osm.py", line 199, in load
>     parser.parse(self.filename)
>   File "/usr/lib/python2.5/xml/sax/expatreader.py", line 107, in parse
>     xmlreader.IncrementalParser.parse(self, source)
>   File "/usr/lib/python2.5/xml/sax/xmlreader.py", line 123, in parse
>     self.feed(buffer)
>   File "/usr/lib/python2.5/xml/sax/expatreader.py", line 207, in feed
>     self._parser.Parse(data, isFinal)
>   File "/usr/lib/python2.5/xml/sax/expatreader.py", line 304, in end_element
>     self._cont_handler.endElement(name)
>   File "osmmagellan/osm.py", line 247, in endElement
>     self.add_element(statement, self.tags, coords)
>   File "osmmagellan/osm.py", line 105, in add_element
>     cellelementrefs = layer.addCellElement(cellelement)
>   File "magellan/Layer.py", line 745, in addCellElement
>     assert nincell < 2**16
> AssertionError
>
>
> Cheers,
> Josh
>
>
>
> Quoting henrik johansson <hen...@gm...>:
>
>> Hi Josh,
>>
>> Thanks for testing.
>>
>> I think I know what the problem is. There's a limit of 64k cell
>> elements in a cell. So it's clearly a pymagellan fault since it could
>> just increase the number of cell levels if it reaches this limit. The
>> threshold is set to max 2000 elements per cell but I think this might
>> not be respected for some reason.
>>
>> I added some more asserts that should catch this problem.
>>
>> /Henrik
>
>
>
>
>
> ------------------------------------------------------------------------------
> OpenSolaris 2009.06 is a cutting edge operating system for enterprises
> looking to deploy the next generation of Solaris that includes the latest
> innovations from Sun and the OpenSource community. Download a copy and
> enjoy capabilities such as Networking, Storage and Virtualization.
> Go to: http://p.sf.net/sfu/opensolaris-get
> _______________________________________________
> Pymag-devel mailing list
> Pym...@li...
> https://lists.sourceforge.net/lists/listinfo/pymag-devel
>

Re: [Pymag-devel] Bounding box

[Josh <jo...@fe...>]

Hi Henrik,

It looks like your guess was right - here's the output from svn trunk  
28 on the same map build as yesterday:

Traceback (most recent call last):
   File "/usr/bin/osmmag.py", line 244, in <module>
     main()
   File "/usr/bin/osmmag.py", line 240, in main
     fromdb=options.fromdb
   File "/usr/bin/osmmag.py", line 84, in osmmag
     data.load()
   File "osmmagellan/osm.py", line 199, in load
     parser.parse(self.filename)
   File "/usr/lib/python2.5/xml/sax/expatreader.py", line 107, in parse
     xmlreader.IncrementalParser.parse(self, source)
   File "/usr/lib/python2.5/xml/sax/xmlreader.py", line 123, in parse
     self.feed(buffer)
   File "/usr/lib/python2.5/xml/sax/expatreader.py", line 207, in feed
     self._parser.Parse(data, isFinal)
   File "/usr/lib/python2.5/xml/sax/expatreader.py", line 304, in end_element
     self._cont_handler.endElement(name)
   File "osmmagellan/osm.py", line 247, in endElement
     self.add_element(statement, self.tags, coords)
   File "osmmagellan/osm.py", line 105, in add_element
     cellelementrefs = layer.addCellElement(cellelement)
   File "magellan/Layer.py", line 745, in addCellElement
     assert nincell < 2**16
AssertionError


Cheers,
Josh



Quoting henrik johansson <hen...@gm...>:

> Hi Josh,
>
> Thanks for testing.
>
> I think I know what the problem is. There's a limit of 64k cell
> elements in a cell. So it's clearly a pymagellan fault since it could
> just increase the number of cell levels if it reaches this limit. The
> threshold is set to max 2000 elements per cell but I think this might
> not be respected for some reason.
>
> I added some more asserts that should catch this problem.
>
> /Henrik

Re: [Pymag-devel] Bounding box

[henrik j. <hen...@gm...>]

Hi Josh,

Thanks for testing.

I think I know what the problem is. There's a limit of 64k cell
elements in a cell. So it's clearly a pymagellan fault since it could
just increase the number of cell levels if it reaches this limit. The
threshold is set to max 2000 elements per cell but I think this might
not be respected for some reason.

I added some more asserts that should catch this problem.

/Henrik

2009/6/2 Josh <jo...@fe...>:
- Show quoted text -

2009/6/2 Josh <jo...@fe...>:
> Quoting henrik johansson <hen...@gm...>:
>
>> Hi,
>>
>> I have now added bounding box support for the database reader.
>>
>> Think it's high time to make another release. It would be nice if you
>> could test the latest trunk to see if it still works for you.
>>
>> /Henrik
>
> Hi Henrik,
>
> I've done a few tests with the current svn version today, and it looks good.
> I haven't tried the maps on a GPSr yet but at least they don't crash
> mapsend ;)
> Most polygons are rendering now. I haven't tested the bounding box or
> pgsql yet.
>
>
> On another note, I've been getting an error for a while on one
> particular map I've been building:
>
> magellan/DBUtil.py:190: DeprecationWarning: 'H' format requires 0 <=
> number <= 65535
>   return apply(struct.pack, (structendian+types,) + data)
> magellan/Layer.py:646: DeprecationWarning: 'H' format requires 0 <=
> number <= 65535
>   return struct.pack(prefix+types, *data)
>
> I *think* this is because there are more than 64k
> residential/unclassified  roads in the osm extract going in to a
> single layer.
> Is this a limitation of the magellan map format, or in pymag?
> In an attempt to overcome this in the past I've split the roads into
> multiple layers, and that builds successfuly.
> The map can be opened in Mapsend and Vantagepoint, but locks up GPSRs
> (tested on half a dozen explorists) as soon as you select the detail
> map.
> I don't know if this is related or a different problem.
> Any hints on diagnosing the problem other than trial-and-error
> exclusion of objects?
>
> Cheers,
> Josh
>
>
>
>
>
>
> ------------------------------------------------------------------------------
> OpenSolaris 2009.06 is a cutting edge operating system for enterprises
> looking to deploy the next generation of Solaris that includes the latest
> innovations from Sun and the OpenSource community. Download a copy and
> enjoy capabilities such as Networking, Storage and Virtualization.
> Go to: http://p.sf.net/sfu/opensolaris-get
> _______________________________________________
> Pymag-devel mailing list
> Pym...@li...
> https://lists.sourceforge.net/lists/listinfo/pymag-devel
>