[geos-devel] question on prospectives of python bindings to opensource geospatial libraries

Sean Gillies sgillies at frii.com
Mon Mar 12 17:03:16 EDT 2007

N.J. Hardebol wrote:
> Hello,
> In the email of Feb. 28 I posed some questions on python GEOS
> implementations to which Sean Gillies replied saying also that Python
> wrappers for GEOS are not actively maintained.
> http://geos.refractions.net/pipermail/geos-devel/2007-February/002848.html
> http://geos.refractions.net/pipermail/geos-devel/2007-March/002851.html
> We ( = group of structural geologist - working on some academic coding -
> see below) were expecting that with the advances in spatial open source
> community, spatial statistics on scripting level (i.e. using python) would
> be a feasible development path. Although also ESRI has made certain steps
> allowing its topology functions to be called from within python we rather
> prefer to rely on open source libraries.
> For this we indeed require that GEOS and OGR developments would get better
> supported to higher scripting languages. Although not well documented use
> of swig python bindings appeared feasible to us. Further we noticed some
> new initiatives like PCL - cartography.
> You denoted that 'Python module in GEOS is not actively maintained' which
> is indeed a warning.
> Could you may give some more reflection and perspective on this?
> Are there any alternatives in which GEOS is somehow embedded and extended
> to python scripting level. e.g. initiatives like PCL?
> It might be helpful to explain some more what we are in to. We're a
> research-group of structural geologists working on new field acquisition
> and post field data processing procedures. We try to characterize the
> fracture patterns in rocks on cms to 100s of meters scale. For this we
> acquire fractures digitized as polylines using a tablet pc and a
> customized arcpad application. In the post-field processing we prefer to
> remain independent of esri solutions and we've developed a small fracture
> statistics routine. This routine is written with python and reads fracture
> data from shapefiles
> using the ogr-library and then some basic spatial statistics (fracture
> spacing etc...) is being calculated using GEOS-functionality. In a pilot
> project, using linux and with shapefiles as data-repository this worked
> well.  Further improving coding and aiming for advanced processing my
> migrating data to postgresql+postgis and adding user friendliness by
> porting code to windows environment we can still obtain a working code.
> However have run in some problems and also when cleaning up of code we've
> noticed some untidy, mixed and inconsistent use of OGR and GEOS libraries.
> Any suggestions on devolopers track or getting connected with devolpers
> pursuing comparable goals would help us.
> thanks,
> Nico

Hi Nico,

Since you are working on windows, I recommend that you introduce 
yourself to the GDAL/OGR community, and talk to the developers about 
adding the functionality that you need. PCL (I am the lead programmer) 
has a nicer Python API, but is not as mature as ogr.py, and has no 
current binary distribution for win32.


Sean Gillies

More information about the geos-devel mailing list