[OSGeo-Discuss] OWS Common Python Server. Guidance Required

Dave Sampson samper.d at gmail.com
Sun Jan 3 08:55:03 PST 2010


Sorry for the tardy response, been in the weeds for a while now.

WRT to OWSLIB and QGIS. The CSW client was the first OWS client i made
for QGIS. Mainly a professional development excercise to learn more
about CSW (Python is my language of choice).

in QGIS there is already a C implementation of a WMS and WFS client.

The next clients I am interested in are:
* WPS to consume something like the GRASS WPS implementation
(http://grass.osgeo.org/wiki/WPS)
* OpenLS to consume some location services. I am trying to create a
OpenLS like server using PAGC (http://www.pagcgeo.org/). Currently i
know of no FOSS OpenLS servers.

If I could use something like COWSFramework at some point. right now I
am doing it from scratch to learn about webservices a bit more. then I
would port to a supportable module.

It would also be nice to have python versions of WMS and WFs clients in
Qgis for references.

Anyhow, the thread is a little old but thought I would follow up.

Cheers

On Wed, 2009-08-05 at 10:55 +0100, Dominic Lowe wrote:
> Hi David,
> 
> We have done something along these lines with this:
> http://proj.badc.rl.ac.uk/ndg/wiki/CowsFramework
> 
> I am hesitant to advertise it as an easy API approach to solving your problem 
> as it's not that refined (yet!) but it is an OGC services framework written 
> in python (based on the Pylons web framework) that you can write different 
> backends for and can add new services to. Currently it mostly supports our 
> own data model at the backend but it is in principle fairly generic and could 
> be extended.
> 
> I'm happy to talk more about it if you are interested.
> 
> btw,  I am interested in your OWSLib work on QGIS - is it the CSW support or 
> have you implemented any of the other clients? It would be good to publicise 
> this stuff on the Python GIS list too.
> 
> Cheers,
> 
> Dom
> 
> 
> 
> 
> On Tuesday 04 August 2009 14:44:05 Sampson, David wrote:
> > Hey Folks,
> >
> > I have recently used the OWSLIB library
> > (http://trac.gispython.org/lab/wiki/OwsLib) for some python based
> > plugins in QGIS (http://sourceforge.net/projects/qgiscommunitypl/). I am
> > also involved with PAGC, a public address geocoder project
> > (http://sourceforge.net/projects/pagc/).
> >
> > OWSLIB made the plugins development a lot smoother and I look forward to
> > doing some more (mainly to learn about OGC specs)
> >
> > PAGC is gearing its development towards the OLS (Open Location Service)
> > spec. [http://www.opengeospatial.org/standards/ols]
> >
> > I am interested in creating a small python server for OLS and
> > WPS....Python mainly because it is what I know and I believe a good
> > learning tool. Then I thought I must not be the first to think this.
> >
> > Recent discussion on the PAGC mailing list suggested the possibility of
> > having a generic OLS server that can have various modules plugged into
> > it. For example Geocoding for one and Routing for another.
> >
> > Then I thought back to the OWSLIB approach where they have an OWS common
> > module and then separate modules for the specific specs. But I believe
> > the lib focuses on clients, and not servers.
> >
> > With the agreement between OSGEO and OGC would there be possible work or
> > collaboration on creating a base server (python based?) that could be
> > used to create various servers of all the specs. An easy API approach to
> > creating OGC services?
> >
> > I know this might start the language battle, I am thinking of ways of
> > lowering the educational barrier and provide examples...  Both the C
> > camp (though providing bindings to a c based lib) and Java camp (by
> > providing Jython and python bindings) could have their own performance
> > driven approach.
> >
> >
> > Perhaps this is already done?
> >
> > Anyone intrigued?
> >
> > Cheers
> 
> 
> _______________________________________________
> Discuss mailing list
> Discuss at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/discuss





More information about the Discuss mailing list