Alas, coordinate sequences are stored as CPoint objects in python lists, which is less than ideal for portability but made sense at the time to get it out the door.
It wouldn't take much effort to generalize the library for someone knowledgeable of the shapely interface, and correct some of the lesser design decisions (all of them mine ;). http://github.com/umidev/ligeos/blob/6087087a5da91d2d0813b06084eac1feb5b5bd55/ligeos/linearref.pyx I'd be happy to revisit this with a bit of academic and specific guidance to points where the buffer access is used by shapely. Nino -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Sean Gillies Sent: Thursday, August 20, 2009 8:08 AM To: gispython.org community projects Subject: Re: [Community] Interest in porting PostGIS analytic functions to Python? Nino, that's a neat package. I wonder if there's potential for easy integration with Shapely 1.1, where coordinate sequences are implemented as Python arrays, using Cython buffer access: http://wiki.cython.org/enhancements/buffer Have you used the above at all? On Aug 19, 2009, at 6:29 PM, Nino Walker wrote: > I should also note a similar effort for linear referencing > (locating, substringing, concatenating, distances, nearest etc). > > http://github.com/umidev/ligeos/tree/master > > There are some very specific things about the library: it works with > geographic coordinate systems without needing reprojection; it works > independent of geos/shapely/etc.; and it is compiled to C for higher > performance (well, Cython generated C). > > We've used it extensively for working with/manipulating line > topologies. > > It is stable, used in production, etc. > > I understand GEOS will include some linear referencing tools in the > future, and hopefully this can be superseded by that. But in the > meantime, it works well and fills a void in the stack. > > - Nino > > > -----Original Message----- > From: [email protected] > [mailto:[email protected] > ] On Behalf Of Sean Gillies > Sent: Wednesday, August 19, 2009 3:32 AM > To: gispython.org community projects > Cc: Christopher Helm > Subject: [Community] Interest in porting PostGIS analytic functions > to Python? > > Michael Elsdörfer has ported line_locate_point from PostGIS to Python > > http://bitbucket.org/miracle2k/pyutils/changeset/156c60ec88f8/ > > Is there anybody else who would be interested in porting more of the > functions from > > > http://trac.osgeo.org/postgis/browser/trunk/lwgeom/lwgeom_functions_analytic.c > > to Python, using Shapely and/or Numpy? I don't think it would have to > be a systematic effort, just one that adds the functions as needed, > and then optimizes them (using Cython, for example) when we really > need more performance. > > If there was interest, we'd start a project on bitbucket based on > Michael's code. Would have to be GPL if it's derived from PostGIS. > > -- > Sean Gillies > Programmer > Institute for the Study of the Ancient World > New York University > -- Sean _______________________________________________ Community mailing list [email protected] http://lists.gispython.org/mailman/listinfo/community _______________________________________________ Community mailing list [email protected] http://lists.gispython.org/mailman/listinfo/community
