Dear Pietro, the issue is apparently not related to the GRASS build but to the fact that when there is a mask on the region, the areas outside it are converted to "nan" in the numpy matrix when the map is FCELL, and to the value "-2147483648" when the map is CELL... at least for me! I need to apply a mask because I intend to set up an iterative process in different regions, but I guess it can be just solved by filtering the matrix before the analysis. I am uploading the code to a github repo [1], just in case anyone is interested or it could serve as an example.
Thank you and cheers, Javier [1] https://github.com/javimarlop/eHab_grasspy On Wed, Mar 5, 2014 at 12:15 PM, Pietro <[email protected]> wrote: > Hi Javier, > > On Wed, Mar 5, 2014 at 10:33 AM, Javier Martínez-López > <[email protected]> wrote: >> ... something is wrong with my installation... I cannot even open any >> map from the NC dataset using build 59147, getting the following error >> after interrupting the idle process [1], while doing it with version >> 57982 works fine... any idea or advice? Here is the built >> configuration that was used for the 59147 version: > > [cut] > >> the only difference is "--with-lapack" in the upper one...is there >> something wrong with it? > > I don't think so, I'm using this flag too. > > Try to build everything from scranch, with: > > make distclean > ./configure blabla > make -j4 > > I've run your code using the current svn version (r59193) and it is > working on my laptop. > The only thing is that the region in my case is set to the elevation > raster map, so the shape is different: > > In [1]: import numpy as np > > In [2]: import grass.pygrass as pygrass > > In [3]: lu = pygrass.raster.RasterNumpy('landuse') > > In [4]: lu.shape > Out[6]: (1350, 1500) > > In [5]: lu.open() > > In [6]: lu[:4, :3] > Out[6]: > RasterNumpy([[5, 5, 5], > [5, 5, 5], > [5, 5, 5], > [5, 3, 3]], dtype=int32) > > In [7]: lu.close() _______________________________________________ grass-dev mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/grass-dev
