[SoC] SoC Report: Homerange analyses in QGIS with R

[Anne Ghisla]

Hi all,

this week I fighted a lot with how to make R and Python interact.
The problem I pointed out on qgis-dev and RPy lists [1] is now solved.
I'll put the code on SVN next week, as the machine on which I'm working won't be
connected until Tuesday. 
Now a full MCP analysis is callable through its window. I have to fix how some
options are passed to the plugin, because for the moment the path to R scripts
is peculiar of my personal installation, and the folder where output shapefiles
are written is the R working directory (r.setwd() returns an error). 
In addition, I will protect the execution from errors coming from uncorrect
data. At the moment, Python exceptions are raised but not handled.

Next week I will review the main scheme of the plugin. There will be two ways of
running the analyses: MCP or kernel alone, or a batch mode. This mode will
collect all options for MCP and kernel analyses, and then run them all in a
bunch. This means changing some signals/slots and refactor a bit.

I'm experiencing some difficulties with RPy and Python, but in overall I've got
no blocks.

regards,

Anne Ghisla

[1] http://lists.osgeo.org/pipermail/qgis-developer/2008-July/004270.html