This repo has been archived for historical purposes.

A python toolkit to tame the wild GRASS.

pip install git+https://github.com/perrygeo/mower.git@master#egg=mower

The goal is to provide a dead-simple context manager to handle GRASS Session state, allowing you to run GRASS 7 commands via pygrass without having to deal with ...

Even to GIS analysts who love code, GRASS presents a significant conceptual barrier in it's system of locations and mapsets. They serve a purpose: to organize data into logical bundles based on spatial reference and themes. But many times we don't have a mapset; we just have a raster or vector dataset and we want to use GRASS for some analysis.

The old workflow goes like this:

• create a new blank mapset
• import the dataset (making a copy) and create a new location from it's spatial reference
• exit grass and reenter the new mapset
• perform the analysis
• export the results back to files or databases or their intended format
• clean up by removing the datasets or the LOCATION entirely

All of this required careful management of environment variables, a directory structure to satisfy the GRASS location/mapset requirements and use of shell/system calls even when using Python.

GRASS 7 and the new pygrass module solves the system call issue by providing a python API to wrap the underlying C libs directly. But you still need to perform extensive work to make sure you've set up a location/mapset and that your system is set up to work in that environment.

Why can't it be this easy?

• Start a GRASS session with a dataset (The mapset and location are handled transparently)
• Do some GRASS analysis
• That's it ... the GRASS session is automatically cleaned up

Well now it is. Treat your grass session as a python context manager and use pygrass however you want without dealing with location/mapset cruft:

from mower import GrassSession

DEM = "/tmp/dem_meters.img"

with GrassSession(DEM) as gs:
    from grass.pygrass.modules.shortcuts import raster

    # Import/Link to External GDAL data
    raster.external(input=DEM, output="dem")

    # Perform calculations
    raster.mapcalc(expression="demft=dem*3.28084")

    # Export from GRASS to GDAL
    raster.out_gdal("demft", format="GTiff", output="/tmp/demft.tif", overwrite=True)

For contrast, see the example wiki code for working with GRASS without explicitly starting it.

If you want to work in the shell instead of python, try this built-in trick

grass71 -c dem.tif ~/grassdata/mylocation

Which will create a new PERMANENT mapset in the mylocation location within the grassdata gisdbase and dump you directly into that session. Try with -e if you want to create but immediately exit.