Writing Loop Script for GRASS

Question

I am new to Python and scripting in general.

My overall goal is to create a script that I can cycle through all the raster maps in a mapset and then save the univariate stats to a .csv file for each raster map. I am working on the first part: cycling through all the maps and printing the stats to the terminal window. Later, I will try and save the data to a .csv file. I have seven maps in my practice mapset and I get seven errors, but my problem is that I do not know what variable to assign to the parameter MAP (hopefully, I stated that correctly).

Here is what I have thus far:

import sys
import os
import grass.script as grass
import grass.script.setup as gsetup
gisbase = os.environ['GISBASE']
gisdb="/Users/kc/GrassData"
location="Data_2013"
mapset="methods_check"
gsetup.init(gisbase, gisdb, location, mapset)
grass.run_command("g.list", _type="rast") 
mymaps = grass.parse_command("g.list", _type="rast") 
print len(mymaps) 
for items in mymaps:
 grass.run_command('r.univar', map='items', separator=",")

What should I put for map = 'items' so that it cycles through all the raster maps of a mapset?

I am using GRASS 7.0, but the same thing happens in GRASS 6.5. In Grass 6.5, the parameter separator must be changed to fs. Both versions label the input file "map".

Answer 1

You need to understand the difference between:

• grass.run_command which performs a GRASS command

grass.run_command('g.remove', raster="raster1")

• grass.read_command which is used to receive a result (generally a string, same result as the command in GRASS GIS, of little interest to Python, you need to use the re (Regular Expression) module to process the result)

grass.read_command("g.region",flags="p") 
'projection: 99 (Lambert Conformal Conic)\nzone: 0\ndatum: bel72\nellipsoid: international\nnorth: 112990.07863053\nsouth: 108258.30027318\nwest: 227552.40388837\neast: 235027.97855494\nnsres: 6.35138035\newres: 6.35138035\nrows: 745\ncols: 1177\ncells: 876865\n'

• grass.parse_command that parse the result for Python (generally a dictionary)

grass.parse_command("g.region",flags="p") 
{'ellipsoid: international': None, 'zone: 0': None, 'nsres: 6.35138035': None, 'south: 108258.30027318': None, 'ewres: 6.35138035': None, 'west: 227552.40388837': None, 'rows: 745': None, 'east: 235027.97855494': None, 'projection: 99 (Lambert Conformal Conic)': None, 'cells: 876865': None, 'north: 112990.07863053': None, 'cols: 1177': None, 'datum: bel72': None}

So, in your case, it is not:

grass.run_command("g.list", _type="rast")

whose result is 0

but:

rasters = grass.read_command("g.list",_type="rast")

whose result is a string with all the layers in the variable rasters

type(raster)
<type 'str'>

But it is not interesting at all here, because you only need the names of the layers:

rasters = grass.parse_command("g.list", _type="rast")

whose result is a Python dictionary and for further, see Python Scripts For GRASS GIS: Using "parse_command()"

The result is a list with the layer names:

['raster1', 'raster2', 'raster3']

so the processing is:

for items in mymaps:
 grass.parse_command('r.univar', map=items) 

.

• user22964: Thank you. I am still carefully working through your valuable comments (I am new to this.). If I understand correctly, my primary issue is that I need to parse the dictionary that holds my maps. I need more time to work on it.