GRASS GIS Python scripting with grass.script package¶
Parts of the API¶
Syntax¶
First, try a standard command in Console tab in Layer Manager in GRASS GUI:
r.info map=elevation -g
We are running r.info with an option map
set to elevation
Now,
switch to Python tab and type the same command but in Python syntax:
gs.read_command('r.info', map='elevation', flags='g')
We used function read_command()
from the grass.script
package
which is imported under the name gs
in the Python tab in GRASS
GUI. There are also other functions besides read_command()
most
notably run_command()
, write_command()
and parse_command()
.
The first parameter for functions from this group is the name of the
GRASS module as string. Other parameters are options of the module.
Python keyword arguments syntax is used for the options. Flags can be
passed in a parameter flags
where value of the parameter is a string
containing all the flags we want to set. The general syntax is the
following:
function_name('module.name', option1=value1, option2=..., flags='flagletters')
The function parameters are the same as module options, so you can just use standard module manual page to learn about the interface.
Most of the GRASS functionality is available through modules and all of
them can be called using the functions above. However, in some cases, it
is more advantageous to use specialized Python functions. This is the
case for mapcalc()
function (wrapper for r.mapcalc module) and
list_strings()
function (wrapper for g.list module).
Combining multiple modules¶
To launch a Python script from GUI, use File -> Launch Python script.
import grass.script as gs
def main():
input_raster = 'elevation'
output_raster = 'high_areas'
stats = gs.parse_command('r.univar', map='elevation', flags='g')
raster_mean = float(stats['mean'])
raster_stddev = float(stats['stddev'])
raster_high = raster_mean + raster_stddev
gs.mapcalc('{r} = {a} > {m}'.format(r=output_raster, a=input_raster,
m=raster_high))
if __name__ == "__main__":
main()
Processing many maps¶
import grass.script as gs
def main():
rasters = ['lsat7_2002_10', 'lsat7_2002_20', 'lsat7_2002_30', 'lsat7_2002_40']
max_min = None
for raster in rasters:
stats = gs.parse_command('r.univar', map=raster, flags='g')
if max_min is None or max_min < stats['min']:
max_min = stats['min']
print(max_min)
if __name__ == "__main__":
main()
Providing GRASS module interface to a script¶
#!/usr/bin/env python3
# %module
# % description: Adds the values of two rasters (A + B)
# % keyword: raster
# % keyword: algebra
# % keyword: sum
# %end
# %option G_OPT_R_INPUT
# % key: araster
# % description: Name of input raster A in an expression A + B
# %end
# %option G_OPT_R_INPUT
# % key: braster
# % description: Name of input raster B in an expression A + B
# %end
# %option G_OPT_R_OUTPUT
# %end
import sys
import grass.script as gs
def main():
options, flags = gs.parser()
araster = options['araster']
braster = options['braster']
output = options['output']
gs.mapcalc('{r} = {a} + {b}'.format(r=output, a=araster, b=braster))
return 0
if __name__ == "__main__":
sys.exit(main())
The options which has something like G_OPT_R_INPUT
after the word
option
are called standard options. Their list is accessible
in GRASS GIS C API documentation of STD_OPT
enum from gis.h
file.
Always use standard options if possible. They are not only easier to use
but also ensure consistency across the modules and easier maintanenace
in case of updates to the parameters parsing system.
Typically, you change description
(and/or label
), sometimes key
and answer
. There are also standard flags to be used
with flag
which work in the same way.
The examples of syntax of options and flags (without the G_OPT...
part)
can be obtained from any GRASS module using special --script
flag.
Alternatively, you can use GRASS source code to look how different scripts
actually define and use their parameters.
Note that the previous code samples were missing some whitespace which Python PEP8 style guide requires but this last sample fulfills all the requirements. You should always use pep8 tool to check your syntax and style or set your editor to do it for you. Note also that although a some mistakes in Python code can be discovered only when executing the code due to the dynamic nature of Python, there is a large number of tools such as pep8 or pylint which can help you to identify problems in you Python code.