[GRASS-dev] eval option in r.mapcalc in a python script

[Moritz Lennert]

On 14/01/16 09:34, Paulo van Breugel wrote:
>  From the r.mapcalc manual, the 'eval' option is described to run more
> complex expressions. But I wonder how I apply the example below in a
> python script?


> eval(elev_200 = elevation - 200, \
> elev_5 = 5 * elevation, \
> elev_p = pow(elev_5, 2))
> elevation_result = (0.5 * elev_200) + 0.8 * elev_p

Can't you just put the above in a string variable and feed it to the 
'expression' parameter of grass.mapcalc (or alternatives 
grass.run_command('r.mapcalc', ...)?


>
> Somewhat related, if I have a large number of maps. For each map I need
> to compute a value. At the end, the resulting maps need to be summed up.
> One way is to run a loop, and use r.series to sum up the results. E.g.,
>
> for i in xrange(len(IN)):
>      out = 'shi' + str(i)
>      out.append(out)
>      grass.mapcalc("$out = log($in)", out=out[i], in=in[i])
> grass.run_command("r.series", output="endresult", input=out, method="sum")
>
>
> This generates a lot of intermediate layers, which can potentially take
> up a lot of space on HD. To avoid that, I could do
>
> grass.mapcalc("$out = 0", out=out)
> for i in xrange(len(IN)):
>      grass.mapcalc("$out = $out + log($in)", out=out, in=in[i]),
> overwrite=True
>

ISTR that modifying the contents of a map in-place might be possible, 
but not recommended.

As a compromose you could do something like this:

grass.mapcalc("$oldout = 0")
for i in xrange(len(IN)):
      grass.mapcalc("$newout = $oldout + log($in)", out=out, in=in[i])
      grass.run_command('g.rename', '$newout,$oldout', overwrite=True)


> An alternative is to construct one large expression along the lines of
> expression = "log(in[1]) + log(in[2] + ....". I assume this will be
> faster? It would result in a unwieldy large expression though, and I
> remember something about a limit to the length of the expression?  So
> that is why I wonder if using the 'eval' option would be useful.
>
> Any tips for the best approach?


Maybe one option, as I could image others to have this kind of issue, 
would be to add a "modifier" (or similar name) parameter to r.series 
which would allow to calculate the statistic defined by "method" on a 
modified version of the maps (e.g. squared, square root, log, sin, cos, 
etc.).


Moritz