Hi dev team,
I have been trying to run a few t.rast.mapcalc operations and
encountered the substring dataset name problem described here:
http://trac.osgeo.org/grass/ticket/2735
To my understanding this has never actually been fixed, so I tried to
come up with a fool-proof solution. To prevent successive substitution
of substrings, the dataset names in the expression are first filled with
random string identifiers in descending string length order that can
later be "formated" with exp.format(). See diff attached. This might be
useful for other t.* modules that use the exp.replace() approach.
Since I havent got the latest trunk compiled, I wasnt able to test it in
grass, but the concept should work. Would be great if someone with more
t.* experience than me could test it.
Best regards,
Michel
Index: lib/python/temporal/mapcalc.py
===================================================================
--- lib/python/temporal/mapcalc.py (revision 71405)
+++ lib/python/temporal/mapcalc.py (working copy)
@@ -11,6 +11,8 @@
# i18N
import gettext
import copy
+import string
+import random
from datetime import datetime
from multiprocessing import Process
import grass.script as gscript
@@ -196,6 +198,18 @@
proc_list = []
proc_count = 0
+ # Remove spaces and new lines
+ exp = expression.replace(" ", "")
+ # sort input dataset names by decending length
+ inpr = inputs
+ inpr.sort(lambda x,y: cmp(len(y), len(x)))
+
+ # assign tmp variables in order
+ inputmp = {}
+ for i,n in inpr:
+ inputmp[n] = ''.join(random.choice(string.ascii_letters) for _ in range(12))
+ exp = exp.replace(n, '{%s}' % inputmp[n])
+
# For all samples
for i in range(num):
@@ -205,19 +219,19 @@
# Create the r.mapcalc statement for the current time step
map_name = "{base}_{suffix}".format(base=base,
suffix=gscript.get_num_suffix(count, num))
- # Remove spaces and new lines
- expr = expression.replace(" ", "")
# Check that all maps are in the sample
valid_maps = True
# Replace all dataset names with their map names of the
# current time step
+ fmts = {}
for j in range(len(map_matrix)):
if map_matrix[j][i] is None:
valid_maps = False
break
- # Substitute the dataset name with the map name
- expr = expr.replace(id_list[j], map_matrix[j][i])
+ fmts[inputmp[id_list[j]]] = map_matrix[j][i]
+ # Substitute the dataset name with the map name
+ expr = exp.formate(**fmts)
# Proceed with the next sample
if not valid_maps:
_______________________________________________
grass-dev mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/grass-dev
