niche_vlaanderen version: master version 24/1/2023
Python version: 3
Operating System: win
I tried to replace hab_nich_join.csv by a custom version
hab_niche_join_22_23.csv with all the types previously defined as
type 22 => now type 28
type 23 => now type 28
I get strange results for area_pot/non pot
Using my test vegetation map bwk_fake_extentok.zip with Brasschaat (you will need to adapt the paths):
# run niche model
simplebr = nv.Niche()
simplebr.run_config_file("../tests/data/bwk/niche_brasschaat/simple.yaml")
# validation
from niche_vlaanderen.validation import NicheValidation
no_simplebr_fkok_2223 = NicheValidation(niche=simplebr, map= path_ch + "testcase_extra/brasschaat/bwk/bwk_fake_extentok.shp",
mapping_file= path_ch +"system_tables/hab_niche_join_22_23.csv")
no_simplebr_fkok_2223.overlay()
# (slechte) kaart met potenties én vegetatiekaart voor type 28
output_niche_28 = simplebr.plot(28)
# get bounds van de vegetatiekaart
xmin, ymin, xmax, ymax = no_simplebr_fkok_2223.map.total_bounds
output_niche_28.set_xlim(xmin-25, xmax+25) # 25 = buffer around the limits
output_niche_28.set_ylim(ymin-25, ymax+25)
mymap = no_simplebr_fkok_2223.map.boundary.plot(ax=output_niche_28)
no_simplebr_fkok_2223.map.apply(lambda x: mymap.annotate(text=str(x['id']-1) + "\n" +
str(x['pHAB1']) + "% " + x['HAB1'] + "\n" +
str(x['pHAB2']) + "% " + str(x['HAB2']) + "\n" +
str(x['pHAB3']) + "% " + str(x['HAB3']) + "\n" +
str(round(x['area_shape'])) + " ha",
xy=x.geometry.centroid.coords[0],
ha='center',
color = "magenta"),
axis=1);
# shape 0 50% type 28 (vroeger 50% 22)
# shape 1 20% 28 (vroeger 20% 22)
# shape 2 nvt
# shape 3 95% 28 (vroeger 60% 22 en 35% 23)
# shape 4 40% 28 (vroeger 40% 22)
# shape 5 100 % 28 (vroeger 100% 22 en 20 % 28)
# show tables for area_pot and nonpot
no_simplebr_fkok_2223.area_pot
no_simplebr_fkok_2223.area_nonpot
# compare with values in table with potential_presence
no_simplebr_fkok_2223.potential_presence
shape_id 3 = 119 ha
calculated area_pot 28 = 28 ha (while potential_presence type 28/present was 14)
calculated area_nonpot 28 = 211 ha (while potential_presence type 28/not present was 105)
area_pot 28 + area_nonpot 28 => in total much more than 119 ha
I tried to replace hab_nich_join.csv by a custom version hab_niche_join_22_23.csv with all the types previously defined as
I get strange results for area_pot/non pot
Using my test vegetation map bwk_fake_extentok.zip with Brasschaat (you will need to adapt the paths):
shape_id 3 = 119 ha calculated area_pot 28 = 28 ha (while potential_presence type 28/present was 14) calculated area_nonpot 28 = 211 ha (while potential_presence type 28/not present was 105) area_pot 28 + area_nonpot 28 => in total much more than 119 ha
(maybe related to https://github.com/inbo/niche_vlaanderen/issues/314?)