def calc_caloric_production_on_uri_list(input_uri_list, output_csv_uri):
first_lulc = True
baseline_calories = 0
output = []
for uri in input_uri_list:
row = []
row.append(nd.explode_uri(uri)['file_root'])
af = nd.ArrayFrame(uri)
# nd.pp(nd.enumerate_array_as_odict(af.data))
sum_calories = calc_caloric_production_from_lulc(uri)
row.append(str(sum_calories))
if first_lulc:
baseline_calories = sum_calories
first_lulc = False
else:
row.append(str(sum_calories - baseline_calories))
output.append(row)
hb.python_object_to_csv(output, output_csv_uri)
def calc_caloric_production_on_uri_list(input_uri_list, output_csv_uri):
first_lulc = True
baseline_calories = 0
output = []
af_list = []
for uri in input_uri_list:
row = []
row.append(nd.explode_uri(uri)['file_root'])
af = nd.ArrayFrame(uri)
sum_calories, af = calc_caloric_production_from_lulc(uri)
row.append(str(sum_calories))
# af.show()
af_list.append(af)
if first_lulc:
baseline_calories = sum_calories
first_lulc = False
else:
row.append(str(sum_calories - baseline_calories))
output.append(row)
hb.python_object_to_csv(output, output_csv_uri)
return af_list
def create_scenario_calorie_csv(input_dir, output_uri):
scenario_results = [[
'',
'carbon',
'wy',
'n_export',
'p_export',
'sed_export',
'caloric_production',
'carbon_diff_from_baseline',
'wy_diff_from_baseline',
'n_export_diff_from_baseline',
'p_export_diff_from_baseline',
'sed_export_diff_from_baseline',
'caloric_production_diff_from_baseline',
'carbon_percent_diff_from_baseline',
'wy_percent_diff_from_baseline',
'n_export_percent_diff_from_baseline',
'p_export_percent_diff_from_baseline',
'sed_export_percent_diff_from_baseline',
'caloric_production_percent_diff_from_baseline',
'carbon_diff_from_baseline',
'wy_diff_from_baseline',
'n_export_diff_from_baseline',
'p_export_diff_from_baseline',
'sed_export_diff_from_baseline',
'caloric_production_diff_from_baseline',
'carbon_percent_diff_from_bau',
'wy_percent_diff_from_bau',
'n_export_percent_diff_from_bau',
'p_export_percent_diff_from_bau',
'sed_export_percent_diff_from_bau',
'caloric_production_percent_diff_from_bau',
]]
# Calculate Sum
baseline_results = []
bau_results = []
for scenario in scenarios:
scenario_result = []
scenario_results.append(scenario_result)
scenario_result.append(scenario)
carbon_result_uri = os.path.join(runs_folder, run_name, scenario,
'carbon/tot_c_cur.tif')
carbon = nd.ArrayFrame(carbon_result_uri)
wy = nd.ArrayFrame(
os.path.join(runs_folder, run_name, scenario,
'hydropower_water_yield/output/per_pixel/wyield.tif'))
n_export = nd.ArrayFrame(
os.path.join(runs_folder, run_name, scenario, 'ndr/n_export.tif'))
p_export = nd.ArrayFrame(
os.path.join(runs_folder, run_name, scenario, 'ndr/p_export.tif'))
sed_export = nd.ArrayFrame(
os.path.join(runs_folder, run_name, scenario,
'sdr/sed_export.tif'))
# calories = nd.ArrayFrame(os.path.join(input_dir, scenario, 'caloric_production.tif' ))
calories = nd.ArrayFrame(
os.path.join(runs_folder, run_name, scenario,
'nutritional_adequacy/caloric_production.tif'))
if scenario == 'Baseline':
baseline_results.append(carbon.sum(
)) #np.sum(carbon[carbon.data!=carbon.no_data_value])
baseline_results.append(wy.sum())
baseline_results.append(n_export.sum())
baseline_results.append(p_export.sum())
baseline_results.append(sed_export.sum())
baseline_results.append(calories.sum())
# TODO Generalize BAU
elif scenario == 'Trend':
bau_results.append(carbon.sum())
bau_results.append(wy.sum())
bau_results.append(n_export.sum())
bau_results.append(p_export.sum())
bau_results.append(sed_export.sum())
bau_results.append(calories.sum())
scenario_result.append(str(float(carbon.sum())))
scenario_result.append(str(float(wy.sum())))
scenario_result.append(str(float(n_export.sum())))
scenario_result.append(str(float(p_export.sum())))
scenario_result.append(str(float(sed_export.sum())))
scenario_result.append(str(float(calories.sum())))
if scenario not in ['Baseline']:
scenario_result.append(
str(float(carbon.sum() - baseline_results[0]) * 100))
scenario_result.append(
str(float(wy.sum() - baseline_results[1]) * 100))
scenario_result.append(
str(float(n_export.sum() - baseline_results[2]) * -1.0 *
100)) # -1 means interpret as retention
scenario_result.append(
str(float(p_export.sum() - baseline_results[3]) * -1.0 * 100))
scenario_result.append(
str(
float(sed_export.sum() - baseline_results[4]) * -1.0 *
100))
scenario_result.append(
str(float(calories.sum() - baseline_results[5]) * 100))
scenario_result.append(
str(
float((carbon.sum() - baseline_results[0]) /
baseline_results[0]) * 100))
scenario_result.append(
str(
float(
(wy.sum() - baseline_results[1]) / baseline_results[1])
* 100))
scenario_result.append(
str(
float((n_export.sum() - baseline_results[2]) /
baseline_results[2]) * -1.0 * 100))
scenario_result.append(
str(
float((p_export.sum() - baseline_results[3]) /
baseline_results[3]) * -1.0 * 100))
scenario_result.append(
str(
float((sed_export.sum() - baseline_results[4]) /
baseline_results[4]) * -1.0 * 100))
scenario_result.append(
str(
float((calories.sum() - baseline_results[5]) /
baseline_results[5]) * 100))
if scenario not in ['Baseline', 'BAU', 'Trend']:
scenario_result.append(
str(float(carbon.sum() - bau_results[0]) * 100))
scenario_result.append(str(float(wy.sum() - bau_results[1]) * 100))
scenario_result.append(
str(float(n_export.sum() - bau_results[2]) * -1.0 * 100))
scenario_result.append(
str(float(p_export.sum() - bau_results[3]) * -1.0 * 100))
scenario_result.append(
str(float(sed_export.sum() - bau_results[4]) * -1.0 * 100))
scenario_result.append(
str(float(calories.sum() - bau_results[5]) * 100))
scenario_result.append(
str(
float((carbon.sum() - bau_results[0]) / bau_results[0]) *
100))
scenario_result.append(
str(float((wy.sum() - bau_results[1]) / bau_results[1]) * 100))
scenario_result.append(
str(
float((n_export.sum() - bau_results[2]) / bau_results[2]) *
-1.0 * 100))
scenario_result.append(
str(
float((p_export.sum() - bau_results[3]) / bau_results[3]) *
-1.0 * 100))
scenario_result.append(
str(
float(
(sed_export.sum() - bau_results[4]) / bau_results[4]) *
-1.0 * 100))
scenario_result.append(
str(
float((calories.sum() - bau_results[5]) / bau_results[5]) *
100))
nd.pp(scenario_results)
hazelbean.python_object_to_csv(scenario_results, os.path.join(output_uri))
sed_export.show(output_uri=os.path.join(runs_folder, run_name,
scenario, 'sed_export.png'))
scenario_result.append(str(float(carbon.sum())))
scenario_result.append(str(float(wy.sum())))
scenario_result.append(str(float(n_export.sum())))
scenario_result.append(str(float(p_export.sum())))
scenario_result.append(str(float(sed_export.sum())))
print('carbon sum: ', carbon.sum())
print('wy sum: ', scenario, wy.sum())
print('n_export sum: ', scenario, n_export.sum())
print('p_export sum: ', scenario, p_export.sum())
print('sed_export sum: ', scenario, sed_export.sum())
# differences = []
# print('scenario_results', scenario_results)
# for i in range(len(results_names)):
# bau_difference = scenario_results[1][i] - scenario_results[0][i]
# cons_difference = scenario_results[2][i] - scenario_results[0][i]
# differences.append([bau_difference, cons_difference])
#
# print(results_names[i] + ' bau difference ' + str(bau_difference) + ', cons difference ' + str(cons_difference))
nd.pp(scenario_results)
csv_output_uri = os.path.join(runs_folder, run_name, 'results.csv')
hazelbean.python_object_to_csv(scenario_results,
os.path.join(csv_output_uri))
def zonal_statistics_flex(input_raster,
zone_vector_path,
zone_ids_raster_path=None,
id_column_label=None,
data_type=None,
ndv=None,
zones_ndv=None,
values_ndv=None,
all_touched=None,
assert_projections_same=True,
use_iterblocks=True,
unique_zone_ids=None,
csv_output_path=None,
verbose=True,
rewrite_zone_ids_raster=True):
""" if zone_ids_raster_path is set, use it and/or create it for later processing speed-ups.
Still todo, in the case where there is a STR labeled column, generate some ids. For now it HAS to be ints.
"""
input_path = hb.get_flex_as_path(input_raster)
base_raster_path_band = (input_path, 1)
# Test that input_raster and shapefile are in the same projection. Sillyness results if not.
if assert_projections_same:
hb.assert_gdal_paths_in_same_projection([input_raster, zone_vector_path])
else:
a = hb.assert_gdal_paths_in_same_projection([input_raster, zone_vector_path], return_result=True)
if not a:
L.critical('Ran zonal_statistics_flex but the inputs werent in identical projections.')
# if zone_ids_raster_path is not defined, use the PGP version, which doesn't use a rasterized approach.
if not zone_ids_raster_path and rewrite_zone_ids_raster is False:
to_return = pgp.zonal_statistics(
base_raster_path_band, zone_vector_path,
aggregate_layer_name=None, ignore_nodata=True,
polygons_might_overlap=True, working_dir=None)
if csv_output_path is not None:
hb.python_object_to_csv(to_return, csv_output_path)
return to_return
# if zone_ids_raster_path is defined, then we are using a rasterized approach.
# NOTE that by construction, this type of zonal statistics cannot handle overlapping polygons (each polygon is just represented by its id int value in the raster).
else:
if zones_ndv is None:
zones_ndv = -9999
if ndv is not None:
values_ndv = ndv
if values_ndv is None:
values_ndv = hb.get_raster_info_hb(input_raster)['nodata'][0]
if values_ndv is None:
values_ndv = -9999.0
# if zone_ids_raster_path is given, use it to speed up processing (creating it first if it doesnt exist)
if not os.path.exists(zone_ids_raster_path) and rewrite_zone_ids_raster is not False:
# Calculate the id raster and save it
if verbose:
L.info('Creating id_raster with convert_polygons_to_id_raster')
hb.convert_polygons_to_id_raster(zone_vector_path, zone_ids_raster_path, input_raster, id_column_label=id_column_label, data_type=data_type,
ndv=zones_ndv, all_touched=all_touched)
else:
if verbose:
L.info('Zone_ids_raster_path existed, so not creating it.')
# hb.assert_gdal_paths_have_same_geotransform([zone_ids_raster_path, input_raster])
if verbose:
L.info('Starting zonal_statistics_rasterized using zone_ids_raster_path at ' + str(zone_ids_raster_path))
# Call zonal_statistics_rasterized to parse vars into cython-format and go from there.
unique_ids, sums, counts = hb.zonal_statistics_rasterized(zone_ids_raster_path, input_raster, zones_ndv=zones_ndv, values_ndv=values_ndv,
use_iterblocks=use_iterblocks, unique_zone_ids=unique_zone_ids, verbose=verbose)
df = pd.DataFrame(index=unique_ids, data={'sum': sums})
print(df)
df[df == 0] = np.nan
df.dropna(inplace=True)
if csv_output_path is not None:
df.to_csv(csv_output_path)
# # Convert the raw arrays to a 2d rc odict
# to_return = OrderedDict()
# for i in unique_ids:
# to_return[i] = {'sum': sums[i], 'count': counts[i]}
#
#
# if csv_output_path is not None:
# hb.python_object_to_csv(to_return, csv_output_path, csv_type='cr_2d_odict')
return df