def get_pop():
lspop = pd.read_csv(
files.sharedpath('social/processed/LandScan2011/gcpregions.csv'),
header=17)
dfir = metacsv.read_csv(files.sharedpath("regions/hierarchy_metacsv.csv"))
dfir = dfir.loc[dfir.is_terminal].reset_index(drop=True)
dfir.drop([
'parent-key', 'name', 'alternatives', 'is_terminal', 'gadmid',
'agglomid', 'notes'
],
axis=1,
inplace=True)
dfir.rename(columns={'region-key': 'region'}, inplace=True)
dfir['iso'] = dfir.region.apply(lambda x: x[:3])
df = pd.read_csv(
files.sharedpath(
'social/baselines/population/merged/population-merged.all.csv'))
df.rename(columns={'value': 'pop'}, inplace=True)
df['iso'] = df.region.apply(lambda x: x[:3])
df.drop(['index'], axis=1, inplace=True)
df = df.loc[df.year > 2005]
dff = df.append(get_missing_pop(df, dfir, lspop),
sort=False).sort_values(['region',
'year']).reset_index(drop=True)
return dff
def __init__(self, iam, ssp):
"""iam and ssp should be as described in the files (e.g., iam = 'low', ssp = 'SSP3')"""
super(GDPpcProvider, self).__init__(iam, ssp, baseline_year)
# Load the baseline values, and split by priority of data
df = metacsv.read_csv(files.sharedpath(baseline_filepath))
self.df_baseline_this = df.loc[(df.model == iam) & (df.scenario == ssp)
& (df.year == baseline_year)]
self.df_baseline_anyiam = df.loc[(df.scenario == ssp) & (
df.year == baseline_year)].groupby('iso').median()
self.baseline_global = df.loc[(df.scenario == ssp)
& (df.year == baseline_year)].median()
# Load the growth rates, and split by priority of data
df_growth = metacsv.read_csv(files.sharedpath(growth_filepath))
df_growth['yearindex'] = np.int_((df_growth.year - baseline_year) / 5)
self.df_growth_this = df_growth.loc[(df_growth.model == iam)
& (df_growth.scenario == ssp)]
self.df_growth_anyiam = df_growth.loc[(
df_growth.scenario == ssp)].groupby(['iso', 'year']).median()
self.growth_global = df_growth.loc[(df_growth.scenario == ssp)
& (df_growth.model == iam)].groupby(
['year']).median()
# Load the nightlights
self.df_nightlights = metacsv.read_csv(
files.sharedpath(nightlights_filepath))
# Cache for ISO-level GDPpc series
self.cached_iso_gdppcs = {}
def get_gdppc_all_regions(model, ssp):
print(model + " " + ssp)
os.chdir(os.getenv("HOME") + '/repos/impact-calculations')
moddict = {'high': 'OECD Env-Growth', 'low': 'IIASA GDP'}
df = metacsv.read_csv(files.sharedpath("regions/hierarchy_metacsv.csv"))
tmplist = []
yvect = np.arange(2010, 2101)
svect = np.repeat(ssp, yvect.size)
mvect = np.repeat(moddict[model], yvect.size)
provider = gdppc.GDPpcProvider(model, ssp)
for ii in np.where(df.is_terminal)[0]:
series = provider.get_timeseries(df['region-key'][ii])
rvect = np.repeat(df['region-key'][ii], yvect.size)
tmp = pd.DataFrame({
'region': rvect,
'year': yvect,
'gdppc': series,
'ssp': svect,
'model': mvect
})
tmplist.append(tmp)
dfg = pd.concat(tmplist, sort=False)
return dfg
def test_xyz_high():
if files.sharedpath('testing')[:11] == '/shares/gcp':
helper(provider_high, 'XYZ.1.2', 7065.37128841, 15837.3384374,
16024.4588895)
else:
helper(provider_high, 'XYZ.1.2', 415.320976851, 1653.5117662281498,
1711.9155483103712)
def test_xyz_low():
if files.sharedpath('testing')[:11] == '/shares/gcp':
helper(provider_low, 'XYZ.1.2', 7065.37128841, 12309.8299344,
12417.4055742)
else:
helper(provider_low, 'XYZ.1.2', 415.320976851, 1254.179869326625,
1269.0353748748823)
if df.shape[0] > 0:
return df
if iso in df_anyiam.index:
df = df_anyiam.loc[iso]
if df.shape[0] > 0:
return df
return df_global
if __name__ == '__main__':
# Test the provider
import time
time0 = time.time()
provider = GDPpcProvider('low', 'SSP3')
df = metacsv.read_csv(files.sharedpath("regions/hierarchy_metacsv.csv"))
time1 = time.time()
print "Load time: %s seconds" % (time1 - time0)
for ii in np.where(df.is_terminal)[0]:
xx = provider.get_timeseries(df.iloc[ii, 0])
time2 = time.time()
print "First pass: %s seconds" % (time2 - time1)
for ii in np.where(df.is_terminal)[0]:
xx = provider.get_timeseries(df.iloc[ii, 0])
time3 = time.time()
print "Second pass: %s seconds" % (time3 - time2)