from ddf_utils.factory.ihme import IHMELoader
from ddf_utils.chef.helpers import read_opt
def _corrected_make_query(self, context, version, **kwargs):
"""generate a query with the context, version and all keyword arguments.
if a parameter is mandatory but not provided, it will fill
with default values.
"""
# metadata
if not self.metadata:
self.load_metadata()
metadata = self.metadata
ages = metadata['age']['id'].values
# location: there is a `custom` location. don't include that one.
locations_md = metadata['location']
locations = locations_md[locations_md['id'] != 'custom']['id'].tolist()
sexs = metadata['sex']['id'].tolist()
years = metadata['year']['id'].tolist()
metrics = metadata['metric']['id'].tolist()
measures = metadata['measure']['id'].tolist()
causes = metadata['cause']['id'].tolist()
# risk/etiology/impairment
# There are actually 4 types data in this dictionary:
# risk, etiology, impairment and injury n-codes.
# however injury n-codes is not enabled.
# we might need to take of it later.
rei = metadata['rei']
# risks = rei[rei['type'] == 'risk']['rei_id'].values
# etiologys = rei[rei['type'] == 'etiology']['rei_id'].values
# impairments = rei[rei['type'] == 'impairment']['rei_id'].values
# others metadata filed not include:
# - groups
# - year_range
queries = {}
# create query base on context and user input
if context == 'le':
measure = read_opt(kwargs, 'measure', default=26)
metric = read_opt(kwargs, 'metric', default=5)
cause = read_opt(kwargs, 'cause', default=causes)
queries.update({
'measure[]': measure,
'metric[]': metric,
'cause[]': cause
})
elif context == 'cause':
measure = read_opt(kwargs, 'measure', default=measures)
metric = read_opt(kwargs, 'metric', default=[1, 2, 3])
cause = read_opt(kwargs, 'cause', default=causes)
queries.update({
'measure[]': measure,
'metric[]': metric,
'cause[]': cause
})
elif context in ['risk', 'etiology', 'impairment']:
# TODO: should be split, don't combine these context here.
measure = read_opt(kwargs, 'measure', default=measures)
metric = read_opt(kwargs, 'metric', default=[1, 2, 3])
context_values = rei[rei['type'] == context]['id'].tolist() ################
context_values = read_opt(kwargs, 'rei_id', default=[1, 2, 3]) ##################
cause = read_opt(kwargs, 'cause', default=causes)
queries.update({
'measure[]': measure,
'metric[]': metric,
context+'[]':context_values,
'rei[]': context_values,
'cause[]': cause
})
else:
# SEV/HALE/haqi
print('not supported context.')
raise NotImplementedError
# insert context and version and other configs
rows = read_opt(kwargs, 'rows', default=10000000) # the maximum records we can get
# ^ Note: user guide[1] says it's 500000 row. But actually we can set this to 10000000
# [1]: http://www.healthdata.org/sites/default/files/files/Data_viz/GBD_2017_Tools_Overview.pdf
email = read_opt(kwargs, 'email', default='downloader@gapminder.org')
idsOrNames = read_opt(kwargs, 'idsOrNames', default='ids') # ids / names / both
singleOrMult = read_opt(kwargs, 'singleOrMult', default='multiple') # single / multiple
base = read_opt(kwargs, 'base', default='single')
location = read_opt(kwargs, 'location', default=locations)
age = read_opt(kwargs, 'age', default=ages)
sex = read_opt(kwargs, 'sex', default=sexs)
year = read_opt(kwargs, 'year', default=years)
queries.setdefault('context', context)
queries.setdefault('version', version)
queries.setdefault('rows', rows)
queries.setdefault('email', email)
queries.setdefault('idsOrNames', idsOrNames)
queries.setdefault('singleOrMult', singleOrMult)
queries.setdefault('base', base)
queries.setdefault('location[]', location)
queries.setdefault('age[]', age)
queries.setdefault('sex[]', sex)
queries.setdefault('year[]', year)
return queries
IHMELoader._make_query = _corrected_make_query
