def __init__(self,
variables,
missing=np.nan,
case_sensitive=False,
force_retrieve=False):
self.case_sensitive = case_sensitive
if isinstance(variables, (list, tuple)):
_vars = {}
for v in variables:
v = v.split('/')
if v[0] not in _vars:
_vars[v[0]] = []
if len(v) == 2:
_vars[v[0]].append(v[1])
variables = _vars
dicts = []
for k, v in variables.items():
d = fetch_dictionary(k, force_retrieve=force_retrieve)
if not case_sensitive:
d.index = d.index.str.lower()
if v:
d = d[v]
d.columns = ['{}_{}'.format(k, c) for c in d.columns]
dicts.append(d)
# Make sure none of the dictionaries have duplicate indices
drop_dups = lambda d: d[~d.index.duplicated(keep='first')]
dicts = [d if d.index.is_unique else drop_dups(d) for d in dicts]
dictionary = pd.concat(dicts, axis=1, join='outer', sort=False)
super().__init__(dictionary, missing=missing)
def __init__(self, variables, missing=np.nan, case_sensitive=True):
self.case_sensitive = case_sensitive
if isinstance(variables, (list, tuple)):
_vars = {}
for v in variables:
v = v.split('/')
if v[0] not in _vars:
_vars[v[0]] = []
if len(v) == 2:
_vars[v[0]].append(v[1])
variables = _vars
dicts = []
for k, v in variables.items():
d = fetch_dictionary(k)
if not case_sensitive:
d.index = d.index.str.lower()
if v:
d = d[v]
d.columns = ['%s_%s' % (k, c) for c in d.columns]
dicts.append(d)
dictionary = pd.concat(dicts, axis=1, join='outer')
super(PredefinedDictionaryExtractor, self).__init__(dictionary,
missing=missing)