def fin_cols(col_l, col_r, id_l, id_r):
fin = ['ltable.'+id_l, 'rtable.'+id_r]
l = mg.diff(col_l, [id_l])
l = ['ltable.'+ x for x in l]
fin.extend(l)
r = mg.diff(col_r, [id_r])
r = ['rtable.'+ x for x in r]
fin.extend(r)
return fin
def predict_ex_attrs(self, table, exclude_attrs):
if not isinstance(exclude_attrs, list):
exclude_attrs = [exclude_attrs]
table = table.to_dataframe()
attrs_to_project = mg.diff(table.columns, exclude_attrs)
x = table[attrs_to_project]
y = self.predict_sklearn(x, check_rem=False)
return y
def predict_ex_attrs(self, table, exclude_attrs):
if not isinstance(exclude_attrs, list):
exclude_attrs = [exclude_attrs]
table = table.to_dataframe()
attrs_to_project = mg.diff(table.columns, exclude_attrs)
x = table[attrs_to_project]
y = self.predict_sklearn(x, check_rem=False)
return y
def get_xy_data_ex(table, exclude_attrs, target_attr):
if not isinstance(exclude_attrs, list):
exclude_attrs = [exclude_attrs]
attrs_to_project = mg.diff(table.columns, exclude_attrs)
table = table.to_dataframe()
x = table[attrs_to_project].values
y = table[target_attr].values
y = y.ravel() # to mute warnings from svm and cross validation
return x, y
def get_xy_data_ex(table, exclude_attrs, target_attr):
if not isinstance(exclude_attrs, list):
exclude_attrs = [exclude_attrs]
attrs_to_project = mg.diff(table.columns, exclude_attrs)
table = table.to_dataframe()
x = table[attrs_to_project].values
y = table[target_attr].values
y = y.ravel() # to mute warnings from svm and cross validation
return x, y
def get_data_in_xy_format_given_ex_attrs(table, exclude_attrs, target_attr):
if not isinstance(exclude_attrs, list):
exclude_attrs = [exclude_attrs]
# get a set with key and foreign keys
s = set([table.get_key(), table.get_property('foreign_key_ltable'), table.get_property('foreign_key_rtable')])
exclude_attrs = list(set(exclude_attrs).union(s))
attrs_to_project = diff(table.columns, exclude_attrs)
x = table[attrs_to_project].values
y = table[target_attr].values
return x, y
def fit_ex_attrs(self, table, exclude_attrs, target_attr):
# assume the exclude attrs and target attr is present
if not isinstance(exclude_attrs, list):
exclude_attrs = [exclude_attrs]
attrs_to_project = mg.diff(table.columns, exclude_attrs)
table = table.to_dataframe()
x = table[attrs_to_project]
#print attrs_to_project
y = table[target_attr]
self.fit_sklearn(x, y, check_rem=False)
def fit_ex_attrs(self, table, exclude_attrs, target_attr):
# assume the exclude attrs and target attr is present
if not isinstance(exclude_attrs, list):
exclude_attrs = [exclude_attrs]
attrs_to_project = mg.diff(table.columns, exclude_attrs)
table = table.to_dataframe()
x = table[attrs_to_project]
#print attrs_to_project
y = table[target_attr]
self.fit_sklearn(x, y, check_rem=False)
def get_data_in_xy_format_given_ex_attrs(table, exclude_attrs, target_attr):
if not isinstance(exclude_attrs, list):
exclude_attrs = [exclude_attrs]
# get a set with key and foreign keys
s = set([
table.get_key(),
table.get_property('foreign_key_ltable'),
table.get_property('foreign_key_rtable')
])
exclude_attrs = list(set(exclude_attrs).union(s))
attrs_to_project = diff(table.columns, exclude_attrs)
x = table[attrs_to_project].values
y = table[target_attr].values
return x, y
