def test_split_panda_default_column_names(self):
# Columns can be named as key words in pandas
sad = emp.sad(self.pat2, self.cols2, splits="mean:2", clean=False)
assert_equal(len(sad[1][1]), 2)
sad = emp.sad(self.pat2, self.cols2, splits="mean:2; y:3", clean=True)
assert_equal(len(sad[1][1]), 2)
def test_clean(self):
# No a in second split on x
sad = emp.sad(self.pat1, self.cols1, 'x:2', clean=False)
assert_equal(len(sad[1][1]), 2) # Both spp when clean False
sad = emp.sad(self.pat1, self.cols1, 'x:2', clean=True)
assert_equal(len(sad[1][1]), 1) # Only 'b' when clean True
def test_one_way_uneven_split(self):
# 0.2 should fall in second division of y
sad = emp.sad(self.pat1, self.cols1, 'y:2')
assert_equal(len(sad), 2)
assert_equal(sad[0][1]['spp'].values, ['a'])
assert_equal(sad[0][1]['y'].values, [2])
assert_equal(sad[1][1]['spp'].values, ['a','b'])
assert_equal(sad[1][1]['y'].values, [2,4])
def test_two_way_split(self):
# Complete split generates 6 results
sad = emp.sad(self.pat1, self.cols1, 'x:2; y:3')
assert_equal(len(sad), 6)
# Goes through x then y
assert_equal(sad[0][1]['spp'].values, 'a')
assert_equal(sad[0][1]['y'].values, 2)
assert_equal(sad[1][1]['y'].values, [1,1])
assert_equal(sad[5][1]['spp'].values, 'b')
assert_equal(sad[0][1]['y'].values, 2)
def test_simple_with_cols(self):
# Specify count and spp_col here
sad = emp.sad(self.pat1, self.cols1, None)
assert_array_equal(sad[0][1]['y'], [4,4])
def test_simple(self):
# Falling back on spp_col in metadata, so count 1 for each row
sad = emp.sad(self.pat1, None, None)
assert_array_equal(sad[0][1]['y'], [3,2])
def test_split_categorical(self):
sad = emp.sad(self.pat1, self.cols1, 'year:split; x:2')
assert_equal(sad[0][1]['y'].values, 3)
assert_equal(sad[1][1]['y'].values, [])
assert_equal(sad[2][1]['y'].values, [1,1])
assert_equal(sad[3][1]['y'].values, [3])