def test_head(self):
data = LazyResult(np.array([1, 2, 3]), np.dtype(np.int64), 1)
series = Series(data.expr, np.dtype(np.int64), RangeIndex(0, 2, 1))
expected_result = np.array([1, 2])
result = series.head(2)
np.testing.assert_array_equal(expected_result, result)
def test_std(self):
data = np.array([1, 2, 3, 4])
series = Series(data, np.dtype(np.int64), RangeIndex(0, 4, 1))
expected_result = 1.2909944487358056
result = series.std()
np.testing.assert_equal(expected_result, evaluate_if_necessary(result))
def test_count(self):
data = np.array([1, 2, 3])
series = Series(data, np.dtype(np.int64), RangeIndex(0, 3, 1))
expected_result = 3
result = series.count()
np.testing.assert_equal(expected_result, evaluate_if_necessary(result))
def test_mean(self):
data = np.array([1, 2, 3, 4])
series = Series(data, np.dtype(np.int64), RangeIndex(0, 4, 1))
expected_result = 2.5
result = series.mean()
np.testing.assert_equal(expected_result, evaluate_if_necessary(result))
def test_describe(self):
data = np.array([1, 2, 3])
series = Series(data, np.dtype(np.int64), RangeIndex(0, 3, 1))
expected_result = np.array([1, 3, 2, 1], np.float64)
result = series.describe(['min', 'max', 'mean', 'std'])
np.testing.assert_equal(expected_result, evaluate_if_necessary(result))
def test_comparison(self):
data = np.array([1, 2, 3, 4])
series = Series(data, np.dtype(np.int64), RangeIndex(0, 4, 1))
expected_result = Series(np.array([True, True, False, False]),
np.dtype(np.bool), RangeIndex(0, 4, 1))
result = series < 3
test_equal_series(expected_result, result)
def test_getitem_series(self):
data = np.array([1, 2, 3])
series = Series(data, np.dtype(np.int64), RangeIndex(0, 3, 1))
expected_result = Series(np.array([1, 3]), np.dtype(np.int64),
Index(np.array([0, 2]), np.dtype(np.int64)))
result = series[series != 2]
test_equal_series(expected_result, result)
def test_element_wise_operation(self):
data = np.array([1, 2, 3])
series = Series(data, np.dtype(np.int64), RangeIndex(0, 3, 1))
expected_result = Series(np.array([2, 4, 6]), np.dtype(np.int64),
RangeIndex(0, 3, 1))
result = series * 2
test_equal_series(expected_result, result)
def test_getitem_slice(self):
weld_type = numpy_to_weld_type('int64')
data = LazyResult(np.array([1, 2, 3]), weld_type, 1)
series = Series(data.expr, np.dtype(np.int64), RangeIndex(0, 3, 1))
expected_result = Series(np.array([1, 2]), np.dtype(np.int64),
RangeIndex(0, 2, 1))
result = series[:2]
test_equal_series(expected_result, result)
def test_agg(self):
data = np.array([1, 2, 3, 4])
series = Series(data, np.dtype(np.int64), RangeIndex(0, 4, 1))
expected_result = Series(
np.array([1., 4.], dtype=np.float64), np.dtype('float64'),
Index(np.array(['min', 'max'], dtype=np.str), np.dtype(np.str)))
result = series.agg(['min', 'max'])
test_equal_series(expected_result, result)
def test_array_operation(self):
data = np.array([1, 2, 3])
series = Series(data, np.dtype(np.int64), RangeIndex(0, 3, 1))
expected_result = Series(np.array([3, 5, 7]), np.dtype(np.int64),
RangeIndex(0, 3, 1))
result = series + Series(np.array([2, 3, 4]), np.dtype(np.int64),
RangeIndex(0, 3, 1))
test_equal_series(expected_result, result)
def test_map_weld_code(self):
series = Series(np.array([1, 3, 4]), np.dtype(np.int64),
RangeIndex(0, 3, 1))
weld_template = "map(%(self)s, |e| e + %(scalar)s)"
mapping = {'scalar': '2L'}
result = series.map(weld_template, mapping)
expected_result = Series(np.array([3, 5, 6]), np.dtype(np.int64),
RangeIndex(0, 3, 1))
test_equal_series(expected_result, result)
def test_bitwise_and(self):
data = np.array([True, True, False, False])
series = Series(data, np.dtype(np.bool), RangeIndex(0, 4, 1))
data_other = np.array([True, False, True, False])
series_other = Series(data_other, np.dtype(np.bool),
RangeIndex(0, 4, 1))
expected_result = Series(np.array([True, False, False, False]),
np.dtype(np.bool), RangeIndex(0, 4, 1))
result = series & series_other
test_equal_series(expected_result, result)
def test_map_weld_cudf(self):
import os
WeldObject.load_binary(os.path.dirname(__file__) + '/cudf/udf_c.so')
series = Series(np.array([1, 3, 4]), np.dtype(np.int64),
RangeIndex(0, 3, 1))
weld_template = "cudf[udf_add, vec[i64]](%(self)s, %(scalar)s)"
mapping = {'scalar': '2L'}
result = series.map(weld_template, mapping)
expected_result = Series(np.array([3, 5, 6]), np.dtype(np.int64),
RangeIndex(0, 3, 1))
test_equal_series(expected_result, result)
def _aggregate(self, operation):
aggregated_data = weld_groupby_aggregate(
self.expr, [str(numpy_to_weld_type(k)) for k in self.by_types],
[str(numpy_to_weld_type(k))
for k in self.columns_types], operation)
if len(self.by) == 1:
new_index = Index(weld_get_column(aggregated_data, 0, True),
self.by_types[0], self.by[0])
else:
arrays = [
LazyResult(weld_get_column(aggregated_data, index, True),
numpy_to_weld_type(self.by_types[index]), 1)
for index in xrange(len(self.by))
]
new_index = MultiIndex.from_arrays(arrays, self.by)
new_data = OrderedDict()
for i in xrange(len(self.columns)):
column_name = self.columns[i]
new_data[column_name] = Series(
weld_get_column(aggregated_data, i, False),
self.columns_types[i], new_index, column_name)
return DataFrame(new_data, new_index)