def test_data_with_three_dimensions():
json_pathname = os.path.join(fixture_dir, "dataset", "three_dim_v1.json")
dataset = jsonstat.JsonStatDataSet()
dataset.from_file(json_pathname)
data = dataset.data(one="one_1", two="two_1", three="three_1")
assert data.value == 111
assert data.status is None
data = dataset.data(one="one_2", two="two_2", three="three_2")
assert data.value == 222
# using a bit different file
json_pathname = os.path.join(fixture_dir, "dataset", "three_dim_size_as_string_v1.json")
dataset = jsonstat.JsonStatDataSet()
dataset.from_file(json_pathname)
data = dataset.data(one="one_1", two="two_1", three="three_1")
assert data.value == 111
assert data.status is None
data = dataset.data(one="one_2", two="two_2", three="three_2")
assert data.value == 222
# using a bit different file
json_pathname = os.path.join(fixture_dir, "dataset", "three_dim_size_as_string_v2.json")
dataset = jsonstat.JsonStatDataSet()
dataset.from_file(json_pathname)
data = dataset.data(one="one_1", two="two_1", three="three_1")
assert data.value == 111
assert data.status is None
data = dataset.data(one="one_2", two="two_2", three="three_2")
assert data.value == 222
def test_data_with_unemployment():
json_pathname = os.path.join(fixture_dir, "dataset", "dataset_unemployment_v1.json")
dataset = jsonstat.JsonStatDataSet("canada")
dataset.from_file(json_pathname)
data = dataset.data(area="AU", year="2012")
assert data.value == 11
# using label Australia instead of index AU
data = dataset.data(area="Australia", year="2012")
assert data.value == 11
# using dictionary
data = dataset.data({'area': "Australia", 'year': "2012"})
assert data.value == 11
data = dataset.data({'area': "AU", 'year': "2012"})
assert data.value == 11
data = dataset.data({"OECD countries, EU15 and total": "AU", 'year': '2012'})
assert data.value == 11
data = dataset.data(area="BE", year="2014")
assert data.value == 33
assert data.status is None
def test_to_dataframe():
json_pathname = os.path.join(fixture_dir, "dataset",
"dataset_unemployment_v1.json")
dataset = jsonstat.JsonStatDataSet()
dataset.from_file(json_pathname)
df = dataset.to_table(content='id',
blocked_dims={"area": "IT"},
rtype=pd.DataFrame)
# print(df)
ar = [['serie', '2012', 'IT', 14], ['serie', '2013', 'IT', 24],
['serie', '2014', 'IT', 34]]
expected = pd.DataFrame(ar, columns=['serie', 'year', 'area', 'Value'])
pdt.assert_frame_equal(expected, df)
df = dataset.to_data_frame('year',
content='id',
blocked_dims={"area": "IT"})
# TODO: print(df)
ar = [['serie', 'IT', 14], ['serie', 'IT', 24], ['serie', 'IT', 34]]
expected = pd.DataFrame(ar,
columns=['serie', 'area', 'Value'],
index=['2012', '2013', '2014'])
expected.index.name = 'year'
pdt.assert_frame_equal(expected, df)
def test_exception_no_existent_dimension():
json_pathname = os.path.join(fixture_dir, "dataset", "dataset_unemployment_v1.json")
dataset = jsonstat.JsonStatDataSet("canada")
dataset.from_file(json_pathname)
assert dataset.dimension("year").did == "year"
with pytest.raises(jsonstat.JsonStatException) as excinfo:
dataset.dimension("not existent dim")
expected = "dataset 'canada': unknown dimension 'not existent dim' know dimensions ids are: serie, year, area"
assert expected == str(excinfo.value)
def test_all_pos():
json_pathname = os.path.join(fixture_dir, "dataset", "dataset_unemployment_v1.json")
dataset = jsonstat.JsonStatDataSet("canada")
dataset.from_file(json_pathname)
result = list(dataset.all_pos())
# fist digit is serie always 0
# second digit is year from 0 to 2
# third digit is area from 0 to 3
# order is ["serie", "year", "area"]
expected = [[0, 0, 0], [0, 0, 1], [0, 0, 2], [0, 0, 3], # first digit 0
[0, 1, 0], [0, 1, 1], [0, 1, 2], [0, 1, 3], # first digit 1
[0, 2, 0], [0, 2, 1], [0, 2, 2], [0, 2, 3]] # first digit 2
assert result == expected
def test_all_pos_with_block():
json_pathname = os.path.join(fixture_dir, "dataset", "dataset_unemployment_v1.json")
dataset = jsonstat.JsonStatDataSet("canada")
dataset.from_file(json_pathname)
result = list(dataset.all_pos({"area": "IT"}))
expected = [[0, 0, 3], [0, 1, 3], [0, 2, 3]]
assert result == expected
dataset.generate_all_vec(area="IT")
result = list(dataset.all_pos({"year": "2014"}))
expected = [[0, 2, 0], [0, 2, 1], [0, 2, 2], [0, 2, 3]]
assert result == expected
dataset.generate_all_vec(year='2014')
def test_info():
json_pathname = os.path.join(fixture_dir, "dataset", "dataset_unemployment_v1.json")
dataset = jsonstat.JsonStatDataSet("canada")
dataset.from_file(json_pathname)
expected = (
"name: 'canada'\n"
"label: 'Unemployment rate in the OECD countries'\n"
"source: 'Unemployment rate in the OECD countries'\n"
"size: 12\n"
"+-----+-------+--------------------------------+------+------+\n"
"| pos | id | label | size | role |\n"
"+-----+-------+--------------------------------+------+------+\n"
"| 0 | serie | serie | 1 | |\n"
"| 1 | year | 2012-2014 | 3 | time |\n"
"| 2 | area | OECD countries, EU15 and total | 4 | geo |\n"
"+-----+-------+--------------------------------+------+------+"
)
assert expected, dataset.__str__()
def test_to_table():
json_pathname = os.path.join(fixture_dir, "dataset",
"dataset_unemployment_v1.json")
dataset = jsonstat.JsonStatDataSet()
dataset.from_file(json_pathname)
table = dataset.to_table()
# table len is the size of dataset + 1 for headers
assert len(dataset) + 1 == len(table)
header_expected = [
u'serie', u'2012-2014', u'OECD countries, EU15 and total', u'Value'
]
assert header_expected == table[0]
first_row_expected = [u'serie', u'2012', u'Australia', 11]
assert first_row_expected == table[1]
second_row_expected = [u'serie', u'2012', u'Austria', 12]
assert second_row_expected == table[2]
def test_all_pos_with_three_dim():
json_pathname = os.path.join(fixture_dir, "dataset", "three_dim_v1.json")
dataset = jsonstat.JsonStatDataSet()
dataset.from_file(json_pathname)
# test 1
result = list(dataset.all_pos({'one': 'one_1'}))
expected = [
[0, 0, 0], [0, 0, 1], [0, 0, 2], [0, 0, 3],
[0, 1, 0], [0, 1, 1], [0, 1, 2], [0, 1, 3],
[0, 2, 0], [0, 2, 1], [0, 2, 2], [0, 2, 3]]
assert result == expected
# test 2
dataset.generate_all_vec(one='one_1')
result = list(dataset.all_pos({"two": "two_2"}))
expected = [
[0, 1, 0], [0, 1, 1], [0, 1, 2], [0, 1, 3],
[1, 1, 0], [1, 1, 1], [1, 1, 2], [1, 1, 3]
]
assert result == expected
def test_to_data_frame_year_IT():
json_pathname = os.path.join(fixture_dir, "dataset",
"dataset_unemployment_v1.json")
dataset = jsonstat.JsonStatDataSet()
dataset.from_file(json_pathname)
df = dataset.to_data_frame("year",
content="id",
blocked_dims={'area': "IT"})
# print(df)
# serie area Value
# year
# 2012 serie IT 14
# 2013 serie IT 24
# 2014 serie IT 34
# print(df.columns)
# checking taht colums are ['geo','Value]
assert (df.columns == pd.Series(['serie', 'area', 'Value'])).all()
# pdt.assert_series_equal(df.columns,pd.Series(['area', 'Value']))
assert 34 == df.loc['2014']['Value']
def test_to_table_inverted_order():
json_pathname = os.path.join(fixture_dir, "dataset",
"dataset_unemployment_v1.json")
dataset = jsonstat.JsonStatDataSet()
dataset.from_file(json_pathname)
order = [i.did for i in dataset.dimensions()]
order = order[::-1] # reverse list
order = dataset._from_aidx_to_adim(order)
table = dataset.to_table(order=order)
# table len is the size of dataset + 1 for headers
assert len(dataset) + 1 == len(table)
header_expected = [
'serie', '2012-2014', 'OECD countries, EU15 and total', 'Value'
]
assert header_expected == table[0]
first_row_expected = ['serie', '2012', 'Australia', 11]
assert first_row_expected == table[1]
second_row_expected = ['serie', '2013', 'Australia', 21]
assert second_row_expected == table[2]
def test_missing_dimension(json_missing_dimension):
dataset = jsonstat.JsonStatDataSet("canada")
with pytest.raises(jsonstat.JsonStatMalformedJson) as excinfo:
dataset.from_string(json_missing_dimension)
expected = "dataset 'canada': missing 'dimension' key"
assert expected == str(excinfo.value)
def test_empty_value(json_empty_value):
dataset = jsonstat.JsonStatDataSet("canada")
with pytest.raises(jsonstat.JsonStatMalformedJson) as excinfo:
dataset.from_string(json_empty_value)
expected = "dataset 'canada': field 'value' is empty"
assert expected == str(excinfo.value)
def test_exception_not_valid():
dataset = jsonstat.JsonStatDataSet("canada")
with pytest.raises(jsonstat.JsonStatException):
dataset.data(year="2003", area="CA")
def test_exception_dataset_size(json_incorrect_data_size):
dataset = jsonstat.JsonStatDataSet("canada")
with pytest.raises(jsonstat.JsonStatException) as excinfo:
dataset.from_string(json_incorrect_data_size)
expected = "dataset 'canada': size 4 is different from calculate size 48 by dimension"
assert expected == str(excinfo.value)
def test_name():
dataset = jsonstat.JsonStatDataSet("canada")
assert dataset.name == "canada"
def test_dimensions():
json_pathname = os.path.join(fixture_dir, "dataset", "dataset_unemployment_v1.json")
dataset = jsonstat.JsonStatDataSet("canada")
dataset.from_file(json_pathname)
assert len(dataset.dimensions()) == 3
