def test_normalize_ww_init():
es = EntitySet()
df = pd.DataFrame({
'id': [1, 2, 3, 4],
'col': ['a', 'b', 'c', 'd'],
'df2_id': [1, 1, 2, 2],
'df2_col': [True, False, True, True]
})
df.ww.init(index='id', name='test_name')
es.add_dataframe(dataframe=df)
assert es['test_name'].ww.name == 'test_name'
assert es['test_name'].ww.schema.name == 'test_name'
es.normalize_dataframe('test_name',
'new_df',
'df2_id',
additional_columns=['df2_col'])
assert es['test_name'].ww.name == 'test_name'
assert es['test_name'].ww.schema.name == 'test_name'
assert es['new_df'].ww.name == 'new_df'
assert es['new_df'].ww.schema.name == 'new_df'
def test_normalize_ww_init():
es = EntitySet()
df = pd.DataFrame({
"id": [1, 2, 3, 4],
"col": ["a", "b", "c", "d"],
"df2_id": [1, 1, 2, 2],
"df2_col": [True, False, True, True],
})
df.ww.init(index="id", name="test_name")
es.add_dataframe(dataframe=df)
assert es["test_name"].ww.name == "test_name"
assert es["test_name"].ww.schema.name == "test_name"
es.normalize_dataframe("test_name",
"new_df",
"df2_id",
additional_columns=["df2_col"])
assert es["test_name"].ww.name == "test_name"
assert es["test_name"].ww.schema.name == "test_name"
assert es["new_df"].ww.name == "new_df"
assert es["new_df"].ww.schema.name == "new_df"
def test_extra_woodwork_params(es):
new_es = EntitySet()
sessions_df = es["sessions"].ww.copy()
assert sessions_df.ww.index == "id"
assert sessions_df.ww.time_index is None
assert isinstance(sessions_df.ww.logical_types["id"], Integer)
warning_msg = (
"A Woodwork-initialized DataFrame was provided, so the following parameters were ignored: "
"index, time_index, logical_types, make_index, semantic_tags, already_sorted"
)
with pytest.warns(UserWarning, match=warning_msg):
new_es.add_dataframe(
dataframe_name="sessions",
dataframe=sessions_df,
index="filepath",
time_index="customer_id",
logical_types={"id": Categorical},
make_index=True,
already_sorted=True,
semantic_tags={"id": "new_tag"},
)
assert sessions_df.ww.index == "id"
assert sessions_df.ww.time_index is None
assert isinstance(sessions_df.ww.logical_types["id"], Integer)
assert "new_tag" not in sessions_df.ww.semantic_tags
def test_replace_dataframe():
df = pd.DataFrame({
'id':
range(4),
'full_name':
['Mr. John Doe', 'Doe, Mrs. Jane', 'James Brown', 'Ms. Paige Turner'],
'email': [
'*****@*****.**', np.nan, '*****@*****.**',
'*****@*****.**'
],
'phone_number':
['5555555555', '555-555-5555', '1-(555)-555-5555', '555-555-5555'],
'age':
pd.Series([33, None, 33, 57], dtype='Int64'),
'signup_date': [pd.to_datetime('2020-09-01')] * 4,
'is_registered':
pd.Series([True, False, True, None], dtype='boolean'),
})
df.ww.init(name='table', index='id')
es = EntitySet('es')
es.add_dataframe(df)
original_schema = es['table'].ww.schema
new_df = df.iloc[2:]
es.replace_dataframe('table', new_df)
assert len(es['table']) == 2
assert es['table'].ww.schema == original_schema
def test_extra_woodwork_params(es):
new_es = EntitySet()
sessions_df = es['sessions'].ww.copy()
assert sessions_df.ww.index == 'id'
assert sessions_df.ww.time_index is None
assert isinstance(sessions_df.ww.logical_types['id'], Integer)
warning_msg = (
'A Woodwork-initialized DataFrame was provided, so the following parameters were ignored: '
'index, time_index, logical_types, make_index, semantic_tags, already_sorted'
)
with pytest.warns(UserWarning, match=warning_msg):
new_es.add_dataframe(dataframe_name='sessions',
dataframe=sessions_df,
index='filepath',
time_index='customer_id',
logical_types={'id': Categorical},
make_index=True,
already_sorted=True,
semantic_tags={'id': 'new_tag'})
assert sessions_df.ww.index == 'id'
assert sessions_df.ww.time_index is None
assert isinstance(sessions_df.ww.logical_types['id'], Integer)
assert 'new_tag' not in sessions_df.ww.semantic_tags
def test_replace_dataframe_data_transformation(latlong_df):
initial_df = latlong_df.copy()
initial_df.ww.init(
name='latlongs',
index='string_tuple',
logical_types={col_name: 'LatLong'
for col_name in initial_df.columns})
es = EntitySet()
es.add_dataframe(dataframe=initial_df)
df = to_pandas(es['latlongs'])
expected_val = (1, 2)
if ks and isinstance(es['latlongs'], ks.DataFrame):
expected_val = [1, 2]
for col in latlong_df.columns:
series = df[col]
assert series.iloc[0] == expected_val
es.replace_dataframe('latlongs', latlong_df)
df = to_pandas(es['latlongs'])
expected_val = (3, 4)
if ks and isinstance(es['latlongs'], ks.DataFrame):
expected_val = [3, 4]
for col in latlong_df.columns:
series = df[col]
assert series.iloc[-1] == expected_val
def test_replace_dataframe_different_dataframe_types():
dask_es = EntitySet(id="dask_es")
sessions = pd.DataFrame({
"id": [0, 1, 2, 3],
"user": [1, 2, 1, 3],
"time": [
pd.to_datetime('2019-01-10'),
pd.to_datetime('2019-02-03'),
pd.to_datetime('2019-01-01'),
pd.to_datetime('2017-08-25')
],
"strings": ["I am a string", "23", "abcdef ghijk", ""]
})
sessions_dask = dd.from_pandas(sessions, npartitions=2)
sessions_logical_types = {
"id": Integer,
"user": Integer,
"time": Datetime,
"strings": NaturalLanguage
}
sessions_semantic_tags = {'user': '******'}
dask_es.add_dataframe(dataframe_name="sessions",
dataframe=sessions_dask,
index="id",
time_index="time",
logical_types=sessions_logical_types,
semantic_tags=sessions_semantic_tags)
with pytest.raises(TypeError, match='Incorrect DataFrame type used'):
dask_es.replace_dataframe('sessions', sessions)
def test_normalize_dataframe():
df = pd.DataFrame({
'id':
range(4),
'full_name':
['Mr. John Doe', 'Doe, Mrs. Jane', 'James Brown', 'Ms. Paige Turner'],
'email': [
'*****@*****.**', np.nan, '*****@*****.**',
'*****@*****.**'
],
'phone_number':
['5555555555', '555-555-5555', '1-(555)-555-5555', '555-555-5555'],
'age':
pd.Series([33, None, 33, 57], dtype='Int64'),
'signup_date': [pd.to_datetime('2020-09-01')] * 4,
'is_registered':
pd.Series([True, False, True, None], dtype='boolean'),
})
df.ww.init(name='first_table', index='id', time_index='signup_date')
es = EntitySet('es')
es.add_dataframe(df)
es.normalize_dataframe('first_table',
'second_table',
'age',
additional_columns=['phone_number', 'full_name'],
make_time_index=True)
assert len(es.dataframe_dict) == 2
assert 'foreign_key' in es['first_table'].ww.semantic_tags['age']
def test_dataframe_without_name(es):
new_es = EntitySet()
new_df = es['sessions'].copy()
assert new_df.ww.schema is None
error = 'Cannot add dataframe to EntitySet without a name. Please provide a value for the dataframe_name parameter.'
with pytest.raises(ValueError, match=error):
new_es.add_dataframe(new_df)
def test_woodwork_dataframe_without_name_errors(es):
new_es = EntitySet()
new_df = es['sessions'].ww.copy()
new_df.ww._schema.name = None
assert new_df.ww.name is None
error = 'Cannot add a Woodwork DataFrame to EntitySet without a name'
with pytest.raises(ValueError, match=error):
new_es.add_dataframe(new_df)
def test_woodwork_dataframe_same_name_parameter(es):
new_es = EntitySet()
new_df = es["sessions"].ww.copy()
new_df.ww._schema.name = "df_name"
assert new_df.ww.name == "df_name"
new_es.add_dataframe(new_df, dataframe_name="df_name")
assert new_es["df_name"].ww.name == "df_name"
def test_woodwork_dataframe_same_name_parameter(es):
new_es = EntitySet()
new_df = es['sessions'].ww.copy()
new_df.ww._schema.name = 'df_name'
assert new_df.ww.name == 'df_name'
new_es.add_dataframe(new_df, dataframe_name='df_name')
assert new_es['df_name'].ww.name == 'df_name'
def test_dataframe_with_name_parameter(es):
new_es = EntitySet()
new_df = es['sessions'][['id']]
assert new_df.ww.schema is None
new_es.add_dataframe(new_df,
dataframe_name='df_name',
index='id',
logical_types={'id': 'Integer'})
assert new_es['df_name'].ww.name == 'df_name'
def test_add_secondary_time_index(dates_df):
dates_df.ww.init(name='dates_table',
index='backwards_order',
time_index='dates_backwards')
es = EntitySet('es')
es.add_dataframe(
dates_df,
secondary_time_index={'repeating_dates': ['random_order', 'special']})
assert dates_df.ww.metadata['secondary_time_index'] == {
'repeating_dates': ['random_order', 'special', 'repeating_dates']
}
def test_dataframe_with_name_parameter(es):
new_es = EntitySet()
new_df = es["sessions"][["id"]]
assert new_df.ww.schema is None
new_es.add_dataframe(new_df,
dataframe_name="df_name",
index="id",
logical_types={"id": "Integer"})
assert new_es["df_name"].ww.name == "df_name"
def test_add_secondary_time_index(dates_df):
dates_df.ww.init(name="dates_table",
index="backwards_order",
time_index="dates_backwards")
es = EntitySet("es")
es.add_dataframe(
dates_df,
secondary_time_index={"repeating_dates": ["random_order", "special"]})
assert dates_df.ww.metadata["secondary_time_index"] == {
"repeating_dates": ["random_order", "special", "repeating_dates"]
}
def test_woodwork_dataframe_ignore_conflicting_name_parameter_warning(es):
new_es = EntitySet()
new_df = es['sessions'].ww.copy()
new_df.ww._schema.name = 'df_name'
assert new_df.ww.name == 'df_name'
warning = 'A Woodwork-initialized DataFrame was provided, so the following parameters were ignored: dataframe_name'
with pytest.warns(UserWarning, match=warning):
new_es.add_dataframe(new_df, dataframe_name='conflicting_name')
assert new_es['df_name'].ww.name == 'df_name'
def test_time_type_check_order(dates_df):
dates_df.ww.init(name='dates_table',
index='backwards_order',
time_index='random_order')
es = EntitySet('es')
error = 'dates_table time index is Datetime type which differs from other entityset time indexes'
with pytest.raises(TypeError, match=error):
es.add_dataframe(dates_df,
secondary_time_index={
'repeating_dates': ['random_order', 'special']
})
assert 'secondary_time_index' not in dates_df.ww.metadata
def test_init_with_mismatched_time_types(dates_df):
dates_df.ww.init(name='dates_table',
index='backwards_order',
time_index='repeating_dates')
es = EntitySet('es')
es.add_dataframe(dates_df,
secondary_time_index={'special_dates': ['special']})
assert es.time_type == Datetime
nums_df = pd.DataFrame({'id': [1, 2, 3], 'times': [9, 8, 7]})
nums_df.ww.init(name='numerics_table', index='id', time_index='times')
error = 'numerics_table time index is numeric type which differs from other entityset time indexes'
with pytest.raises(TypeError, match=error):
es.add_dataframe(nums_df)
def test_init_with_mismatched_time_types(dates_df):
dates_df.ww.init(name="dates_table",
index="backwards_order",
time_index="repeating_dates")
es = EntitySet("es")
es.add_dataframe(dates_df,
secondary_time_index={"special_dates": ["special"]})
assert es.time_type == Datetime
nums_df = pd.DataFrame({"id": [1, 2, 3], "times": [9, 8, 7]})
nums_df.ww.init(name="numerics_table", index="id", time_index="times")
error = "numerics_table time index is numeric type which differs from other entityset time indexes"
with pytest.raises(TypeError, match=error):
es.add_dataframe(nums_df)
def test_time_type_check_order(dates_df):
dates_df.ww.init(name="dates_table",
index="backwards_order",
time_index="random_order")
es = EntitySet("es")
error = "dates_table time index is Datetime type which differs from other entityset time indexes"
with pytest.raises(TypeError, match=error):
es.add_dataframe(
dates_df,
secondary_time_index={
"repeating_dates": ["random_order", "special"]
},
)
assert "secondary_time_index" not in dates_df.ww.metadata
def test_add_dataframe_to_es(df):
es1 = EntitySet('es')
assert es1.dataframe_dict == {}
es1.add_dataframe(df,
dataframe_name='table',
index='id',
semantic_tags={'category': 'new_tag'})
assert len(es1.dataframe_dict) == 1
copy_df = df.ww.copy()
es2 = EntitySet('es')
assert es2.dataframe_dict == {}
es2.add_dataframe(copy_df)
assert len(es2.dataframe_dict) == 1
assert es1['table'].ww == es2['table'].ww
def test_add_dataframe_to_es(df):
es1 = EntitySet("es")
assert es1.dataframe_dict == {}
es1.add_dataframe(df,
dataframe_name="table",
index="id",
semantic_tags={"category": "new_tag"})
assert len(es1.dataframe_dict) == 1
copy_df = df.ww.copy()
es2 = EntitySet("es")
assert es2.dataframe_dict == {}
es2.add_dataframe(copy_df)
assert len(es2.dataframe_dict) == 1
assert es1["table"].ww == es2["table"].ww
def test_int_double_time_type(dates_df):
dates_df.ww.init(name='dates_table',
index='backwards_order',
time_index='random_order',
logical_types={
'random_order': 'Integer',
'special': 'Double'
})
es = EntitySet('es')
# Both random_order and special are numeric, but they are different logical types
es.add_dataframe(dates_df,
secondary_time_index={'special': ['dates_backwards']})
assert isinstance(es['dates_table'].ww.logical_types['random_order'],
Integer)
assert isinstance(es['dates_table'].ww.logical_types['special'], Double)
assert es['dates_table'].ww.time_index == 'random_order'
assert 'special' in es['dates_table'].ww.metadata['secondary_time_index']
def test_normalize_dataframe():
df = pd.DataFrame({
"id":
range(4),
"full_name": [
"Mr. John Doe",
"Doe, Mrs. Jane",
"James Brown",
"Ms. Paige Turner",
],
"email": [
"*****@*****.**",
np.nan,
"*****@*****.**",
"*****@*****.**",
],
"phone_number": [
"5555555555",
"555-555-5555",
"1-(555)-555-5555",
"555-555-5555",
],
"age":
pd.Series([33, None, 33, 57], dtype="Int64"),
"signup_date": [pd.to_datetime("2020-09-01")] * 4,
"is_registered":
pd.Series([True, False, True, None], dtype="boolean"),
})
df.ww.init(name="first_table", index="id", time_index="signup_date")
es = EntitySet("es")
es.add_dataframe(df)
es.normalize_dataframe(
"first_table",
"second_table",
"age",
additional_columns=["phone_number", "full_name"],
make_time_index=True,
)
assert len(es.dataframe_dict) == 2
assert "foreign_key" in es["first_table"].ww.semantic_tags["age"]
def test_int_double_time_type(dates_df):
dates_df.ww.init(
name="dates_table",
index="backwards_order",
time_index="random_order",
logical_types={
"random_order": "Integer",
"special": "Double"
},
)
es = EntitySet("es")
# Both random_order and special are numeric, but they are different logical types
es.add_dataframe(dates_df,
secondary_time_index={"special": ["dates_backwards"]})
assert isinstance(es["dates_table"].ww.logical_types["random_order"],
Integer)
assert isinstance(es["dates_table"].ww.logical_types["special"], Double)
assert es["dates_table"].ww.time_index == "random_order"
assert "special" in es["dates_table"].ww.metadata["secondary_time_index"]
def test_replace_dataframe():
df = pd.DataFrame({
"id":
range(4),
"full_name": [
"Mr. John Doe",
"Doe, Mrs. Jane",
"James Brown",
"Ms. Paige Turner",
],
"email": [
"*****@*****.**",
np.nan,
"*****@*****.**",
"*****@*****.**",
],
"phone_number": [
"5555555555",
"555-555-5555",
"1-(555)-555-5555",
"555-555-5555",
],
"age":
pd.Series([33, None, 33, 57], dtype="Int64"),
"signup_date": [pd.to_datetime("2020-09-01")] * 4,
"is_registered":
pd.Series([True, False, True, None], dtype="boolean"),
})
df.ww.init(name="table", index="id")
es = EntitySet("es")
es.add_dataframe(df)
original_schema = es["table"].ww.schema
new_df = df.iloc[2:]
es.replace_dataframe("table", new_df)
assert len(es["table"]) == 2
assert es["table"].ww.schema == original_schema
def test_add_time_index_through_woodwork_different_type(dates_df):
dates_df.ww.init(name='dates_table',
index='backwards_order',
time_index='dates_backwards')
es = EntitySet('es')
es.add_dataframe(
dates_df,
secondary_time_index={'repeating_dates': ['random_order', 'special']})
assert dates_df.ww.metadata['secondary_time_index'] == {
'repeating_dates': ['random_order', 'special', 'repeating_dates']
}
assert es.time_type == Datetime
assert es._check_uniform_time_index(es['dates_table']) is None
dates_df.ww.set_time_index('random_order')
assert dates_df.ww.time_index == 'random_order'
error = 'dates_table time index is numeric type which differs from other entityset time indexes'
with pytest.raises(TypeError, match=error):
es._check_uniform_time_index(es['dates_table'])
def test_add_time_index_through_woodwork_different_type(dates_df):
dates_df.ww.init(name="dates_table",
index="backwards_order",
time_index="dates_backwards")
es = EntitySet("es")
es.add_dataframe(
dates_df,
secondary_time_index={"repeating_dates": ["random_order", "special"]})
assert dates_df.ww.metadata["secondary_time_index"] == {
"repeating_dates": ["random_order", "special", "repeating_dates"]
}
assert es.time_type == Datetime
assert es._check_uniform_time_index(es["dates_table"]) is None
dates_df.ww.set_time_index("random_order")
assert dates_df.ww.time_index == "random_order"
error = "dates_table time index is numeric type which differs from other entityset time indexes"
with pytest.raises(TypeError, match=error):
es._check_uniform_time_index(es["dates_table"])