def test1_smallset(self):
path_input = "test/data/schema_matching/default_matches_cities_input.csv"
df = pd.read_csv(path_input)
pairs_relational = relational_matching(df)[["uri_1","uri_2"]].sort_values(by=["uri_1","uri_2"]).reset_index(drop=True)
pairs_string = string_similarity_matching(df)[["uri_1","uri_2"]].sort_values(by=["uri_1","uri_2"]).reset_index(drop=True)
pairs_schema = label_schema_matching(df)[["uri_1","uri_2"]].sort_values(by=["uri_1","uri_2"]).reset_index(drop=True)
pairs_overlap = value_overlap_matching(df)[["uri_1","uri_2"]].sort_values(by=["uri_1","uri_2"]).reset_index(drop=True)
assert all([pairs_relational.equals(x) for x in [pairs_string, pairs_schema, pairs_overlap]])
def test3_numeric_data(self):
path_input = "test/data/schema_matching/value_matches_cities_numeric_input.csv"
df = pd.read_csv(path_input)
path_expected = "test/data/schema_matching/value_matches_cities_numeric_expected.csv"
expected_matches = pd.read_csv(path_expected)
output_matches = value_overlap_matching(df)
output_matches['value_overlap'] = pd.to_numeric(output_matches['value_overlap'])
pd.testing.assert_frame_equal(
output_matches, expected_matches, check_like=True)
def test4_no_matches_numeric_data(self):
df = pd.DataFrame({
'city' : [1, 1, 0, 1],
'entity' : ['Bremen', 'Hamburg', 'Denmark', 'Berlin'],
'new_link_1': ['http://dbpedia.org/resource/Bremen', 'http://dbpedia.org/resource/Hamburg', 'http://dbpedia.org/resource/Denmark', 'http://dbpedia.org/resource/Berlin'],
'Link_Out_numeric_http://dbpedia.org/ontology/PopulatedPlace/areaMetro': [1, 0, 0, 0],
'Link_Out_numeric_http://dbpedia.org/ontology/abstract': [12, 12, 11, 12]
})
expected_result_df = pd.DataFrame({
'uri_1' : ['http://dbpedia.org/ontology/PopulatedPlace/areaMetro'],
'uri_2' : ['http://dbpedia.org/ontology/abstract'],
'value_overlap': [0.0]
})
result = value_overlap_matching(df)
pd.testing.assert_frame_equal(
result, expected_result_df, check_like=True)
def test2_no_matches_boolean_data(self):
df = pd.DataFrame({
'city' : [1, 1, 0, 1],
'entity' : ['Bremen', 'Hamburg', 'Denmark', 'Berlin'],
'new_link_1': ['http://dbpedia.org/resource/Bremen', 'http://dbpedia.org/resource/Hamburg', 'http://dbpedia.org/resource/Denmark', 'http://dbpedia.org/resource/Berlin'],
'new_link_in_boolean_http://dbpedia.org/resource/Category:German_state_capitals': [True, True, False, True],
'new_link_in_boolean_http://dbpedia.org/resource/Category:Countries_in_Europe': [False, False, True, False]
})
expected_result_df = pd.DataFrame({
'uri_1' : ['new_link_in_boolean_http://dbpedia.org/resource/Category:Countries_in_Europe'],
'uri_2' : ['http://dbpedia.org/resource/Category:German_state_capitals'],
'value_overlap': [0.0]
})
result = value_overlap_matching(df)
pd.testing.assert_frame_equal(
result, expected_result_df, check_like=True)
def test2_bigset(self):
#WARNING: Takes long to run!
path_input = "test/data/schema_matching/pair_equality_test2_bigset.csv"
df = pd.read_csv(path_input)
pairs_relational = relational_matching(df)[["uri_1","uri_2"]].sort_values(by=["uri_1","uri_2"]).reset_index(drop=True)
pairs_string = string_similarity_matching(df)[["uri_1","uri_2"]].sort_values(by=["uri_1","uri_2"]).reset_index(drop=True)
pairs_schema = label_schema_matching(df)[["uri_1","uri_2"]].sort_values(by=["uri_1","uri_2"]).reset_index(drop=True)
pairs_overlap = value_overlap_matching(df)[["uri_1","uri_2"]].sort_values(by=["uri_1","uri_2"]).reset_index(drop=True)
assert all([pairs_relational.equals(x) for x in [pairs_string, pairs_schema, pairs_overlap]])