def test_init():
space = Space('tests/example.n3')
assert_equal(space._path_to_rdf, 'file:tests/example.n3')
assert_equal(space._format, 'ntriples')
assert_equal(space._property,
'http://www.w3.org/2004/02/skos/core#broader')
assert_equal(
space._direct_parents, {
'http://dbpedia.org/resource/Category:Categories_named_after_television_series':
['http://dbpedia.org/resource/Category:Foo'],
'http://dbpedia.org/resource/Category:Star_Trek': [
'http://dbpedia.org/resource/Category:Categories_named_after_television_series',
],
'http://dbpedia.org/resource/Category:Futurama': [
'http://dbpedia.org/resource/Category:Categories_named_after_television_series',
'http://dbpedia.org/resource/Category:New_York_City_in_fiction',
],
})
assert_equal(
space._index, {
'http://dbpedia.org/resource/Category:Categories_named_after_television_series':
0,
'http://dbpedia.org/resource/Category:New_York_City_in_fiction':
1,
'http://dbpedia.org/resource/Category:Foo':
2,
})
assert_equal(space._size, 3)
def test_sum_weighted_uris():
space = Space('tests/example.n3')
s = space.sum_weighted_uris([
('http://dbpedia.org/resource/Category:Futurama', 2),
('http://dbpedia.org/resource/Category:Star_Trek', 1)
])
np.testing.assert_allclose(np.asarray(s.todense()), [[
2 / np.sqrt(2 + 0.9**2) + 1 / np.sqrt(1 + 0.9**2),
2 / np.sqrt(2 + 0.9**2),
2 * 0.9 / np.sqrt(2 + 0.9**2) + 0.9 / np.sqrt(1 + 0.9**2)
]])
def test_parents():
space = Space('tests/example.n3')
assert_equal(
space.parents('http://dbpedia.org/resource/Category:Futurama'), [
('http://dbpedia.org/resource/Category:Categories_named_after_television_series',
1),
('http://dbpedia.org/resource/Category:New_York_City_in_fiction',
1),
('http://dbpedia.org/resource/Category:Foo', 0.9),
])
assert_equal(
space.parents('http://dbpedia.org/resource/Category:Star_Trek'), [
('http://dbpedia.org/resource/Category:Categories_named_after_television_series',
1),
('http://dbpedia.org/resource/Category:Foo', 0.9),
])
assert_equal(space.parents('http://dbpedia.org/resource/Category:Foo'), [])
space = Space('tests/london.n3')
Space.max_depth = 0
assert_equal(space.parents(
'http://dbpedia.org/resource/Category:London'
), [(
'http://dbpedia.org/resource/Category:NUTS_1_statistical_regions_of_England',
1), ('http://dbpedia.org/resource/Category:M4_corridor', 1)])
Space.max_depth = 1
print space.parents('http://dbpedia.org/resource/Category:London')
assert_equal(space.parents(
'http://dbpedia.org/resource/Category:London'
), [('http://dbpedia.org/resource/Category:Regional_planning_in_England',
0.9),
('http://dbpedia.org/resource/Category:Regions_of_England', 0.9),
('http://dbpedia.org/resource/Category:NUTS_1_statistical_regions_of_the_European_Union',
0.9), ('http://dbpedia.org/resource/Category:England', 0.9),
('http://dbpedia.org/resource/Category:NUTS_1_statistical_regions_of_the_United_Kingdom',
0.9), ('http://dbpedia.org/resource/Category:Regions_of_Wales', 0.9),
('http://dbpedia.org/resource/Category:NUTS_1_statistical_regions_of_England',
1),
('http://dbpedia.org/resource/Category:Local_government_in_England',
0.9), ('http://dbpedia.org/resource/Category:M4_corridor', 1)])
Space.max_depth = 10
def test_centroid_weighted_uris():
space = Space('tests/example.n3')
centroid = space.centroid_weighted_uris([
('http://dbpedia.org/resource/Category:Futurama', 2),
('http://dbpedia.org/resource/Category:Star_Trek', 1)
])
np.testing.assert_allclose(
centroid.todense(),
np.array([[
(2 / np.sqrt(2 + 0.9**2) + 1 / np.sqrt(1 + 0.9**2)) / 2,
(1 / np.sqrt(2 + 0.9**2)),
(2 * 0.9 / np.sqrt(2 + 0.9**2) + 0.9 / np.sqrt(1 + 0.9**2)) / 2
]]))
def test_cache_vectors():
space = Space('tests/example.n3')
space.cache_vectors()
assert space._uri_to_vector.has_key(
'http://dbpedia.org/resource/Category:Futurama'
) # Checking that we cached the vectors
assert space._uri_to_vector.has_key(
'http://dbpedia.org/resource/Category:Star_Trek')
np.testing.assert_array_equal(
space._uri_to_vector['http://dbpedia.org/resource/Category:Futurama'],
space.to_vector('http://dbpedia.org/resource/Category:Futurama'))
np.testing.assert_array_equal(
space._uri_to_vector['http://dbpedia.org/resource/Category:Star_Trek'],
space.to_vector('http://dbpedia.org/resource/Category:Star_Trek'))
def test_similarity_all():
space = Space('tests/example.n3')
m = lil_matrix((2, 3))
m[0, 0] = 1 / np.sqrt(1 + 2 * 2 + 3 * 3)
m[0, 1] = 2 / np.sqrt(1 + 2 * 2 + 3 * 3)
m[0, 2] = 3 / np.sqrt(1 + 2 * 2 + 3 * 3)
m[1, 0] = 4 / np.sqrt(4 * 4 + 5 * 5 + 6 * 6)
m[1, 1] = 5 / np.sqrt(4 * 4 + 5 * 5 + 6 * 6)
m[1, 2] = 6 / np.sqrt(4 * 4 + 5 * 5 + 6 * 6)
v = m[0, :]
m = m.tocsr()
similarities = space.similarity_all(m, v)
assert_equal(similarities[0], 1)
assert_equal(
similarities[1],
((1 * 4 + 2 * 5 + 3 * 6) /
(np.sqrt(1 + 2 * 2 + 3 * 3) * np.sqrt(4 * 4 + 5 * 5 + 6 * 6))))
def test_to_vector():
space = Space('tests/example.n3')
np.testing.assert_array_equal(
space.to_vector(
'http://dbpedia.org/resource/Category:Futurama').todense(), [[
1 / np.sqrt(2 + 0.9**2), 1 / np.sqrt(2 + 0.9**2),
0.9 / np.sqrt(2 + 0.9**2)
]])
np.testing.assert_array_equal(
space.to_vector(
'http://dbpedia.org/resource/Category:Star_Trek').todense(),
[[1 / np.sqrt(1 + 0.9**2), 0, 0.9 / np.sqrt(1 + 0.9**2)]])
assert space._uri_to_vector.has_key(
'http://dbpedia.org/resource/Category:Futurama'
) # Checking that we cached the vectors when generating them
assert space._uri_to_vector.has_key(
'http://dbpedia.org/resource/Category:Star_Trek')
np.testing.assert_array_equal(
space._uri_to_vector['http://dbpedia.org/resource/Category:Futurama'],
space.to_vector('http://dbpedia.org/resource/Category:Futurama'))
np.testing.assert_array_equal(
space._uri_to_vector['http://dbpedia.org/resource/Category:Star_Trek'],
space.to_vector('http://dbpedia.org/resource/Category:Star_Trek'))
import rdfsim
from rdfsim.space import Space
space = Space('skos_categories_1000.nt',
property='http://www.w3.org/2004/02/skos/core#broader')
print "Distance betwen Dwarves and Elves:"
print space.similarity_uri(
'http://dbpedia.org/resource/Category:Middle-earth_Dwarves',
'http://dbpedia.org/resource/Category:Middle-earth_Elves')
print "Distance between Dwarves and Futurama:"
print space.similarity_uri(
'http://dbpedia.org/resource/Category:Middle-earth_Dwarves',
'http://dbpedia.org/resource/Category:Futurama')
def test_similarity_uri():
space = Space('tests/example.n3')
np.testing.assert_allclose(
space.similarity_uri('http://dbpedia.org/resource/Category:Futurama',
'http://dbpedia.org/resource/Category:Star_Trek'),
(1 + 0.9 * 0.9) / (np.sqrt(2 + 0.9**2) * np.sqrt(1 + 0.9**2)))