def test_similarity():
#sign(dist(a,b))==+
pdb_ids = [276, 4629]
active_sites = []
for id in pdb_ids:
filepath = os.path.join("data", "%i.pdb" % id)
active_sites.append(io.read_active_site(filepath))
features = find_features.calc_features(active_sites)
assert scoring.L2(features['276'], features['4629']) > 0
def test_distance_to_self():
#dist(a,a)==0
pdb_ids = [276]
active_sites = []
for id in pdb_ids:
filepath = os.path.join("data", "%i.pdb" % id)
active_sites.append(io.read_active_site(filepath))
features = find_features.calc_features(active_sites)
x = scoring.L2(features['276'], features['276'])
assert x == 0.0
def test_recip_distance():
#dist(a,b)==dist(b,a)
pdb_ids = [276, 4629]
active_sites = []
for id in pdb_ids:
filepath = os.path.join("data", "%i.pdb" % id)
active_sites.append(io.read_active_site(filepath))
features = find_features.calc_features(active_sites)
assert scoring.L2(features['276'],
features['4629']) == scoring.L2(features['4629'],
features['276'])
def test_hierarchical_clustering():
# tractable subset
pdb_ids = [276, 4629, 10701]
active_sites = []
for id in pdb_ids:
filepath = os.path.join("data", "%i.pdb" % id)
active_sites.append(io.read_active_site(filepath))
features = find_features.calc_features(active_sites)
for res, feature_vect in features.items():
hier.hierarchical("V" + res, feature_vect, metric='Euclidian')
assert hier.hierarchical.Cluster()
def test_partition_clustering():
# tractable subset
pdb_ids = [276, 4629, 10701]
active_sites = []
for id in pdb_ids:
filepath = os.path.join("data", "%i.pdb" % id)
active_sites.append(io.read_active_site(filepath))
features = find_features.calc_features(active_sites)
data = []
for res, feature_vect in features.items():
data.append(kmeans.data_point(label=res, data=tuple(feature_vect)))
km = kmeans.kmeans(data=data, k=3, threshold=3)
km.cluster()
# update this assertion
assert len(km.centroids.items()) == 3