def query_for_split_decision(self, inds1, inds2, vectorizer, vectors, vector_names):
print display_vector_index_details(inds1, vectors, vector_names, vectorizer)
print display_vector_index_details(inds2, vectors, vector_names, vectorizer)
print "Overlap:"
print display_shared_vector_indeces_details([inds1, inds2], vectors, vector_names, vectorizer)
response = raw_input("Should {0} and {1} be in the same cluster? (Y/N)".format(vector_names[inds1], vector_names[inds2]))
if response.upper() == "Y" or response.upper() == "YES":
return False # If users think the two should be in one cluster, we don't split, so return False
if response.upper() == "N" or response.upper() == "NO":
return True # vice versa
def select_next_cluster():
cluster = _non_fixed_clusters[-1]
set_cluster_statistics(cluster)
size = np.ones(n_samples) * 15
size[cluster.index] = 28
alpha = np.ones(n_samples) * 0.6
alpha[cluster.index] = 0.85
cluster_commonality.text = generate_display_string(display_shared_vector_indeces_details(
cluster.index,
clusteringnmap.interactive_state._vectors,
clusteringnmap.interactive_state._vector_names,
clusteringnmap.interactive_state._vectorizer
).replace("\n", "<br/>"))
source.data["size"] = size
source.data["alpha"] = alpha
def select_next_cluster():
cluster = _non_fixed_clusters[-1]
set_cluster_statistics(cluster)
size = np.ones(n_samples) * 15
size[cluster.index] = 28
alpha = np.ones(n_samples) * 0.6
alpha[cluster.index] = 0.85
cluster_commonality.text = generate_display_string(
display_shared_vector_indeces_details(
cluster.index, clusteringnmap.interactive_state._vectors,
clusteringnmap.interactive_state._vector_names,
clusteringnmap.interactive_state._vectorizer).replace(
"\n", "<br/>"))
source.data["size"] = size
source.data["alpha"] = alpha
