# along with GraphBrain. If not, see <http://www.gnu.org/licenses/>.
import operator
import numpy as np
from sklearn.cluster import DBSCAN
import gb.tools.json as json_tools
import gb.hypergraph.edge as ed
import gb.nlp.parser as par
from gb.explore.similarity import edge_similarity
if __name__ == '__main__':
print('creating parser...')
par = par.Parser()
print('parser created.')
edge_data = json_tools.read('edges_similar_concepts.json')
extra_edges = {}
for item in edge_data:
edge = ed.str2edge(item['edge'])
matched = [ed.str2edge(match[1]) for match in item['matches']]
for part in edge[1:]:
if part not in matched:
key = ed.edge2str(part)
if key in extra_edges:
extra_edges[key] += 1
else:
extra_edges[key] = 1
sorted_edges = sorted(extra_edges.items(),
key=operator.itemgetter(1),
# read data
# edge_data = json_tools.read('edges_similar_concepts.json')
# build extra edges list
# extra_edges = []
# full_edges = []
# for it in edge_data:
# e = ed.str2edge(it['edge'])
# full_edges.append(e)
# matched = [ed.str2edge(match[1]) for match in it['matches']]
# for part in e[1:]:
# if part not in matched:
# extra_edges.append(part)
edge_data = json_tools.read('all.json')
# build full edges list
extra_edges = []
for it in edge_data:
extra_edges.append(ed.without_namespaces(ed.str2edge(it['edge'])))
full_edges = extra_edges
ag = AtomGroups(par)
print('set edges')
ag.set_edges(extra_edges)
print('generate_atoms')
ag.generate_atoms()
print('generate synonyms')
ag.generate_synonyms()
print('generate atom groups')
ag.generate_atom_groups()