from knock41 import Chunk, get_neko_list
for line in get_neko_list():
for chunk in line:
verb = ''
for morph in chunk.get_morphs():
if morph.get_pos() == '動詞':
verb = morph.get_base()
if verb == '':
continue
particle_list = list()
for src in chunk.get_srcs():
src_morph = line[src].get_word_only_morphs()
if len(src_morph) > 0 and src_morph[-1].get_pos() == '助詞':
pair = [src_morph[-1].get_base(), line[src].get_word_only()]
particle_list.append(pair)
if len(particle_list) > 0:
# 中身をlist型にしておかないとpopができない
particle, phrase = list(map(list, zip(*sorted(particle_list))))
if 'を' in particle:
wo_case = phrase.pop(particle.index('を'))
particle.pop(particle.index('を'))
print('{}\t{}\t{}'.format(wo_case + verb, ' '.join(particle),
' '.join(phrase)))
from knock41 import Chunk, get_neko_list
from graphviz import Digraph
for i, line in enumerate(get_neko_list()):
if i > 5:
break
graph = Digraph(format='png')
graph.attr('node', shape='circle')
for chunk in line:
origin = chunk.get_string()
destination = line[chunk.get_dst()].get_string()
graph.node(origin)
if chunk.get_dst() != -1:
graph.edge(origin, destination)
graph.render('knock44_graph{}'.format(i))