def write_kakariuke_tree_oneline(i, oneline):
graph = pydot.Dot(graph_type='digraph')
map_id_and_node_graph_includes = {}
for i_s, src in enumerate(oneline):
if src.dst() == -1:
continue
chunk_src_string = NLP42.concat_morphs(src.morphs())
chunk_dst_string = NLP42.concat_morphs(oneline[src.dst()].morphs())
if len(chunk_src_string) == 0 or len(chunk_dst_string) == 0:
continue
i_d = src.dst()
if i_s in map_id_and_node_graph_includes:
src_node = map_id_and_node_graph_includes[i_s]
else:
src_node = pydot.Node(str(i_s), label = chunk_src_string)
map_id_and_node_graph_includes[i_s] = src_node
graph.add_node(src_node)
if i_d in map_id_and_node_graph_includes:
dst_node = map_id_and_node_graph_includes[i_d]
else:
dst_node = pydot.Node(str(i_d), label = chunk_dst_string)
map_id_and_node_graph_includes[i_d] = dst_node
graph.add_node(pydot.Node(str(i_d), label = chunk_dst_string))
graph.add_edge(pydot.Edge(src_node, dst_node))
graph.write_png(output_dir + '/' + str(i) + '.png')
def write_kakariuke_oneline_2(oneline):
for src in oneline:
if src.dst() == -1:
continue
if not (has_chunk_noun(src) and has_chunk_verb(oneline[src.dst()])):
continue
chunk_src_string = NLP42.concat_morphs(src.morphs())
chunk_dst_string = NLP42.concat_morphs(oneline[src.dst()].morphs())
if len(chunk_src_string) == 0 or len(chunk_dst_string) == 0:
continue
print(chunk_src_string, chunk_dst_string, sep = '\t')
def write_sa_verb_and_pattern_of_particle_and_morphs_oneline(i, oneline):
map_sa_verb_and_pattern_of_particle_and_morphs = {}
for i_d, dst in enumerate(oneline):
if not NLP43.is_verb(dst.morphs()[0]):
continue
for i_s in dst.srcs():
chunk = oneline[i_s].morphs()
if len(chunk) < 2:
continue
last_morph_of_src = chunk[-1]
second_last_morph_of_src = chunk[-2]
if not (is_particle_wo(last_morph_of_src) and is_noun_connected_sa(second_last_morph_of_src)):
continue
key_base = NLP42.concat_morphs(chunk) + dst.morphs()[0].base()
for i_src_of_s in dst.srcs():
if i_src_of_s == i_s:
continue
c = oneline[i_src_of_s].morphs()
c = omit_punctual_from_morphs(c)
if len(c) < 1:
continue
last_morph_of_i_src_of_s = c[-1]
if NLP45.is_particle(last_morph_of_i_src_of_s):
if key_base in map_sa_verb_and_pattern_of_particle_and_morphs:
d = map_sa_verb_and_pattern_of_particle_and_morphs[key_base]
d['particles'].append(copy.deepcopy(last_morph_of_i_src_of_s.surface()))
d['morphs'].append(copy.deepcopy(c))
else:
new_dict = dict(particles = copy.deepcopy([last_morph_of_i_src_of_s.surface()]), morphs = copy.deepcopy([c]))
map_sa_verb_and_pattern_of_particle_and_morphs[key_base] = new_dict
write_verb_and_pattern_of_particle_and_morphs_result(i, map_sa_verb_and_pattern_of_particle_and_morphs)
def write_verb_and_pattern_of_particle_and_morphs_result(i, map_verb_and_pattern_of_particle_and_morphs):
#print(i)
for key_base in map_verb_and_pattern_of_particle_and_morphs:
d = map_verb_and_pattern_of_particle_and_morphs[key_base]
particles = d['particles']
morphs = d['morphs']
chunk_strings = list(map(lambda x: NLP42.concat_morphs(x), morphs))
if len(particles) > 0:
print(key_base, ' '.join(map(lambda x: str(x), particles)), ' '.join(chunk_strings), sep = '\t')
def get_morph_chain(nouns, chunks):
if isinstance(nouns, list):
chains = []
for i in nouns:
chains.append(get_morph_chain(i, chunks))
return chains
else:
if nouns is None:
return ''
chunk = chunks[nouns]
omitted = NLP47.omit_punctual_from_morphs(chunk.morphs())
omitted_string = NLP42.concat_morphs(omitted)
if chunk.dst() == -1:
return [omitted_string]
else:
return [omitted_string] + get_morph_chain(chunk.dst(), chunks)