def __init__(self,
word_processor=word_processor.BaseWordProcessor(),
lang='en',
corenlp: my_corenlp.StanfordCoreNLP = None):
super().__init__(word_processor)
self.stanford_nlp = corenlp
if not self.stanford_nlp:
self.stanford_nlp = corenlp_factory.create_corenlp_server(
lang=lang)
first_word = coref_relation[i]
second_word = coref_relation[j]
for k in range(first_word[1], first_word[2]):
first_index = sentence_start_indices[first_word[0] -
1] + k - 1
for l in range(second_word[1], second_word[2]):
second_index = sentence_start_indices[second_word[0] -
1] + l - 1
print(
'first_word: %s second_word: %s' %
(all_words[first_index], all_words[second_index]))
word_relation_graph[first_index][second_index] = 1
word_relation_graph[second_index][first_index] = 1
return word_relation_graph
if __name__ == '__main__':
sentence = 'When the food came, it was almost good.'
core_nlp = corenlp_factory.create_corenlp_server()
coref_graph = get_coref_edges(sentence, core_nlp)
spacy_nlp = spacy.load("en_core_web_sm")
doc = spacy_nlp(sentence)
for token in doc:
children = list(token.children)
print(token)
# neuralcoref.add_to_pipe(spacy_nlp)
graph = create_dependency_graph_for_dgl(sentence, spacy_nlp)
plot_dgl_graph(graph)
print('')
def __init__(self,
word_processor=word_processor.BaseWordProcessor(),
lang='en'):
super().__init__(word_processor)
self.stanford_nlp = corenlp_factory.create_corenlp_server(lang=lang)
words, postags, arcs, child_dict_list, format_parse_list = self.sentence_parse_result[
dependency_key]
return words, postags, arcs, child_dict_list, format_parse_list
else:
words, postags, arcs, child_dict_list, format_parse_list = self._inner_parser_main(
sentence)
self.sentence_parse_result[dependency_key] = [
words, postags, arcs, child_dict_list, format_parse_list
]
self.cache_keys.append(dependency_key)
if len(self.cache_keys) > self.max_cache_sentence_num:
cache_key = self.cache_keys.pop(0)
self.sentence_parse_result.pop(cache_key)
return words, postags, arcs, child_dict_list, format_parse_list
else:
words, postags, arcs, child_dict_list, format_parse_list = self._inner_parser_main(
sentence)
return words, postags, arcs, child_dict_list, format_parse_list
if __name__ == '__main__':
parser = CorenlpParser(corenlp_factory.create_corenlp_server())
# sentence = sys.argv[1]
sentence = '以及公司内部的党派之争最终导致了尚阳科技在2006年梦碎当场。'
words, postags, arcs, child_dict_list, format_parse_list = parser.parser_main(
sentence)
print(postags, len(postags))
print(arcs, len(arcs))
print(child_dict_list, len(child_dict_list))
print(format_parse_list, len(format_parse_list))
if not constituency_parser_result.startswith('('):
leaf_node = TreeNode(constituency_parser_result)
return leaf_node
else:
constituency_parser_result = constituency_parser_result[1: -1]
first_whitespace_index = constituency_parser_result.index(' ')
value = constituency_parser_result[: first_whitespace_index]
parent_node = TreeNode(value)
constituency_parser_result = constituency_parser_result[first_whitespace_index + 1:]
sub_constituency_parser_results = sub_constituency_parser_result_generator(constituency_parser_result)
children = []
for sub_constituency_parser_result in sub_constituency_parser_results:
parse_result = parse_corenlp_parse_result(sub_constituency_parser_result)
children.append(parse_result)
for child in children:
child.parent = parent_node
parent_node.children = children
return parent_node
if __name__ == '__main__':
with corenlp_factory.create_corenlp_server(lang='zh', start_new_server=True) as nlp:
sentence = '【多国隐形战机“暗战”巴黎航展】图片说明：航空工业展台的隐形战机模型引发关注。'
sentence = re.sub('[\(\)]', '', sentence)
result = nlp.parse(sentence)
print(result)
constituency_tree = parse_corenlp_parse_result(result)
print(TreeNode.find_corresponding_node(constituency_tree, '模型', 20))
print('end')
