def fetch_all_content_result(document_list):
print("in fetch")
content_all_list = []
result_all_list = []
i = 0
myseg = MySegment()
for document in document_list:
content, result = content_result(myseg, document)
content_all_list.append(content)
result_all_list.append(result)
i += 1
print(i)
myseg.close()
return content_all_list, result_all_list
def seg_document(document_list):
content_list = []
result_list = []
i = 0
myseg = MySegment()
mypos = MyPostagger()
for document in document_list:
# print(i)
# i = i + 1
content_wordlist, result_wordlist = content_result(myseg, document)
content_wordlist = mypos.words2pos(content_wordlist,
['n', 'nl', 'ns', 'v'])
result_wordlist = mypos.words2pos(result_wordlist,
['n', 'nl', 'ns', 'v'])
content_list.append(content_wordlist)
result_list.append(result_wordlist)
myseg.close()
mypos.close()
print("----------------------------------------------")
return content_list, result_list
def seg_document(document_list):
content_list = []
result_list = []
myseg = MySegment()
i = 0
for document in document_list:
print(i)
content_wordlist, result_wordlist = content_result(myseg, document)
content_list.append(content_wordlist)
result_list.append(result_wordlist)
i += 1
return content_list, result_list
if __name__ == "__main__":
criminal_list = [
'交通肇事罪', # 危险驾驶罪（危险 驾驶罪）
'过失致人死亡罪', # 故意杀人罪（故意 杀人 杀人罪） 故意伤害罪（故意 伤害 伤害罪）
'故意杀人罪',
'故意伤害罪',
'过失致人重伤罪',
'抢劫罪',
#'诈骗罪', #（诈骗 诈骗罪 诈骗案）
'拐卖妇女儿童罪'
]
myseg = MySegment()
dataH = dataHelper()
case_dict = dict()
save_list = list()
reason_list = list()
for criminal in criminal_list:
print("~~~~~~~~~~~~~~~~~~{}~~~~~~~~~~~~~~~~~~".format(criminal))
# case_dict = dict()
# save_list = list()
# reason_list = list()
file_dir = BasePath + criminal + "/"
dir_list = os.listdir(file_dir)
# print(' '.join(list(return_word_set)))
return list(return_word_set)
def get_details_words(myseg, mypos, document):
behavior_set = {"自首", "坦白", "累犯", "谅解", "和解"}
# print("+++++++++++++++++++++++++++++++")
decode_document = get_details(document)
# print(decode_document)
details_words = myseg.sen2word(decode_document.encode('utf8'))
# print(' '.join(details_words))
sdetails_words = list(
set(mypos.words2pos(details_words, ['n', 'nl', 'ns', 'v'])) -
behavior_set)
# print(' '.join(sdetails_words))
return sdetails_words
if __name__ == "__main__":
file_path = BasePath + "/data/judgment_yishen.txt"
print(file_path)
myseg = MySegment()
mypos = MyPostagger()
document_list = read_document(file_path)
for document in document_list:
print("___________________________________________________")
print(get_details(document))
myseg.close()
mypos.close()
len(document_all_id_list)))
# 语料向量
x_sample = np.loadtxt(BasePath +
"/word2vec_model/corpus_w2v_full_finance_average.txt")
print("load the corpus vector in : {}".format(
BasePath + "/word2vec_model/corpus_w2v_full_average.txt"))
# # 随机森林训练
clf_filepath = BasePath + "/data/clf_model_full_average.m"
if os.path.exists(clf_filepath):
print("the model already exists in :{}".format(clf_filepath))
clf = joblib.load(clf_filepath)
else:
print("No model loaded!")
# 分词模块
myseg = MySegment()
opt_Document = DocumentsOnMysql()
def rf_similarity(path_vec):
fea_size = len(path_vec)
sim_vec = np.zeros([fea_size, fea_size])
k = 0
for i in range(0, fea_size):
for j in range(i, fea_size):
print(k)
k = k + 1
num_path = np.sum(path_vec[i] & path_vec[j])
sim_vec[i][j] = num_path
sim_vec[j][i] = num_path
return sim_vec / sim_vec.diagonal().T