def sim_two_question():
"""测试一下两个问题的相似句子"""
from FeatureProject.bert.extract_keras_bert_feature import KerasBertVector
from sklearn import preprocessing
from math import pi
import numpy as np
import time
import math
def cosine_distance(v1, v2): # 余弦距离
if type(v1) == list:
v1 = np.array(v1)
if type(v2) == list:
v2 = np.array(v2)
if v1.all() and v2.all():
return np.dot(v1, v2) / (np.linalg.norm(v1) * np.linalg.norm(v2))
else:
return 0
def scale_zoom(rate): # sig 缩放
zoom = (1 + np.exp(-float(rate))) / 2
return zoom
def scale_triangle(rate): # sin 缩放
triangle = math.sin(rate / 1 * pi / 2 - pi / 2)
return triangle
bert_vector = KerasBertVector()
print("bert start ok!")
while True:
print("input ques-1: ")
ques_1 = input()
print("input ques_2: ")
ques_2 = input()
vector_1 = bert_vector.bert_encode([ques_1])
vector_2 = bert_vector.bert_encode([ques_2])
sim = cosine_distance(vector_1[0], vector_2[0])
# sim_list = [sim, 0, 0.2, 0.4, 0.6, 0.8, 1.0]
# sim = preprocessing.scale(sim_list)[0]
# sim = preprocessing.MinMaxScaler(feature_range=(0, 1)).fit_transform(sim_list)[0]
# sim_1 = preprocessing.normalize(sim_list, norm='l1')[0]
# sim_2 = preprocessing.normalize(sim_list, norm='l2')[0]
# sim = scale_zoom(sim)
# sim = scale_triangle(sim)
# print(sim_1)
# print(sim_2)
print(sim)
def calculate_count():
"""
统计一下1000条测试数据的平均耗时
:return:
"""
from FeatureProject.bert.extract_keras_bert_feature import KerasBertVector
import time
bert_vector = KerasBertVector()
print("bert start ok!")
time_start = time.time()
for i in range(10):
vector = bert_vector.bert_encode(["jy，你知道吗，我一直都很喜欢你呀，在一起在一起在一起，哈哈哈哈"])
time_end = time.time()
time_avg = (time_end - time_start) / 10
print(vector)
print(time_avg)
def chatbot_sentence_vec_by_bert_own():
"""bert encode is writted by my own"""
from FeatureProject.bert.extract_keras_bert_feature import KerasBertVector
from conf.path_config import chicken_and_gossip_path
from utils.text_tools import txtRead
import numpy as np
# 读取数据和一些参数，这里只取了100个标准问题
topk = 5
matrix_ques_save_path = "doc_vecs_chicken_and_gossip"
questions = txtRead(chicken_and_gossip_path, encodeType='utf-8')
ques = [ques.split('\t')[0] for ques in questions][0:100]
# 生成标准问题的bert句向量
bert_vector = KerasBertVector()
ques_basic_vecs = bert_vector.bert_encode(ques)
# 线上你可以生成，直接调用，然后直接load就好
np.savetxt(matrix_ques_save_path, ques_basic_vecs)
# matrix_ques = np.loadtxt(matrix_ques_save_path)
query_bert_vec = bert_vector.bert_encode(["小姜机器人是什么"])[0]
query_bert_vec = np.array(query_bert_vec)
print(query_bert_vec)
# 矩阵点乘，很快的，你也可以用annoy等工具，计算就更加快了
qq_score = np.sum(query_bert_vec * ques_basic_vecs, axis=1) / np.linalg.norm(ques_basic_vecs, axis=1)
topk_idx = np.argsort(qq_score)[::-1][:topk]
for idx in topk_idx:
print('小姜机器人回答检索： %s\t%s' % (qq_score[idx], questions[idx]))
while True:
print("你的问题:")
query = input()
query_bert_vec = bert_vector.bert_encode([query])[0]
query_bert_vec = np.array(query_bert_vec)
# 矩阵点乘，很快的，你也可以用annoy等工具，计算就更加快了
qq_score = np.sum(query_bert_vec * ques_basic_vecs, axis=1) / np.linalg.norm(ques_basic_vecs, axis=1)
topk_idx = np.argsort(qq_score)[::-1][:topk]
for idx in topk_idx:
print('小姜机器人回答检索： %s\t%s' % (qq_score[idx], questions[idx]))