main.py

# from gensim.models import Word2Vec
# import numpy as np
# TRAIN=r'D:\Machinelearn\envs\pynlp\pyhanlp-master\pyhanlp\static\data\test\icwb2-data\training\pku_training.txt'
# TEST=r'D:\Machinelearn\envs\pynlp\pyhanlp-master\pyhanlp\static\data\test\icwb2-data\testing\pku_test.txt'
# test_vector='test_vector.txt'
# out_w2v='out_w2v.txt'
# f1=open(TEST,'r')
# sentences=[]
# for line in f1:
#     line=line.strip('\n')
#     sentence=''.join(line.split())
#     word_list=[word for word in sentence]
#     sentences.append(word_list)
# model=Word2Vec(sentences,vector_size=100,window=5,min_count=1,workers=4)
# # f=open('output_file.txt', 'w', encoding='utf-8')
# # model.save_word2vec_format(f, binary=False)
# # f.close()
# # model.save(out_w2v)
# # model = Word2Vec.load('out_w2v.txt')
# model.wv.save_word2vec_format(test_vector, binary=False)
# f1.close()
#
# import json
#
# TRAIN=r'D:\Machinelearn\envs\pynlp\pyhanlp-master\pyhanlp\static\data\test\icwb2-data\training\pku_training.txt'
# TEST=r'D:\Machinelearn\envs\pynlp\pyhanlp-master\pyhanlp\static\data\test\icwb2-data\gold\pku_test_gold.txt'
#
# f=open(TEST,'r')
#
# words={}
# text=f.read()
# text=text.replace('\n',' ')
# text_list=text.split()
#
# for word in text_list:
#     if len(word)==1:
#         words[word]=words.get(word,dict())
#         words[word]['S']=words[word].get('S',0)
#         words[word]['S']+=1
#     else:
#         words[word[0]]=words.get(word[0],dict())
#         words[word[0]]['B']=words[word[0]].get('B',0)
#         words[word[0]]['B']+=1
#
#         for i in word[1:-1]:
#             words[i] = words.get(i, dict())
#             words[i]['M'] = words[i].get('M', 0)
#             words[i]['M'] += 1
#
#         words[word[-1]]=words.get(word[-1],dict())
#         words[word[-1]]['E']=words[word[-1]].get('E',0)
#         words[word[-1]]['E']+=1
#
# with open('test_label.json','w') as f:
#     json.dump(words,f)
#
# f.close()
#
# import json
#
# f1=open('test_label.json','r')
# data=json.load(f1)
# f1.close()
#
# labels=[]
# f2=open('test_vector.txt','r',encoding='utf-8')
# outvectors=f2.readlines()
# for vector in outvectors[1:]:
#     if vector[0] in data:
#         labels.append((vector[0],max(data[vector[0]])))
# f2.close()
#
# f3=open('test_gold.txt','w',encoding='utf-8')
# for word in labels:
#     f3.write(' '.join(word)+'\n')
# f3.close()
#
# import numpy as np
# TRAIN=r'D:\Machinelearn\envs\pynlp\pyhanlp-master\pyhanlp\static\data\test\icwb2-data\training\pku_training.txt'
#
# train_label=[]
# with open(TRAIN,'r') as f:
#     t_text=f.read()
#     t_text=t_text.replace('\n',' ')
#     test_split=t_text.split()
#     for word in test_split:
#         if len(word)==1 :
#             train_label.append('S')
#         else:
#             train_label.append('B')
#             for i in range(len(word[1:-1])):
#                 train_label.append('M')
#             train_label.append('E')
#
# trans=np.zeros((4,4))
# l=['B','M','E','S']
# for i in range(1,len(train_label)):
#     f=l.index(train_label[i-1])
#     s=l.index(train_label[i])
#     trans[f,s]+=1
#
# trans/=len(train_label)-1
# np.save('trans.npy',trans)
#
# data=np.load('trans.npy')
# print(data)
#
# import matplotlib.pyplot as plt
# x=np.linspace(-5,5,100)
# plt.subplot(1,2,1)
# y1=1/(1+np.exp(-x))
# plt.plot(x,y1)
# plt.subplot(1,2,2)
# y2=np.tanh(x)
# plt.plot(x,y2)
# plt.show()
# import numpy as np
# import matplotlib.pyplot as plt
# from mpl_toolkits.mplot3d import Axes3D
# fig = plt.figure()
# ax = fig.add_subplot(111, projection='3d')
# w1=np.array([1,2,3]).reshape(1,3)
# print(w1.T)
import os
import random
import shutil

test_rate = 0.005  # 检测样本的百分比
file_folder = './Annotations'
file_names = os.listdir(file_folder)[2:]
full_file_names = [os.path.splitext(file_name)[0] for file_name in file_names]
test_num = round(test_rate * len(full_file_names))

with open('./ImageSets/Main/test.txt', 'w') as f_test, open('./ImageSets/Main/trainval.txt', 'w') as f_trainval:
    for i in range(test_num):
        random_w = random.random()
        test_name_tmp = full_file_names[round(random_w * len(full_file_names))]
        test_name = test_name_tmp.strip()
        f_test.write(test_name + '\n')
        full_file_names.remove(test_name_tmp)

    for full_file_name in full_file_names:
        f_trainval.write(full_file_name + '\n')

new_folder = '../test_image'
if not os.path.exists(new_folder):
    os.mkdir(new_folder)

test_nums = [line.strip() for line in open('./ImageSets/Main/test.txt').readlines()]
for test_num in test_nums:
    test_p = os.path.join('./JPEGImages', test_num + '.jpg')
    shutil.copy(test_p, os.path.join(new_folder, test_num + '.jpg'))