import multiprocessing
import os, time, random
import numpy as np
import cv2
import os
import sys
from time import ctime
import tensorflow as tf
 
image_dir = r"D:/sxl/處理圖片/漢字分類/train10/"    #圖像文件夾路徑
data_type = 'test'
save_path = r'E:/sxl_Programs/Python/CNN/npy/'    #存儲路徑
data_name = 'Img10'                #npy文件名
 
char_set = np.array(os.listdir(image_dir))      #文件夾名稱列表
np.save(save_path+'ImgShuZi10.npy',char_set)     #文件夾名稱列表
char_set_n = len(char_set)              #文件夾列表長度
 
read_process_n = 1  #進程數
repate_n = 4     #隨機移動次數
data_size = 1000000  #1個npy大小
 
shuffled = True   #是否打亂
 
#可以讀取帶中文路徑的圖
def cv_imread(file_path,type=0):
  cv_img=cv2.imdecode(np.fromfile(file_path,dtype=np.uint8),-1)
  # print(file_path)
  # print(cv_img.shape)
  # print(len(cv_img.shape))
  if(type==0):
    if(len(cv_img.shape)==3):
      cv_img = cv2.cvtColor(cv_img, cv2.COLOR_BGR2GRAY)
  return cv_img
 
#多個數組按同一規則打亂數據
def ShuffledData(features,labels):
  '''
  @description:隨機打亂數據與標簽，但保持數據與標簽一一對應
  '''
  permutation = np.random.permutation(features.shape[0])
  shuffled_features = features[permutation,:] #多維
  shuffled_labels = labels[permutation]    #1維
  return shuffled_features,shuffled_labels
 
#函數功能：簡單網格
#函數要求：1.無關圖像大小；2.輸入圖像默認為灰度圖;3.參數只有輸入圖像
#返回數據：1x64*64維特征
def GetFeature(image):
 
  #圖像大小歸一化
  image = cv2.resize(image,(64,64))
  img_h = image.shape[0]
  img_w = image.shape[1]
 
  #定義特征向量
  feature = np.zeros(img_h*img_w,dtype=np.int16)
 
  for h in range(img_h):
    for w in range(img_w):
      feature[h*img_h+w] = image[h,w]
 
  return feature
 
# 寫數據進程執行的代碼:
def read_image_to_queue(queue):
  print('Process to write: %s' % os.getpid())
  for j,dirname in enumerate(char_set): # dirname 是文件夾名稱
    label = np.where(char_set==dirname)[0][0]   #文件夾名稱對應的下標序號
    print('序號：'+str(j),'讀 '+dirname+' 文件夾...時間：',ctime() )
    for parent,_,filenames in os.walk(os.path.join(image_dir,dirname)):
      for filename in filenames:
        if(filename[-4:]!='.jpg'):
          continue
        image = cv_imread(os.path.join(parent,filename),0)
 
        # cv2.imshow(dirname,image)
        # cv2.waitKey(0)
        queue.put((image,label))
  
  for i in range(read_process_n):
    queue.put((None,-1))
 
  print('讀圖結束!')
  return True
    
# 讀數據進程執行的代碼:
def extract_feature(queue,lock,count):
  '''
  @description:從隊列中取出圖片進行特征提取
  @queue:先進先出隊列
   lock：鎖，在計數時上鎖，防止沖突
   count:計數
  '''
 
  print('Process %s start reading...' % os.getpid())
 
  global data_n
  features = [] #存放提取到的特征
  labels = [] #存放標簽
  flag = True #標志著進程是否結束
  while flag:
    image,label = queue.get() #從隊列中獲取圖像和標簽
 
    if len(features) >= data_size or label == -1:  #特征數組的長度大于指定長度，則開始存儲
 
      array_features = np.array(features) #轉換成數組
      array_labels = np.array(labels)
 
      array_features,array_labels = ShuffledData(array_features,array_labels) #打亂數據
      
      lock.acquire()  # 鎖開始
 
      # 拆分數據為訓練集，測試集
      split_x = int(array_features.shape[0] * 0.8)
      train_data, test_data = np.split(array_features, [split_x], axis=0)   # 拆分特征數據集
      train_labels, test_labels = np.split(array_labels, [split_x], axis=0) # 拆分標簽數據集
 
      count.value += 1  #下標計數加1
      str_features_name_train = data_name+'_features_train_'+str(count.value)+'.npy'
      str_labels_name_train = data_name+'_labels_train_'+str(count.value)+'.npy'
      str_features_name_test = data_name+'_features_test_'+str(count.value)+'.npy'
      str_labels_name_test = data_name+'_labels_test_'+str(count.value)+'.npy'
 
      lock.release()  # 鎖釋放
 
      np.save(save_path+str_features_name_train,train_data)
      np.save(save_path+str_labels_name_train,train_labels)
      np.save(save_path+str_features_name_test,test_data)
      np.save(save_path+str_labels_name_test,test_labels)
      print(os.getpid(),'save:',str_features_name_train)
      print(os.getpid(),'save:',str_labels_name_train)
      print(os.getpid(),'save:',str_features_name_test)
      print(os.getpid(),'save:',str_labels_name_test)
      features.clear()
      labels.clear()
 
    if label == -1:
      break
 
    # 獲取特征向量，傳入灰度圖
    feature = GetFeature(image)
    features.append(feature)
    labels.append(label)
 
    # # 隨機移動4次
    # for itime in range(repate_n):
    #   rMovedImage = randomMoveImage(image)
    #   feature = SimpleGridFeature(rMovedImage) # 簡單網格
    #   features.append(feature)
    #   labels.append(label)
  
  print('Process %s is done!' % os.getpid())
 
if __name__=='__main__':
  time_start = time.time() # 開始計時
 
  # 父進程創建Queue，并傳給各個子進程：
  image_queue = multiprocessing.Queue(maxsize=1000) #隊列
  lock = multiprocessing.Lock()           #鎖
  count = multiprocessing.Value('i',0)        #計數
 
  #將圖寫入隊列進程
  write_sub_process = multiprocessing.Process(target=read_image_to_queue, args=(image_queue,))
 
  read_sub_processes = []              #讀圖子線程
  for i in range(read_process_n):
    read_sub_processes.append(
      multiprocessing.Process(target=extract_feature, args=(image_queue,lock,count))
    )
 
  # 啟動子進程pw，寫入:
  write_sub_process.start()
 
  # 啟動子進程pr，讀取:
  for p in read_sub_processes:
    p.start()
 
  # 等待進程結束:
  write_sub_process.join()
  for p in read_sub_processes:
    p.join()
 
  time_end=time.time()
  time_h=(time_end-time_start)/3600
  print('用時：%.6f 小時'% time_h)
  print ("讀圖提取特征存npy,運行結束！")