import os
import cv2
import numpy as np
import matplotlib.pyplot as plt
import h5py

def get_file(file_dir):
    a_image = []
    a_label = []
    b_image = []
    b_label = []
    c_image = []
    c_label = []
    d_image = []
    d_label = []
    first0 = True
    first1 = True
    first2 = True
    first3 = True
    for file in os.listdir(file_dir+'/option1'):
            a_image.append(file_dir+'/option1/'+file)
            a_label.append(0)
    del a_image[0]
    del a_label[0]

    for file in os.listdir(file_dir+'/option2'):
            b_image.append(file_dir+'/option2/'+file)
            b_label.append(1)
    del b_image[0]
    del b_label[0]

    for file in os.listdir(file_dir+'/option3'):
            c_image.append(file_dir+'/option3/'+file)
            c_label.append(2)
    del c_image[0]
    del c_label[0]

    for file in os.listdir(file_dir+'/option4'):
            d_image.append(file_dir+'/option4/'+file)
            d_label.append(3)
    del d_image[0]
    del d_label[0]

    image_list = np.hstack((a_image,b_image,c_image,d_image))
    print(image_list)
    label_list = np.hstack((a_label,b_label,c_label,d_label))
    print(label_list)

    temp = np.array([image_list , label_list])
    temp = temp.transpose()
    np.random.shuffle(temp)

    image_list = list(temp[:,0])
    label_list = list(temp[:,1])
    label_list = [float(i) for i in label_list]
    return image_list ,label_list

    image_list  , label_list = get_file('gesture_picture')

    image = np.zeros((len(image_list) , 28 ,28))
label = np.zeros(len(label_list))
for i in range(len(image)):
    image[i] = np.array(cv2.imread(image_list[i],cv2.IMREAD_GRAYSCALE))
    label[i] = np.array(label_list[i])
    label = label.astype(np.int32)
import matplotlib.pyplot as plt
def plot_image(images , labels , idx , num = 10):
    fig = plt.gcf()
    fig.set_size_inches(12 , 14)
    for i in range(num):
        ax = plt.subplot(5,5,i+1)
        ax.imshow(image[i] , cmap = 'binary')
        ax.set_title(label[i])
        ax.set_xticks([])
        ax.set_yticks([])
        plt.show()
plot_image(image , label , 0 , 25)

msk = np.random.randn(len(image))<0.8
img_train = image[msk]
label_train = label[msk]
img_test = image[~msk]
label_test = label[~msk]

 def printshape():
    print(img_train.shape)
    print(label_train.shape)
    print(img_test.shape)
    print(label_test.shape)

    printshape()
from keras.utils import np_utils

img_train_nor = np.reshape(img_train/255.0 , (len(img_train),28, 28 ,1))
img_test_nor = np.reshape(img_test/255.0 , (len(img_test),28 ,28 ,1))

label_train_hot = np_utils.to_categorical(label_train)
label_test_hot = np_utils.to_categorical(label_test)

img_train_nor.shape
img_test_nor.shape

from keras.models import Sequential
from keras.layers import Dense
from keras.layers import Dense ,Dropout , Activation , Flatten
from keras.layers import Conv2D , MaxPooling2D , ZeroPadding2D

model = Sequential()
model.add(Conv2D(filters = 32 , kernel_size = (3,3) , input_shape = (28,28,1) , activation = 'relu' , padding = 'same'))
model.add(Dropout(0.25))
model.add(MaxPooling2D(pool_size = (2 , 2)))
model.add(Conv2D(64 , (3,3) , activation='relu', padding='same'))
model.add(Dropout(0.25))
model.add(MaxPooling2D(pool_size = (2 , 2)))
model.add(Flatten())
model.add(Dropout(0.25))
model.add(Dense(1024 , activation = 'relu'))
model.add(Dropout(0.25))
model.add(Dense(4 , activation = 'softmax'))

model.compile(
                loss = 'categorical_crossentropy',
                optimizer = 'adam',
                metrics = ['accuracy']
             )
model.fit(img_train_nor , label_train_hot , epochs =3 ,batch_size = 50 ,validation_split = 0.2)


from keras.models import load_model
model.save('my_model_cnn.h5')  # creates a HDF5 file 'my_model.h5'

import numpy as np
import matplotlib.pyplot as plt
import cv2
import os
cap = cv2.VideoCapture(0)
opened = cap.isOpened
while(opened):
        n = 0
        (flag , frame) = cap.read(0)
        gray = cv2.cvtColor(frame , cv2.COLOR_BGR2GRAY)
        img = cv2.resize(gray,(28,28))

        #====
        if(n%150 == 0):
            img = img/255.0
            img = np.array(img)
            img = np.reshape(img , (1, 28 , 28 , 1))
            k = model.predict(img)
            if k.argmax()==0:
                print('空白')
            elif k.argmax()==1:
                print('剪刀')
            elif k.argmax()==2:
                print('石頭')
            elif k.argmax()==3:
                print('布')
        #====

        cv2.imshow('frame',gray)
        if cv2.waitKey(1) & 0xFF == ord('q'):
            break
        #plt.imshow(frame)
        #plt.show()
        n+=1
cv2.destroyAllWindows()

from keras.models import load_model
model = load_model('my_model.h5')
scores = model.evaluate(img_test_nor , label_test_hot)