from IPython.display import display
from biosppy.signals import ecg
from wfdb import processing
import matplotlib.pyplot as plt
import numpy as np
import os
import shutil
import wfdb
import tools as st
from keras.utils import plot_model

from keras.models import Sequential
from keras.layers import Dense, Embedding
from keras.layers import Dropout, Flatten, BatchNormalization
from keras.layers import Conv1D, MaxPooling1D, LSTM
from keras.models import model_from_json



'''list = []
record_result = []
input = open('records.txt', 'r')
for line in input:
    line1 = line.replace("\n","")
    arr = line1.split('/')
    _, fields = wfdb.rdsamp(arr[1], pb_dir='ptbdb/' + arr[0] + '/')
    if 'Healthy control' in fields['comments'][4]:
        list.append(1)
        record_result.append(line)
    if 'Myocardial infarction' in fields['comments'][4]:
        if 'inferior' in fields['comments'][5]:
            list.append(0)
            record_result.append(line)
f = open('output.txt','w')
ff = open('output_result.txt','w')
f.write("\n".join([str(x) for x in record_result]))
ff.write("\n".join([str(x) for x in list]))
f.close()
ff.close()'''


'''list = []
input = open('output.txt', 'r')
for line in input:
    line = line.replace("\n","")
    arr = line.split('/')
    record2 = wfdb.rdrecord(arr[1], pb_dir='ptbdb/' + arr[0] + '/', channels=[1,2,5])
    f = open('data.txt', 'w')
    f.write("# Simple Text Format\n")
    f.write("# Sampling Rate (Hz):= 1000.00\n")
    f.write("# Resolution:= 12\n")
    f.write("# Labels:= ECG\n")
    print(np.array(record2.p_signal).shape)
    for x in record2.p_signal:
        f.write(str(x[0]) + " " + str(x[1]) + " " + str(x[2]) + "\n")
    f.close()
    xxxs = ""
    xxx = open("data.txt")
    s=xxx.readlines()[4:]
    signal0 = np.loadtxt("data.txt", usecols = (0))
    out0 = ecg.ecg(signal=signal0, sampling_rate=1000., show = False)["templates"]
    signal1 = np.loadtxt("data.txt", usecols = (1))
    out1 = ecg.ecg(signal=signal1, sampling_rate=1000., show = False)["templates"]
    signal2 = np.loadtxt("data.txt", usecols = (2))
    out2 = ecg.ecg(signal=signal2, sampling_rate=1000., show = False)["templates"]
    ff = open('test.txt','w')
    b = []
    for x in range(len(out0[0])):
        b.append(sum(out0[:,x])/len(out0[:,x]))
    for x in range(len(out1[0])):
        b.append(sum(out1[:,x])/len(out1[:,x]))
    for x in range(len(out2[0])):
        b.append(sum(out2[:,x])/len(out2[:,x]))
    for i in range(len(b)):
        ff.write(str(b[i])+ "\n")
    ff.close()
    a = []
    for i in range(1,1801):
        a.append(i)
    load_data = np.array(b)
    list.append(load_data)
output = open('result_data.txt','w')
output.write("\n".join([",".join([str(x) for x in np_arr.tolist()]) for np_arr in list]))
output.close()'''


def model(test_data, res_data, validation_data, validation_res):
    model = Sequential()
    model.add(Conv1D(filters=25, kernel_size=3,activation='relu',input_shape=(600,3)))
    model.add(MaxPooling1D(pool_size = 2))
    model.add(Dropout(0.2))
    model.add(Conv1D(filters=20, kernel_size=3,activation='relu'))
    model.add(MaxPooling1D(pool_size = 2))
    model.add(Dropout(0.2))
    model.add(Conv1D(filters=15, kernel_size=3,activation='relu'))
    model.add(MaxPooling1D(pool_size = 2))
    model.add(Dropout(0.2))
    model.add(Conv1D(filters=10, kernel_size=3,activation='relu'))
    model.add(MaxPooling1D(pool_size = 2))
    model.add(Dropout(0.2))
    model.add(Flatten())
    model.add(Dense(600, activation='relu', input_dim = 600))
    model.add(Dense(50, activation='relu'))
    model.add(Dense(10, activation='relu'))
    model.add(Dropout(0.5))
    model.add(Dense(1, activation='sigmoid'))
    model.compile(loss="binary_crossentropy", optimizer="adam", metrics=['accuracy'])
    model.fit(np.array(test_data), np.array(res_data), epochs = 20, batch_size=8, verbose=1,validation_data=(np.array(validation_data),np.array(validation_res)))
    scores = model.evaluate(np.array(validation_data),np.array(validation_res), verbose=0)
    pred = model.predict(np.array(np.array(validation_data[:]).reshape(1,600,3)),verbose = 0)
    plot_model(model, to_file='model.png', show_shapes=True)
    return scores[1]


fi = open('result_data.txt')
test_data = [np.array([float(xs) for xs in x.split(",")]) for x in fi.readlines()]
test_data = np.array(test_data).reshape((169,600,3))

fu = open('output_result.txt','r')
res_data = [[int(x)] for x in fu.readlines()]
neural_res = []
k = 0
for i in range(len(res_data)):
    validation_data = np.array(test_data[i]).reshape(1,600,3)
    tst_data = test_data[:i]
    tst_data = np.array(np.append(tst_data, test_data[i+1:])).reshape(168,600,3)

    validation_res = list(res_data[i])
    rs_data = res_data[:i]
    rs_data = np.append(rs_data, res_data[i+1:])
    k = k + 1
    print(k)
    kk = model(tst_data,rs_data,validation_data,validation_res)
    neural_res.append(kk)
print(neural_res)
for i in (neural_res):
    itog = sum(neural_res)/len(neural_res)
print(itog)