# coding: utf-8

# In[1]:

import numpy as np

import matplotlib.pylab as plt

from collections import Counter

import IOutils

import random

# from neuralnetworks.templates import BasicCNN

from sklearn.metrics import confusion_matrix, classification_report

import theano

import lasagne

from lasagne import layers

from lasagne.updates import nesterov_momentum

from nolearn.lasagne import NeuralNet

from neuralnetworks.tools import LayerFactory

from preprocessing.preprocess_utils import window_generator_ND

from metrics.custom import multiple_auc

# In[2]:

WINDOW_SIZE = 300

# In[3]:

LF = LayerFactory()

layers_list = [

]

# In[4]:

nn = NeuralNet(layers_list,

**LF.kwargs)

# In[5]:

training_ds = IOutils.data_streamer(patients_list=[2], series_list=range(1,7))

# nn = BasicCNN(input_shape=(None,42), output_num_units=12, max_epochs=50, hidden=[256, 120], add_drops=[1,1])

vt = IOutils.VectorTransformer()

# In[ ]:

n_repeat_sampling = 1

dataset_count = 0

for X,Y in training_ds:

X = X.astype(np.float)

X[np.isnan(X)] = 0

X = X/X.max()

wg = window_generator_ND(X, window_size=WINDOW_SIZE)

dataset_count += 1

# transform the Ys

Y_train = vt.transform(Y)

time_point = 0

X_train = np.array(list(wg))

print X_train.shape

any_nans = np.isnan(X_train).any()

any_infs = np.isinf(X_train).any()

if any_nans or any_infs:

print('NANS WERE FOUND')

break

# print X_train.shape

# fit the classifier

nn.fit(X_train, Y_train)

#end windower

#endfor

# In[ ]:

from metrics.custom import multiple_auc

X_test, Y_test = IOutils.data_streamer(patients_list=[2], series_list=[8]).next()

# X_test, Y_test = testing_ds.next()

X_test = X_test.astype(np.float)

X_test[np.isnan(X_test)] = 0

X_test = X_test/X_test.max()

Y_test = vt.transform(Y_test).astype(np.int32).reshape(-1)

wg = window_generator_ND(X_test, window_size=WINDOW_SIZE)

# predicted = nn.predict(X_test)

# print confusion_matrix(Y_test, predicted)

# # In[ ]:

# print classification_report(Y_test, predicted)

# print multiple_auc(Y_test, predicted)

# In[ ]:

testing_data = np.array(list(wg))

# In[ ]:

testing_data.shape

import cPickle as pickle

pickle.dump(nn, open('./CNN_perpatient.pickle', 'w'))

print('Pickle saved')

print('terminate')

# In[ ]: