import urllib.request
import os
import numpy
import pandas as pd
from sklearn import preprocessing
from keras.models import Sequential
from keras.layers import Dense,Dropout
import matplotlib.pyplot as plt

url = 'http://biostat.mc.vanderbilt.edu/wiki/pub/Main/DataSets/titanic3.xls'
filepath = 'titanic3.xls'

if not os.path.isfile(filepath):
    result = urllib.request.urlretrieve(url,filepath)
    print('download:',result)

all_df = pd.read_excel(filepath)
print(all_df[:2])

def PreprocessData(raw_df):
    raw_df = raw_df.drop(['name'], axis=1)
    age_mean = raw_df['age'].mean()
    raw_df['age'] = raw_df['age'].fillna(age_mean)
    fare_mean = raw_df['fare'].mean()
    raw_df['fare'] = raw_df['fare'].fillna(fare_mean)
    raw_df['sex'] = raw_df['sex'].map({'female': 0, 'male': 1}).astype(int)
    x_OneHot_df = pd.get_dummies(data=raw_df, columns=['embarked'])

    ndarray = x_OneHot_df.values
    Label = ndarray[:, 0]
    Features = ndarray[:, 1:]

    minmax_scale = preprocessing.MinMaxScaler(feature_range=(0, 1))
    scaledFeatures = minmax_scale.fit_transform(Features)

    return scaledFeatures,Label

cols = ['survived','name','pclass','sex','age','sibsp','parch','fare','embarked']
all_df = all_df[cols]
print(all_df[:2])

msk = numpy.random.rand(len(all_df)) < 0.8
train_df = all_df[msk]
test_df = all_df[~msk]

print('total:',len(all_df),'train:',len(train_df),'test:',len(test_df))

train_Features,train_Label = PreprocessData(train_df)
test_Features,test_Label = PreprocessData(test_df)

model = Sequential()

model.add(Dense(units=40,input_dim=9,kernel_initializer='uniform',activation='relu'))

model.add(Dense(units=30,kernel_initializer='uniform',activation='relu'))

model.add(Dense(units=1,kernel_initializer='uniform',activation='sigmoid'))

model.compile(loss='binary_crossentropy',optimizer='adam',metrics=['accuracy'])

train_history = model.fit(x=train_Features,y=train_Label,validation_split=0.1,epochs=30,batch_size=30,verbose=2)

def show_train_history(train_history, train, validation):
    plt.plot(train_history.history[train])
    plt.plot(train_history.history[validation])
    plt.title('Train History')
    plt.ylabel(train)
    plt.xlabel('Epoch')
    plt.legend(['train', 'validation'], loc='upper left')
    plt.show()


show_train_history(train_history, 'acc', 'val_acc')
show_train_history(train_history, 'loss', 'val_loss')

scores = model.evaluate(x=test_Features,y=test_Label)
print(scores[1])

Jack = pd.Series([0,'Jack',3,'male',23,1,0,5.0000,'S'])
Rose = pd.Series([1,'Rose',1,'female',20,1,0,100.0000,'S'])
JR_df = pd.DataFrame([list(Jack),list(Rose)],
                     columns=['survived','name','pclass','sex','age','sibsp','parch','fare','embarked'])

JR_df = pd.concat([all_df,JR_df])
JR_Features,JR_Label = PreprocessData(JR_df)
JR_probability = model.predict(JR_Features)

print(JR_probability)

pd = JR_df
pd.insert(len(JR_df.columns),'probability',JR_probability)

print(pd)