# Estudo de Caso 1 - SVMs

####### Importando e Preparando o Dataset #######

# Import pandas
import pandas as pd

# Carrega o dataset
df = pd.read_csv('iris.csv')

# Remove a coluna ID
df = df.drop(['Id'],axis=1)

# Define a variável alvo
target = df['Species']

# Cria um conjunto
s = set()

# Valores possíveis da variável target
for val in target:
    s.add(val)

# Converte para lista
s = list(s)

# Separa em dados de treino e de teste
rows = list(range(100,150))
df = df.drop(df.index[rows])


####### Plot dos Dados #######

# Import matplotlib
import matplotlib.pyplot as plt

# Define x e y (entrada e saída)
x = df['SepalLengthCm']
y = df['PetalLengthCm']

# Dados referentes a uma classe de flores
setosa_x = x[:50]
setosa_y = y[:50]

# Dados referentes a outra classe de flores
versicolor_x = x[50:]
versicolor_y = y[50:]

# Plot
plt.figure(figsize=(8,6))
plt.scatter(setosa_x,setosa_y,marker='+',color='green')
plt.scatter(versicolor_x,versicolor_y,marker='_',color='red')
plt.show()


####### Algoritmo SVM #######

from sklearn.utils import shuffle
from sklearn.cross_validation import train_test_split
import numpy as np

# Descarta o restante dos recursos e extraia os valores de destino
df = df.drop(['SepalWidthCm','PetalWidthCm'],axis=1)
Y = []
target = df['Species']

for val in target:
    if(val == 'Iris-setosa'):
        Y.append(-1)
    else:
        Y.append(1)
df = df.drop(['Species'],axis=1)
X = df.values.tolist()

# Embaralhe e divida os dados de treinamento e testes
X, Y = shuffle(X,Y)
x_train = []
y_train = []
x_test = []
y_test = []

x_train, x_test, y_train, y_test = train_test_split(X, Y, train_size=0.9)

# Converte tudo para array NumPy
x_train = np.array(x_train)
y_train = np.array(y_train)
x_test = np.array(x_test)
y_test = np.array(y_test)

# Ajusta o shape
y_train = y_train.reshape(90,1)
y_test = y_test.reshape(10,1)

# Dados de treino das flores 1 e 2
train_f1 = x_train[:,0]
train_f2 = x_train[:,1]

# Ajusta o shape dos dados das flores 1 e 2
train_f1 = train_f1.reshape(90,1)
train_f2 = train_f2.reshape(90,1)

# Cria o vetor de pesos
w1 = np.zeros((90,1))
w2 = np.zeros((90,1))

# Hiperparâmetros
epochs = 1
alpha = 0.0001

# Treinamento
while(epochs < 10000):

    y = w1 * train_f1 + w2 * train_f2

    prod = y * y_train

    print(epochs)

    count = 0

    for val in prod:
        if(val >= 1):
            cost = 0
            w1 = w1 - alpha * (2 * 1/epochs * w1)
            w2 = w2 - alpha * (2 * 1/epochs * w2)

        else:
            cost = 1 - val
            w1 = w1 + alpha * (train_f1[count] * y_train[count] - 2 * 1/epochs * w1)
            w2 = w2 + alpha * (train_f2[count] * y_train[count] - 2 * 1/epochs * w2)
        count += 1
    epochs += 1


####### Avalia o Modelo #######

from sklearn.svm import SVC
from sklearn.metrics import accuracy_score
import warnings
warnings.filterwarnings("ignore")


# Agora, cortamos os pesos, pois os dados de teste contêm apenas 10 pontos de dados
index = list(range(10,90))
w1 = np.delete(w1,index)
w2 = np.delete(w2,index)

# Ajusta o shape
w1 = w1.reshape(10,1)
w2 = w2.reshape(10,1)

# Extrair os recursos de dados de teste
test_f1 = x_test[:,0]
test_f2 = x_test[:,1]

test_f1 = test_f1.reshape(10,1)
test_f2 = test_f2.reshape(10,1)

# Previsões
y_pred = w1 * test_f1 + w2 * test_f2
predictions = []
for val in y_pred:
    if(val > 1):
        predictions.append(1)
    else:
        predictions.append(-1)

print("\nAcurácia nos Dados de Teste (Algoritmo Personalizado)", accuracy_score(y_test, predictions))


# Testando agora com o algoritmo pronto do Scikit-Learn
clf = SVC(kernel='linear')
clf.fit(x_train,y_train)
y_pred = clf.predict(x_test)
print("\nAcurácia nos Dados de Teste (Algoritmo Pronto do Scikit-Learn)",accuracy_score(y_test, y_pred))
print("\n")