import pandas as pd
from sklearn.neighbors import  KNeighborsClassifier
from sklearn.model_selection import train_test_split
from sklearn import preprocessing
import sklearn.metrics as met
from termcolor import colored

def prediciton_info(x_train, y_train,  x_test, y_test, y_pred, indices, distances):

    k=len(indices[0])
    for i in range(0, len(x_test)):
        print(colored("test_point: ", "blue"),
              colored(x_test.iloc[i: (i + 1)], "blue"), sep="\n")

        if (y_test.iloc[i] == y_pred[i]):
            color = "green"
        else:
            color = "red"

        print("actual: ",  colored(y_test.iloc[i], color))
        print("predicted: ", colored(y_pred[i], color))
        print("\n", colored("neighbours: ", "blue"))
        for j in range(0, k):
            print(x_train.iloc[indices[i][j]: (indices[i][j] + 1)])
            print("class:", colored(y_train.iloc[indices[i][j]], "yellow"))
            print("distance: ", distances[i][j], "\n")

        print("\n")


def class_info(clf, x_train, y_train, x_test, y_test):

    clf.fit(x_train, y_train)
    distances, indices = clf.kneighbors(x_test)
    y_pred = clf.predict(x_test)

    cnf_matrix = met.confusion_matrix(y_test, y_pred)
    print("Matrica konfuzije", cnf_matrix, sep="\n")
    print("\n")

    accuracy = met.accuracy_score(y_test, y_pred)
    print("Preciznost", accuracy)
    print("\n")

    class_report = met.classification_report(y_test, y_pred, target_names=df["Species"].unique())
    print("Izvestaj klasifikacije", class_report, sep="\n")

    option = input("Da li zelite informacije o klasifikacije svake instance? (1 za da, 0 za ne)")
    print(option)
    if(option=="1"):
        prediciton_info(x_train, y_train, x_test, y_test, y_pred, indices, distances)


df = pd.read_csv("iris_pandas.csv")

featurs = df.columns[:4].tolist()

#ako zelimo da izdvojimo odredjene atribute za klasifikaciju
#featurs = ["Petal_Length",  "Petal_Width"]

x_original=df[featurs]

#standardizacija atributa
x=pd.DataFrame(preprocessing.scale(x_original))


#normalizacija
#x=pd.DataFrame(preprocessing.MinMaxScaler().fit_transform(x_original))

#dodeljivanje imena kolonama
x.columns = featurs
y=df["Species"]


#podela na trening i test skup
x_train, x_test, y_train, y_test = train_test_split(x, y, train_size=0.8, stratify=y)

"""
Paramerti za KNN:
n_neighbors : broj suseda
              default=5
weights : tezine suseda
              default='uniform'
              moguce:
              'uniform' : svi susedi imaju podjednak uticaj
              'distance' : blizi susedi imaju veci uticaj na odredjivanje klase
algorithm:
          default: 'auto'
          moguce:
          'brute'
          'kd_tree'
          'ball_tree'
          'auto'
leaf_size: velicina listova u drvetu (za 'kd_tree' i 'ball_tree')

metric:  metrika
       default : 'minkowski'
p: parametar za Minkowski rastojanje (p=1 za Menhetn, p=2 za Euklidsko)
"""

k_values = range(3,10)
p_values = [1, 2]
weights_values = ['uniform', 'distance']

for k in k_values:
    for p in p_values:
        for weight in weights_values:
            clf = KNeighborsClassifier(n_neighbors=k,
                                        p=p,
                                        weights=weight)

            print(colored("k="+ str(k), "blue"))
            print(colored("p="+str(p), "blue"))
            print(colored("weight=" + weight, "blue") )

            class_info(clf, x_train, y_train, x_test, y_test)