import numpy as np
import matplotlib.pyplot as plt
from sklearn import neighbors
import pandas as pd
import sys
import os

data = pd.read_csv(
    "C:/Users/Piotr/Desktop/dane.csv",
    sep=';',
    usecols=['X', 'Y'],
    dtype={"X": np.float64, "Y": np.float64}
)

X = data["X"].tolist()
Y = data["Y"].tolist()

test=[]
test2=[]

for elem in X:
    tmp=[]
    tmp.append(elem)
    test.append(tmp)

for elem in Y:
    tmp=[]
    tmp.append(elem)
    test2.append(tmp)

X = test
y = test2

T = np.linspace(1900, 2020, 1000)[:, np.newaxis]

n_neighbors = 5

plt.figure(figsize=(20,20))

knn = neighbors.KNeighborsRegressor(n_neighbors, weights='uniform')
y_ = knn.fit(X, y).predict(T)
plt.subplot(4, 1, 1)
plt.scatter(X, y, color='darkorange', label='data')
plt.plot(T, y_, color='navy', label='prediction')
plt.axis([1899, 2018, -0.7, 1.4])
plt.legend(prop={'size': 30})
plt.title("KNeighborsRegressor (k = %i, weights = '%s')" % (n_neighbors, weights), fontsize=30)

knn = neighbors.KNeighborsRegressor(n_neighbors, weights='distance')
y_ = knn.fit(X, y).predict(T)
plt.subplot(4, 1, 2)
plt.scatter(X, y, color='darkorange', label='data')
plt.plot(T, y_, color='navy', label='prediction')
plt.axis([1899, 2018, -0.7, 1.4])
plt.legend(prop={'size': 30})
plt.title("KNeighborsRegressor (k = %i, weights = '%s')" % (n_neighbors, weights), fontsize=30)

T = np.linspace(2010, 2016, 1000)[:, np.newaxis]
knn = neighbors.KNeighborsRegressor(n_neighbors, weights='uniform')
y_ = knn.fit(X, y).predict(T)
plt.subplot(4, 1, 3)
plt.scatter(X, y, color='darkorange', label='data')
plt.plot(T, y_, color='navy', label='prediction')
plt.axis([2010, 2016, -0.7, 1.4])
plt.legend(prop={'size': 30})
plt.title("KNeighborsRegressor (k = %i, weights = '%s')" % (n_neighbors, weights), fontsize=30)

T = np.linspace(2010, 2016, 1000)[:, np.newaxis]
knn = neighbors.KNeighborsRegressor(n_neighbors, weights='distance')
y_ = knn.fit(X, y).predict(T)
plt.subplot(4, 1, 4)
plt.scatter(X, y, color='darkorange', label='data')
plt.plot(T, y_, color='navy', label='prediction')
plt.axis([2010, 2016, -0.7, 1.4])
plt.legend(prop={'size': 30})
plt.title("KNeighborsRegressor (k = %i, weights = '%s')" % (n_neighbors, weights), fontsize=30)

plt.tight_layout()
plt.show()