# Import packages
%matplotlib inline
import matplotlib.pyplot as plt
import seaborn as sns; sns.set()  # for plot styling
import numpy as np

#need to average the x and y that are cloest. this becomes my new cluster point
# Generate Samples
from sklearn.datasets.samples_generator import make_blobs
X, y_true = make_blobs(n_samples=300, centers=4,
                       cluster_std=0.60, random_state=0)
plt.scatter(X[:, 0], X[:, 1], s=50);
x = X[:, 0]
y = X[:, 1]
#print(x)
#print(y)
#print(len(y))
#print(len(x))

###############################################
# YOUR CODE GOES HERE
# Put some code to find clusters here
# Assign the clusters and labels in your code
###############################################
#print(range(len(X)))

# Uncomment to display clusters and cluster centers
#plt.scatter(X[:, 0], X[:, 1], c=labels,
#            s=50, cmap='viridis');
#plt.scatter(centers[:, 0], centers[:, 1], c='black', s=200, alpha=0.5);

k = 4
# k_mean_x_curr = np.zeros((k, k))
k_mean_curr = np.zeros((k, 2))
k_mean_y_curr = np.zeros((k, k))
#print(k_mean_x_curr)

k_mean_prev = np.zeros((2, k))
#print(k_mean_curr)


#Initialize cluster centroids

centers = np.zeros((k, 2))
for i in range(k):
    centers[i,0] = x[int(len(X)/k*i)]
    centers[i,1] = y[int(len(X)/k*i)]
#print(centers)

#Plot initial clusters
plt.scatter(centers[:, 0], centers[:, 1], c='black', s=100, alpha=0.5);
#plt.show()

#Calculate euclidean distance from centroids
dist = np.zeros((len(X), k+2))

for j in range(len(X)): # for each point
    for i in range(k): # for each centroid

        dist[j][i] = np.sqrt((x[j] - centers[i][0])**2 + (y[j] - centers[i][1])**2)
    # populate minimum distance vector
    dist[j][4] = list(dist[j]).index(min(dist[j][0], dist[j][1], dist[j][2], dist[j][3]))

print(dist)
#Update Cluster Means
for j in range(len(X)):
    n = int(dist[j][k+1])
    #print(n)
    #print(x[j])
    k_mean_x_curr[0][n] = x[j] + k_mean_x_curr[0][n] #x sum
    k_mean_y_curr[0][n] = y[j] + k_mean_x_curr[0][n] #y sum

    k_mean_x_curr[1][n] += 1 #x count
    k_mean_y_curr[1][n] += 1 #y count

# #Calculate mean
for i in range(k):
    if(k_mean_x_curr[1][i]==0):
      k_mean_x_curr[2][i] = 0
    else:
      k_mean_x_curr[2][i] = k_mean_x_curr[0][i] / k_mean_x_curr[1][i]

    if(k_mean_y_curr[1][i]==0):
      k_mean_y_curr[2][i] = 0
    else:
      k_mean_y_curr[2][i] = k_mean_y_curr[0][i] / k_mean_y_curr[1][i]


# print(k_mean_x_curr)
# print(k_mean_y_curr)



#Update centers array with new means
for i in range(k):
#     #print('before: ',centers[i][0])
#     #print(centers[i][1])
    centers[i][0] = k_mean_x_curr[2][i] #x
    centers[i][1] = k_mean_y_curr[2][i] #y
#print('after: ',centers[i][0])
#print(centers[i][1])

plt.scatter(centers[:, 0], centers[:, 1], c='red', s=100, alpha=0.5);
#plt.show()