import pylab

import numpy as np

import math

import sys

datax = [map(int,i.strip().split()) for i in open(sys.argv[1]).readlines()]

dt = [i for i in open(sys.argv[2]).readlines()]

m=len(dt)

datay=[(0 if dt[i]=='Alaska\n' else 1) for i in range(m)]

X=np.asarray(datax)

Y1=np.asarray(datay)

Y=Y1.reshape((Y1.size,1))

sum1=sum(Y[Y==1])

phi=1.0*sum1/m

X0=X[np.ix_(Y1==0,)]

X1=X[np.ix_(Y1==1,)]

mu0=1.0*(sum(X0))/(m-sum1)

mu1=1.0*(sum(X1))/sum1

mu0=mu0.reshape((1,2))

mu1=mu1.reshape((1,2))

sigma0=np.zeros((2,2))

sigma1=np.zeros((2,2))

for x in X0:

sigma0=sigma0+np.dot((x-mu0).T,(x-mu0))

for x in X1:

sigma1=sigma1+np.dot((x-mu1).T,(x-mu1))

sigma=1.0*(sigma0+sigma1)/m

# print(phi)

# print(mu0)

# print(mu1)

# print(sigma)

left=min(X[:,0])

right=max(X[:,0])

siginv=np.linalg.inv(sigma)

num=np.dot(mu1,(np.dot(siginv,mu1.T)))-np.dot(mu0,(np.dot(siginv,mu0.T)))- 2*(math.log((phi)/(1-phi)))

num=num/2

mat=np.dot((mu1-mu0),siginv)

lefty=(np.asscalar(num-mat[0][0]*left)/mat[0][1])

righty=np.asscalar((num-mat[0][0]*right)/mat[0][1])

plotx=[left,right]

ploty=[lefty,righty]

pylab.scatter(X0[:,0], X0[:,1], s=10, label='Alaska', c='r', marker="^")

pylab.scatter(X1[:,0], X1[:,1], s=10, label='Canada', c='b', marker=".")

# pylab.plot(datax,datay,[Y==0]'o')

pylab.plot(plotx,ploty, '-k')

pylab.show()

# pylab.savefig('c.png')