from sklearn import tree, ensemble, neighbors, svm, linear_model

import math

import numpy as np

import pickle

def GenGlobal(errors, sizes):

errors = np.array(errors)

sizes = np.array(sizes)

errors = np.reshape(errors, (sizes.shape[0], sizes.shape[0]))

weights = []

sortedErrors = np.sort(errors, axis = 1)

mydict = dict()

rows = errors.shape[0]

cols = errors.shape[1]

for i in range(rows):

firstValue = 1

if(sortedErrors[i][0] == -np.inf):

firstValue = 0

for j in range(cols):

key = mydict.get(str(sortedErrors[i][j]))

if key == None:

mydict[str(sortedErrors[i][j])] = firstValue

firstValue += 1

for j in range(cols):

errors[i][j] = mydict[str(errors[i][j])]

mydict.clear()

for i in range(len(sizes)):

currWeight = 0.0

if sizes[i] != 0.0:

for j in range(cols):

if(errors[i][j] != 0):

currWeight += math.pow(2, -errors[i][j])

currWeight *= sizes[i]

weights.append(currWeight)

weights = np.array(weights)

weights[weights < np.median(weights)] = 0.0

weights = weights / np.sum(weights)

return weights