# -*- coding: utf-8 -*-
# @ author hulei 2016-5-3
from numpy import *
import operator
from os import listdir

import sys
reload(sys)
sys.setdefaultencoding('utf8')

# x,y=getDataSet_dz('iris.data.txt',4)
    
def getDataSet(filename,numberOfFeature):       #將資料集讀入記憶體  
    fr = open(filename)
    numberOfLines = len(fr.readlines())         #get the number of lines in the file  file.readlines()是把檔案的全部內容讀到記憶體，並解析成一個list
    returnMat = zeros((numberOfLines,numberOfFeature))        #prepare matrix to return  3代表資料集中特徵數目###
    classLabelVector = []                       #prepare labels return   
    fr = open(filename)
    index = 0
    for line in fr.readlines():
        line = line.strip()                    #strip() 引數為空時，預設刪除空白符（包括'\n', '\r',  '\t',  ' ')
        listFromLine = line.split(',')         #split 以什麼為標準分割一次  分成陣列中的每個元素
        returnMat[index,:] = listFromLine[0:numberOfFeature] 
        #classLabelVector.append(int(listFromLine[-1]))   #append() 方法向列表的尾部新增一個新的元素
        if listFromLine[-1] == 'Iris-setosa' :
            classLabelVector.append(1)
        elif listFromLine[-1] == 'Iris-versicolor' :
            classLabelVector.append(2)
        else:
        #elif listFromLine[-1] == 'Iris-virginica' :
            classLabelVector.append(3)
        index += 1
    return returnMat,classLabelVector
    
def getDataSet_dz(filename,numberOfFeature):    #改進版，可以消除資料中的空白行
    numberOfLines = 0
    mx = []     #將資料集 去除空行後存入
    fr = open(filename)
    for line in fr.readlines():      
        line = line.strip()   
        if line != ''   :    #去除空白行    
            numberOfLines+=1
            mx.append( line.split(',') )
    returnMat = zeros((numberOfLines,numberOfFeature))
    classLabelVector = [] 
    for index in range(numberOfLines)   :
        returnMat[index,:] = mx[index][0:numberOfFeature] 
        if mx[index][-1] == 'Iris-setosa' :
            classLabelVector.append(1)
        elif mx[index][-1] == 'Iris-versicolor' :
            classLabelVector.append(2)
        else:
        #elif listFromLine[-1] == 'Iris-virginica' :
            classLabelVector.append(3)
    return returnMat,classLabelVector