#coding:utf-8

from numpy import *
import operator
from collections import Counter
import matplotlib
import matplotlib.pyplot as plt


###匯入特徵資料
def file2matrix(filename):
    fr = open(filename)
    contain = fr.readlines()###讀取檔案的所有內容
    count = len(contain)
    returnMat = zeros((count,3))
    classLabelVector = []
    index = 0
    for line in contain:
        line = line.strip() ###擷取所有的回車字元
        listFromLine = line.split('\t')
        returnMat[index,:] = listFromLine[0:3]###選取前三個元素，儲存在特徵矩陣中
        classLabelVector.append(listFromLine[-1])###將列表的最後一列儲存到向量classLabelVector中
        index += 1
    
    ##將列表的最後一列由字串轉化為數字，便於以後的計算
    dictClassLabel = Counter(classLabelVector)
    classLabel = []
    kind = list(dictClassLabel)
    for item in classLabelVector:
        if item == kind[0]:
            item = 1
        elif item == kind[1]:
            item = 2
        else:
            item = 3
        classLabel.append(item)
    return returnMat,classLabel#####將文字中的資料匯入到列表

##繪圖（可以直觀的表示出各特徵對分類結果的影響程度）
datingDataMat,datingLabels = file2matrix('D:\python\Mechine learing in Action\KNN\datingTestSet.txt')
fig = plt.figure()
ax = fig.add_subplot(111)
ax.scatter(datingDataMat[:,0],datingDataMat[:,1],15.0*array(datingLabels),15.0*array(datingLabels))
plt.show()

## 歸一化資料,保證特徵等權重
def autoNorm(dataSet):
    minVals = dataSet.min(0)
    maxVals = dataSet.max(0)
    ranges = maxVals - minVals
    normDataSet = zeros(shape(dataSet))##建立與dataSet結構一樣的矩陣
    m = dataSet.shape[0]
    for i in range(1,m):
        normDataSet[i,:] = (dataSet[i,:] - minVals) / ranges
    return normDataSet,ranges,minVals

##KNN演算法
def classify(input,dataSet,label,k):
    dataSize = dataSet.shape[0]
    ####計算歐式距離
    diff = tile(input,(dataSize,1)) - dataSet
    sqdiff = diff ** 2
    squareDist = sum(sqdiff,axis = 1)###行向量分別相加，從而得到新的一個行向量
    dist = squareDist ** 0.5
    
    ##對距離進行排序
    sortedDistIndex = argsort(dist)##argsort()根據元素的值從大到小對元素進行排序，返回下標

    classCount={}
    for i in range(k):
        voteLabel = label[sortedDistIndex[i]]
        ###對選取的K個樣本所屬的類別個數進行統計
        classCount[voteLabel] = classCount.get(voteLabel,0) + 1
    ###選取出現的類別次數最多的類別
    maxCount = 0
    for key,value in classCount.items():
        if value > maxCount:
            maxCount = value
            classes = key
    return classes

##測試(選取10%測試）
def datingTest():
    rate = 0.10
    datingDataMat,datingLabels = file2matrix('D:\python\Mechine learing in Action\KNN\datingTestSet.txt')
    normMat,ranges,minVals = autoNorm(datingDataMat)
    m = normMat.shape[0]
    testNum = int(m * rate)
    errorCount = 0.0
    for i in range(1,testNum):
        classifyResult = classify(normMat[i,:],normMat[testNum:m,:],datingLabels[testNum:m],3)
        print("分類後的結果為:,", classifyResult)
        print("原結果為：",datingLabels[i])
        if(classifyResult != datingLabels[i]):
                                  errorCount += 1.0
    print("誤分率為:",(errorCount/float(testNum)))
                                  
###預測函式
def classifyPerson():
    resultList = ['一點也不喜歡','有一丟丟喜歡','灰常喜歡']
    percentTats = float(input("玩視訊所佔的時間比?"))
    miles = float(input("每年獲得的飛行常客里程數?"))
    iceCream = float(input("每週所消費的冰淇淋公升數?"))
    datingDataMat,datingLabels = file2matrix('D:\python\Mechine learing in Action\KNN\datingTestSet2.txt')
    normMat,ranges,minVals = autoNorm(datingDataMat)
    inArr = array([miles,percentTats,iceCream])
    classifierResult = classify((inArr-minVals)/ranges,normMat,datingLabels,3)
    print("你對這個人的喜歡程度:",resultList[classifierResult - 1])