task9

"""
    绘制树图形

"""
import pandas as pd
#from task8 import createTree
import numpy as np
import matplotlib.pyplot as plt
from collections import Counter

def plotNode(text, centerPt, parentPt, nodeType):
    """
        绘制注解，带箭头
        annotate 函数用来绘制注解
        parentPt： 父节点位置
        centerPt: 被指向的位置
        nodeType: 节点类型
    """
    createPlot.ax1.annotate(text, xy=parentPt,  xycoords='axes fraction',
    xytext=centerPt, textcoords='axes fraction',
    va="center", ha="center", bbox=nodeType, arrowprops= {'arrowstyle': '<-'} )


# 内部节点文本框样式
decisionNode = dict(boxstyle="sawtooth", fc="0.8")
# 叶节点文本框样式
leafNode = dict(boxstyle="round4", fc="0.8")

def getNumLeafs(myTree):
    """
        获取叶节点的数目
        确定横轴x的长度
    """
    numLeafs = 0
    firstStr = list(myTree.keys())[0]
    secondDict = myTree[firstStr]
    for key in secondDict:
        if isinstance(secondDict[key], dict):
            numLeafs += getNumLeafs(secondDict[key])
        else:
            numLeafs +=1
    return numLeafs


def getTreeDepth(myTree):
    """
        获取树的深度
        确定纵轴y的长度
    """
    maxDepth = 0
    firstStr = list(myTree.keys())[0]
    secondDict = myTree[firstStr]
    for key in secondDict:
        if isinstance(secondDict[key], dict):
            thisDepth = 1 + getTreeDepth(secondDict[key])
        else:
            thisDepth = 1
        if thisDepth > maxDepth:
            maxDepth = thisDepth
    return maxDepth


def createPlot(tree):
    """
        创建画布
    """
    fig = plt.figure(1, facecolor='gray')
    fig.clf()
    axprops = dict(xticks=[], yticks=[])
    createPlot.ax1 = plt.subplot(111, frameon=False, **axprops)
    plotTree.totalW = float(getNumLeafs(tree))
    plotTree.totalD = float(getTreeDepth(tree))
    plotTree.xOff = -0.5/plotTree.totalW
    plotTree.yOff = 1.0
    plotTree(tree, (0.5,1.0), '')
    plt.show()


def plotMidText(cntrPt, parentPt, txtString):
    """
        连线添加文字
    """
    xMid = (parentPt[0]-cntrPt[0])/2.0 + cntrPt[0]
    yMid = (parentPt[1]-cntrPt[1])/2.0 + cntrPt[1]
    createPlot.ax1.text(xMid, yMid, txtString, va="center", ha="center", rotation=0)


def plotTree(myTree, parentPt, text):
    """
        绘制决策树
    """
    numLeafs = getNumLeafs(myTree)
    firstStr = list(myTree.keys())[0]
    cntrPt = (plotTree.xOff + (1.0 + float(numLeafs))/2.0/plotTree.totalW, plotTree.yOff)
    plotMidText(cntrPt, parentPt, text)
    plotNode(firstStr, cntrPt, parentPt, decisionNode)
    secondDict = myTree[firstStr]
    plotTree.yOff = plotTree.yOff - 1.0/plotTree.totalD
    for key in secondDict:
        if isinstance(secondDict[key], dict):
            plotTree(secondDict[key],cntrPt,str(key))
        else:
            plotTree.xOff = plotTree.xOff + 1.0/plotTree.totalW
            plotNode(secondDict[key], (plotTree.xOff, plotTree.yOff), cntrPt, leafNode)
            plotMidText((plotTree.xOff, plotTree.yOff), cntrPt, str(key))
    plotTree.yOff = plotTree.yOff + 1.0/plotTree.totalD

def calcShannonEnt(dataSet):
    """ 计算信息熵
    """
    # 获取最后一列的数据
    labels = dataSet[:,-1]
    # 统计所有类别对应出现的次数
    labelCounts = Counter(labels)
    # 数据已准备好，计算熵
    shannonEnt = 0.0
    dataLen = len(dataSet)
    for key in labelCounts:
        pro = labelCounts[key] / dataLen
        shannonEnt -= pro * np.log2(pro)
    return shannonEnt

def chooseFeature(dataSet):
    """
        选择最优属性
        gain = baseEntropy - newEntropy
    """
    baseEntropy = calcShannonEnt(dataSet) # 整个数据集的信息熵
    bestInfoGain = 0.0
    bestFeature = -1
    # 遍历所有属性
    for i in range(len(dataSet[0]) -1):
        splitDict = Counter(dataSet[:, i])
        newEntropy = 0.0
        for v in splitDict:
            subDataSet = dataSet[dataSet[:, i]==v]
            prob = splitDict[v]/float(len(dataSet))
            newEntropy += prob * calcShannonEnt(subDataSet)
        gain = baseEntropy - newEntropy
        if gain > bestInfoGain:
            bestInfoGain = gain
            bestFeature = i
    return bestFeature

def createTree(dataSet, feature_labels):
    """
        生成决策树
        返回字典树
        dataSet: 数据集
        feature_labels： 属性标签
    """
    labels = dataSet[:, -1]
    # 数据集样本类别相同
    if len(set(labels)) == 1:
        return labels[0]
    # 属性值为空或者唯一属性值相同，返回样本数最多的类别
    if len(dataSet[0]) == 1 or (len(dataSet[0]) == 2 and len(set(dataSet[:, 0])) == 1):
        resDict = dict(Counter(labels))
        sortedClassCount = sorted(resDict.items(), key=lambda item: item[1], reverse=True)
        return sortedClassCount[0][0]

    # 选择最优属性
    bestFeat = chooseFeature(dataSet)
    bestFeatLabel = feature_labels[bestFeat]
    myTree = {bestFeatLabel:{}}
    del(feature_labels[bestFeat])
    # 对选择属性进行划分
    for v in Counter(dataSet[:, bestFeat]):
        # 划分后的子集不应该包含我们选择属性对应的列
        subDataSet = np.delete(dataSet[dataSet[:, bestFeat]==v], bestFeat, axis=1)
        subLabels = feature_labels[:]
        # 生成子树
        myTree[bestFeatLabel][v] = createTree(subDataSet,subLabels)
    return myTree

if __name__ == '__main__':
    # 读取数据集，这里忽略 ids 及 连续属性列
    dataset = pd.read_csv("watermelon_3a.csv", usecols=['color', 'root', 'knocks', 'texture', 'navel', 'touch', 'label'])
    feature_labels = list(dataset.columns)
    dataset = dataset.values
    res = createTree(dataset, feature_labels)
    print(res)
    createPlot(res)