def __init__(self):
filePath = "F:/test/workload.xlsx"
wrFile = WRFile()
self.workload = wrFile.readDataFromExcel(filePath=filePath,
sheet_name="1")
data = self.evaluateBurst()
data = data / np.max(data)
wrFile.writeDataIntoExcel(data=data, filePath="F:/test/avgsampEn.xlsx")
def analyzePredictionPrecision():
objFileName = "svr_rbf"
precision = []# 列表头分别为 ratio,max,mean
wrFile = WRFile()
objFilePath = "F:\\FIFA\\predict\\"+objFileName+"/precision_over.xlsx"
predict = wrFile.readDataFromExcel(objFilePath,min_cols = 1,max_cols = 3)
predict = predict.reshape(3,8)
result = [np.average(predict[0]),np.average(predict[1]),np.average(predict[2])]
wrFile.writeDataIntoExcel(data = result,filePath = "F:\\FIFA\\predict\\"+objFileName+"/precision_over_evaluate.xlsx" )
def analyzeStats():
qList = [2, 3, 4, 5, 6]
wrFile = WRFile()
result = []
for q in qList:
fileName = "F:\data\experiment/Delay_SQ_q" + str(q) + ".xlsx"
data = wrFile.readDataFromExcel(filePath=fileName,
sheet_name="1",
min_cols=4,
max_cols=4)
r = getStatisticAttribute(data)
result.append(r)
#print("ATBM is",result)
wrFile.writeDataIntoExcel(
result, filePath="F:\data\experiment/Delay_SQ_stats.xlsx")
def compareModel():
fileStart = "F:\\one\\final\\inmin/workload"
fileEnd = "inmin.xlsx"
wrFile = WRFile()
#fft = FFTPredict()
#noBurst_pre = fft.FFTofNoBurst(data51 , data52 )
#WFD_pre = fft.FFTofWFD(data51 , data52 )
periods = 10
for i in range(53, 61):
result = TestGreyModel(periods, filePath=fileStart + str(i) + fileEnd)
predict = result[1]
print(i)
wrFile.writeDataIntoExcel(data=predict,
filePath="F:/one/predict/GM/wc" + str(i) +
".xlsx")
def analyzeResult(q, better, worse):
axeRange = q
record = []
print(axeRange)
for x in range(axeRange):
for y in range(axeRange):
for z in range(axeRange):
if better[x][y][z] > 0 or worse[x][y][z] > 0:
print()
#print("better:"+str(better[x][y][z]))
#print("worse:"+str(worse[x][y][z]))
record.append([
str(x) + "<c1<=" + str(x + 1) + "," + str(y) +
"<c2<=" + str(y + 1) + "," + str(z) + "<c3<=" +
str(z + 1), better[x][y][z], worse[x][y][z]
])
#print("=======================")
wrFile = WRFile()
filePath = "F:\\data\\experiment/Delay_distribute_q" + str(q) + ".xlsx"
wrFile.writeDataIntoExcel(data=record, filePath=filePath)
def splitOnlineData():
shopName = ["京东商城", "天猫商城", "亚马逊商城", "淘宝商城"]
#读出所有数据，然后都存放到
# 1.生成日期数据
wrFile = WRFile()
data = wrFile.readDataFromExcel(filePath="F:\\data\\orginal\\online.xlsx",
sheet_name="online",
cols=3)
#2. 对数据进行切片
jingdong = []
tmall = []
amazon = []
taobao = []
for i in range(len(data)):
count = i % 4
if count == 0:
if data[i] > 0:
jingdong.append(data[i])
else:
jingdong.append(data[i - 4])
elif count == 1:
if data[i] > 0:
tmall.append(data[i])
else:
tmall.append(data[i - 4])
elif count == 2:
if data[i] > 0:
amazon.append(data[i])
else:
amazon.append(data[i - 4])
elif count == 3:
if data[i] > 0:
taobao.append(data[i])
else:
taobao.append(data[i - 4])
#3.将数据写入excel
fileRoot = "F:\\data\\online/"
filesuffix = ".xlsx"
wrFile.writeDataIntoExcel(data=jingdong,
filePath=fileRoot + "jingdong" + filesuffix)
wrFile.writeDataIntoExcel(data=tmall,
filePath=fileRoot + "tmall" + filesuffix)
wrFile.writeDataIntoExcel(data=amazon,
filePath=fileRoot + "amazon" + filesuffix)
wrFile.writeDataIntoExcel(data=taobao,
filePath=fileRoot + "taobao" + filesuffix)
c = moveDown(c)
result = a[i]
if result == 0:
equal[c[0]][c[1]][c[2]] += 1
elif result == 1:
better[c[0]][c[1]][c[2]] += 1
else:
worse[c[0]][c[1]][c[2]] += 1
#print("better is",better)
#print("worse is",worse)
#print("equal is",equal)
analyzeResult(q, better, worse)
#divideSpace(q)
#getStatisticAttribute(q)
#analyzeStats()
wrFile = WRFile()
data = wrFile.readDataFromExcel(
filePath="D:\\cloudsim\\log\\workload1/taobao.xlsx",
min_cols=1,
max_cols=1,
sheet_name="1")
data = np.floor((np.array(data) / 100))
#print(data)
wrFile.writeDataIntoExcel(
data=data, filePath="D:\\cloudsim\\log\\workload1/deplete_taobao.xlsx")
print(np.percentile(np.array(data), 80))
class ClusterData:
def __init__(self):
self.wrFile = WRFile()
'''
timestamp second,new_timestamp
["2016-11-17 06:08:42.692"]===>[42,2016-11-17 06:08:42]
["2016-11-17 06:08:42"]===>[42,2016-11-17 06:08:42]
'''
def getSecond_timestamp(self, timestamp):
loc = timestamp.rindex(":")
try:
dot_loc = timestamp.rindex(".")
new_item = timestamp[:dot_loc]
second = timestamp[loc + 1:dot_loc]
except ValueError:
new_item = timestamp
second = timestamp[loc + 1:]
return {"second": second, "timestamp": new_item}
def transformQtoXYZ(self, q_x, q_y, q_z, q_w):
x = round(2 * q_x * q_z + 2 * q_y * q_w, 2)
y = round(2 * q_y * q_z - 2 * q_x * q_w, 2)
z = round(1 - 2 * q_x * q_x - 2 * q_y * q_y, 2)
return [x, y, z]
def mergeDataInSeconds(self, num, video_num):
path = "F:\\project\\dataset\\vr\\Formated_Data\\Experiment_1/" + str(
num) + "/VD_1.xlsx"
#readDataFromExcel(self,filePath,cols=1,sheet_name = "1",min_col =1,max_col = 1)
#readDataFromExcel2(self,filePath,rows = 1,sheet_name = "1",min_row =1,max_row = 1):
result = self.wrFile.readDataFromExcel2(path,
rows=2,
sheet_name="Sheet",
max_row=0)
#Timestamp PlaybackTime UnitQuaternion.x UnitQuaternion.y UnitQuaternion.z UnitQuaternion.w HmdPosition.x HmdPosition.y HmdPosition.z
max_rows = len(result) #最大行-1
min_rows = 1
count = 1
sum_q_x = float(result[1][2])
sum_q_y = float(result[1][3])
sum_q_z = float(result[1][4])
sum_q_w = float(result[1][5])
time_xyz = []
for r in range(min_rows + 1, max_rows):
item = result[r]
item = numpy.array(item[2:], dtype="float64")
current_time = self.getSecond_timestamp(result[r][0])
last_time = self.getSecond_timestamp(result[r - 1][0])
if (current_time["second"] == last_time["second"]):
#print("same")
sum_q_x = sum_q_x + item[0]
sum_q_y = sum_q_y + item[1]
sum_q_z = sum_q_z + item[2]
sum_q_w = sum_q_w + item[3]
count = count + 1
else:
average_q_x = round(sum_q_x / count, 2)
average_q_y = round(sum_q_y / count, 2)
average_q_z = round(sum_q_z / count, 2)
average_q_w = round(sum_q_w / count, 2)
xyz = self.transformQtoXYZ(average_q_x, average_q_y,
average_q_z, average_q_w)
#print(xyz)
count = 1
sum_q_x = item[0]
sum_q_y = item[1]
sum_q_z = item[2]
sum_q_w = item[3]
temp = [current_time["timestamp"]]
temp.extend(xyz)
time_xyz.append(temp)
#结算最后一秒
average_q_x = round(sum_q_x / count, 2)
average_q_y = round(sum_q_y / count, 2)
average_q_z = round(sum_q_z / count, 2)
average_q_w = round(sum_q_w / count, 2)
xyz = self.transformQtoXYZ(average_q_x, average_q_y, average_q_z,
average_q_w)
timestamp = [
self.getSecond_timestamp(result[max_rows - 1][0])["timestamp"]
]
timestamp.extend(xyz)
time_xyz.append(timestamp)
#(self,data,filePath,describe ="no description",cols = 0,sheet_name = "1")#
rows = self.wrFile.writeDataIntoExcel(
data=time_xyz,
filePath="F:\\project\\dataset\\vr\\Formated_Data\\Experiment_1/" +
str(num) + "/u" + str(num) + "_v" + str(video_num) +
"_second.xlsx")
return rows
'''
data = 48*204*3
48 stands for the total number of users
204 represents there are 204 seconds of this video
3 describe 3 dimension data needs to be recorded
'''
def mergeDataFromUsers(self, video):
data = numpy.zeros((48, 204, 3))
#print(data)
'''=====store the data of users in a multi-dimension matrix Data====='''
for user in range(1, 48 + 1):
filePath = "F:\\project\\dataset\\vr\\Formated_Data\\Experiment_1/" + str(
user) + "/u" + str(user) + "_v" + str(
video) + "_second.xlsx" #u1_v1_second.xlsx
user_data = numpy.array(self.wrFile.readDataFromExcel2(filePath))
#print(user_data.shape)
user_rows = len(user_data)
row_standard = 204
if (user_rows >= row_standard):
data[user - 1, 0:row_standard] = user_data[:row_standard, 1:]
else: #user_rows<row_standard needs to be added some rows
data[user - 1, :user_rows] = user_data[:, 1:]
for i in range(1, row_standard - user_rows + 1):
data[user - 1,
user_rows - 1 + i, :] = user_data[user_rows - 1, 1:]
'''====================================='''
#print(data[18,:,:])
second_data = data[:, 0, :]
for r in range(1, 204):
#if r==18:
#print(data[:,r,:])
second_data = numpy.hstack((second_data, data[:, r, :]))
print(second_data.shape)
self.wrFile.writeDataIntoExcel(
data=second_data,
filePath=
"F:\\project\\dataset\\vr\\Formated_Data\\Experiment_1/second.xlsx"
)
def predictMotion():
pass
def getClassifer(self, Bmax):
#1.randomly select 43 samples of the data, and
#2.depend on samples to train classfier
filePath = "F:\\project\\dataset\\vr\\Formated_Data\\Experiment_1/cluster_data.xlsx"
data = numpy.array(
self.wrFile.readDataFromExcel2(filePath=filePath, max_row=48))
accuracy = numpy.zeros(203).reshape((203, 1))
for col in range(0, 203): # 应该从后往前选
X = data[:, col * 4:(col + 1) * 4]
if col >= 199:
label = numpy.array(data[:, col * 4 + 3 + 1 * 4]).reshape(
(48, 1))
else:
label = numpy.array(data[:, col * 4 + 3 + Bmax * 4]).reshape(
(48, 1))
# get samples
whole_index = range(0, 48)
sample_index = ran.sample(whole_index, 43)
test_index = list(set(whole_index) - set(sample_index))
sample_data = X[sample_index][:, :3]
sample_label = label[sample_index]
test_data = X[test_index][:, :3]
test_label = label[test_index]
#carry out training
try:
clf = SVC(gamma='auto')
clf.fit(sample_data, sample_label)
accuracy_temp = clf.score(test_data, test_label)
accuracy[col] = accuracy_temp
except ValueError:
accuracy[col] = 1
#print(accuracy.shape)
filePath = "F:\\project\\dataset\\vr\\Formated_Data\\Experiment_1/prediction_accuracy.xlsx"
self.wrFile.writeDataIntoExcel(data=accuracy, filePath=filePath)
def plotScatter(self, x, y, z, slot):
plt.scatter(x=x, y=y, c=z)
plt.xlabel("x", fontsize=20)
plt.ylabel("y", fontsize=20)
plt.legend("z", fontsize=20)
plt.show()
plt.savefig(
"F:\\project\\dataset\\vr\\Formated_Data\\Experiment_1\\image/xyz"
+ str(slot) + ".png")
def plotPie(self, data):
#3 11 34
labels = ["class-1", "class-2", "class-3"]
sizes = data
plt.pie(sizes, labels=labels)
plt.show()
def plotPlot(self, x, y):
plt.plot(x, y, "-.")
plt.xlabel("time/second")
plt.ylabel("prediction precision")
plt.show()
def clusterData(self, filePath):
data = numpy.array(
self.wrFile.readDataFromExcel2(filePath=filePath, max_row=48))
#X = data[:,:3]
#cluster_result = numpy.zeros((204,3))
label_result = None
for col in range(0, 204):
X = data[:, col * 3:(col + 1) * 3]
#samples = numpy.sample(range(0,48),43)
#test = set(range(0,48))-set(samples)
# 计算向量间的欧氏距离
#print(" col ",col)
#dist = DistanceMetric.get_metric('euclidean')
#result = dist.pairwise(X)
#print(result)
# 进行聚类
eps = 0.6
clustering = DBSCAN(eps=eps, min_samples=5).fit(X)
labels = numpy.array(clustering.labels_).reshape((48, 1))
#print(clustering.labels_)
# 记录聚类结果
X = numpy.hstack((X, labels))
if col == 0:
label_result = X
else:
label_result = numpy.hstack((label_result, X))
filePath = "F:\\project\\dataset\\vr\\Formated_Data\\Experiment_1/cluster_statistic.xlsx"
def writeAndNormalize(self):
data = self.m / (np.max(self.m))
wrFile = WRFile()
wrFile.writeDataIntoExcel(data=data,
filePath="F:/test/ATBM_" + str(self.q) +
"burst.xlsx")
end = (J + 1) * timeIncrement
if end > num:
end = num
day_begin = max(end - 60 * 60, 0)
m = numpy.mean(data[begin:end])
v = numpy.var(data[begin:end])
Y.append(v / m)
tol = abs(1 - Y[len(Y) - 1] / Y[len(Y) - 2])
J += 1
idc_data[start:end] = Y[len(Y) - 1]
#print("收敛的idc",Y[len(Y)-1],"end",end)
tol = 1
start = end
'''
plt.plot(numpy.arange(0,len(self.data)),self.data/400,"g-",label = "workload/400")
plt.plot(numpy.arange(0,len(idc_data)),idc_data,"m",label = "IDC")
plt.legend(loc = "upper right")
plt.title("Bursty Evaluation of IDC")
plt.xlabel("Time(seconds)")
plt.ylabel("Bursty Intensity")
plt.grid(True)
'''
return idc_data
filePath = "F:/test/workload.xlsx"
burstResult = "F:/test/idc_burst.xlsx"
data = AnalyseIDC(filePath).caculIDC()
wrFile = WRFile()
wrFile.writeDataIntoExcel(data=data, filePath=burstResult)