import json
import math
import numpy as np
from scipy import stats
import matplotlib.pyplot as plt

data_list = []
with open('test_130380_73.log','r') as f:
    data = f.readlines()
    #readlines按行读取返回数组，遍历数组元素
    for i in data:
        item = i.split('<<<<<')
        each = item[1].split('>')
        data_list.append(json.loads(each[0]) )

cache = []
mysql = []
emqt = []
push = []
all = []
for dic in data_list:
    if dic['type'] == 'mysql':
        mysql.append(dic['use'])
    elif dic['type'] == 'cache':
        cache.append(dic['use'])
    elif dic['type'] == 'emqt':
        emqt.append(dic['use'])
    elif dic['type'] == 'push':
        push.append(dic['use'])
    elif dic['type'] == 'all':
        all.append(dic['use'])

# print('cache用时：',cache)
# print('mysql时间：',mysql)
# print('emqt时间：',emqt)
# print('push时间：',push)
# print('all时间：',all)

#stats.t.interval(置信度，自由度，均值，标准差)   得到置信区间
mysql_ci = stats.t.interval(0.95, df=len(mysql)-1, loc=np.mean(mysql), scale=stats.sem(mysql))
cache_ci = stats.t.interval(0.95, df=len(cache)-1, loc=np.mean(cache), scale=stats.sem(cache))
emqt_ci = stats.t.interval(0.95, df=len(emqt)-1, loc=np.mean(emqt), scale=stats.sem(emqt))
push_ci = stats.t.interval(0.95, df=len(push)-1, loc=np.mean(push), scale=stats.sem(push))
all_ci = stats.t.interval(0.95, df=len(all)-1, loc=np.mean(all), scale=stats.sem(all))

print('mysql_ci',mysql_ci)
print('cache_ci',cache_ci)
print('emqt_ci',emqt_ci)
print('push_ci',push_ci)
print('all_ci',all_ci)
print('cache样本数：',len(cache),'最大用时：',max(cache),'最小用时：',min(cache),'cache均值：','%.2f' % np.mean(cache))
print('emqt样本数',len(emqt),'最大用时：',max(emqt),'最小用时：',min(emqt),'emqt均值：','%.2f' % np.mean(emqt))
print('push样本数',len(push),'最大用时：',max(push),'最小用时：',min(push),'push均值：','%.2f' % np.mean(push))
print('all样本数',len(all),'最大用时：',max(all),'最小用时：',min(all),'all均值：','%.2f' % np.mean(all))
print('mysql样本数',len(mysql),'最大用时：',max(mysql),'最小用时：',min(mysql),'mysql均值：','%.2f' % np.mean(mysql))