# 针对不同的docker（6个）画出cpu和mem的统计图
# coding:utf-8
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import os
file_path = '../data2/docker_output_2019.5.6.new.csv'
docker_dict = {}
img_save_path = '../img/docker_output_data2/'
def read_file(file_path, docker_dict):
#读取csv文件中的数据
file_path = '../data2/docker_output_2019.5.6.new.csv'
data = pd.read_csv(file_path)
# 按照docker进行分组，并且赋值到相应的字典
groups = data.groupby(data['docker'])
docker_dict = {}
for i,group in enumerate(groups):
docker_dict['docker_'+str(i)] = group
return data, docker_dict
def draw_mean_data(data):
# 画出平均所有docker的数据图
docker_mean = data.groupby(data['docker']).mean()
docker_mean.index = ['docker_'+ str(i) for i in np.arange(6)]
docker_mean_cpu = docker_mean['cpu']
docker_mean_cpu.plot.bar()
plt.title('mean docker CPU')
plt.xticks(rotation=360)
plt.savefig(img_save_path + 'all_CPU_mean.jpg')
plt.show()
docker_mean_mem = docker_mean['mem']
docker_mean_mem.plot.bar()
plt.title('mean docker MEM')
plt.xticks(rotation=360)
plt.savefig(img_save_path + 'all_MEM_mean.jpg')
plt.show()
def draw_docker(docker_name, docker_dict):
# 输入序号，如:docker_1
# tuple注意，每项有两个，第一项docker的id，第二项是对应的dataframe，所以取docker_dict[docker_name][1]
data0 = docker_dict[docker_name][1]
# cpu
data_cpu = data0['cpu']
# 重新设置series的index
data_cpu.index = np.arange(len(data_cpu))
# mem
data_mem = data0['mem']
data_mem.index = np.arange(len(data_mem))
# 分开画CPU和MEM的，分别存储
data_cpu.plot()
plt.title(docker_name+":CPU")
plt.savefig(img_save_path + docker_name + '_CPU.jpg')
plt.show()
data_mem.plot()
plt.title(docker_name+":MEM")
# plt.show()
plt.savefig(img_save_path + docker_name + '_MEM.jpg')
plt.show()
# 画CPU和MEM一起的，一起存储
data_cpu.plot(label='CPU')
plt.title(docker_name+":CPU+MEM")
data_mem.plot(label='MEM')
plt.legend()
plt.savefig(img_save_path + docker_name + '_CPU_MEM.jpg')
plt.show()
if __name__ == '__main__':
data, docker_dict = read_file(file_path, docker_dict)
draw_mean_data(data)
groups = data.groupby(data['docker'])
for i in range(len(groups)):
docker_name = 'docker_'+ str(i)
draw_docker(docker_name, docker_dict)
# 打包需要的文件
os.system('zip -q -r ' + img_save_path + 'docker_output_img.zip '+ img_save_path)