96 lines
2.3 KiB
Python
96 lines
2.3 KiB
Python
"""
|
|
* Copyright (C) 2022 Intel Corporation.
|
|
*
|
|
* SPDX-License-Identifier: BSD-3-Clause
|
|
"""
|
|
|
|
from flask import Flask
|
|
from flask import send_file
|
|
|
|
import numpy as np
|
|
import pandas
|
|
import random
|
|
|
|
import matplotlib.pyplot as plt
|
|
|
|
import posix_ipc as ipc
|
|
import mmap
|
|
|
|
app = Flask(__name__)
|
|
|
|
#Runs when the user goes to our ip address
|
|
@app.route('/')
|
|
def histapp():
|
|
|
|
#Create and save the figure
|
|
create_hist()
|
|
|
|
#Send the histogram as a webpage to the user
|
|
return send_file("/root/hist.png", mimetype='image/png')
|
|
|
|
#Creates the user histogram and saves to hist.png
|
|
def create_hist():
|
|
|
|
#Go to the beginning of the shared memory region
|
|
shm.seek(0)
|
|
|
|
#Get the data
|
|
web_sem.acquire()
|
|
data = shm.readline();
|
|
web_sem.release()
|
|
|
|
#Transform the data into an array that matplotlib can understand
|
|
count, dataset = transform_data(data)
|
|
|
|
#Clear the figure and recreate from the new data
|
|
plt.clf()
|
|
|
|
#Setup the figure and save it
|
|
bins=np.arange(min(dataset),max(dataset)+2)-0.5
|
|
figure = plt.hist(dataset,bins,rwidth=0.5)
|
|
|
|
plt.title("ACRN Sample Application cyclictest display (unoptimized)")
|
|
plt.xlabel("Latency Value (microseconds)")
|
|
plt.ylabel("Count Percentage " + f"{count:,}")
|
|
plt.savefig("/root/hist.png")
|
|
|
|
return figure
|
|
|
|
def transform_data(data_string):
|
|
|
|
#Holds the transformed data
|
|
transformed_data_values = []
|
|
|
|
str_data = data_string.decode("utf-8")
|
|
str_data = str_data.replace('\n',"")
|
|
|
|
data_values = str_data.split()
|
|
|
|
#Holds the count of latencies that we have
|
|
total_count = data_values[0]
|
|
|
|
#Used for transforming the data values
|
|
data_percentages = data_values[1:]
|
|
if (len(data_percentages) % 2 != 0):
|
|
return transformed_data_values
|
|
|
|
#Transform the data into a list that can be fed to matplotlib
|
|
for i in range(0, int(len(data_percentages) / 2)):
|
|
transformed_data_values += ([int(data_percentages[i*2])] * int(data_percentages[i*2 + 1]))
|
|
|
|
return int(total_count), transformed_data_values
|
|
|
|
if __name__ == '__main__':
|
|
|
|
#Set up shared memory between userapp and the webserver
|
|
shm_path = "/pyservershm"
|
|
shm_size = 1048576
|
|
shm_f = ipc.SharedMemory(shm_path, 0, size=shm_size)
|
|
shm = mmap.mmap(shm_f.fd, shm_size)
|
|
|
|
#Set up the semaphore to maintaine synchronization
|
|
web_sem = ipc.Semaphore("/pyserversem", 0, 0o0774)
|
|
|
|
#Run the webserver
|
|
app.run(host="0.0.0.0", port=80, debug=False)
|