Flask stuck loading after opening a certain route with a function
Multi tool use
Flask stuck loading after opening a certain route with a function
I'm relatively new to Python and Flask. I've been trying to create a web application that reads data readings from a .txt file and plots them onto a matplotlib plot. In this web-app, I have a front page with 3 buttons. These buttons redirect to different routes with functions that read data and plot them onto a matplotlib plot. The web-app works perfectly only for the first time I go to either of these routes. After that nothing loads anymore. I guess I have an infinite loop of some sort but I can't figure it out. Also, after the website gets stuck the Python process starts consuming more resources.
The problem persists only when I open this route on the web-app:
@app.route("/temperature/")
This loads without problems on the web-page, but only for one time and the whole web-app gets stuck and I cannot access any of the other routes either.
Thanks in advance!
EDIT 1 - Whole code below
cloudapp.py (Python sourcecode that runs Flask and the functions)
from flask import Flask
from flask import render_template
from flask import request
import numpy as np
import matplotlib.pyplot as plt, mpld3
from datetime import datetime
app = Flask(__name__, template_folder='C:UsersValtteriDesktopcloudappTemplatesHTML')
global all_lines
@app.route("/")
def frontpage():
return render_template('frontpage.html')
@app.route("/temperature/")
def temperature():
f = open('C:/Email/file.txt', 'r')
cpt = 0 # Line amount value
all_lines = # List that has every Nth value
for line in f:
cpt += 1 # Goes through every line and adds 1 to the counter
# How often values are plotted (every Nth value)
if cpt%100 == 0:
all_lines.append(line) # When 30th line is counted, add that line to all_lines list
if cpt == 500: # How many values are plotted (counts from first)
break
dates = [str(line.split(';')[0]) for line in all_lines]
date = [datetime.strptime(x,'%Y.%m.%d_%H:%M') for x in dates]
y = [float(line.split(';')[1]) for line in all_lines]
z = [float(line.split()[2]) for line in all_lines]
fig, ax = plt.subplots()
ax.plot_date(date, y, 'r-')
f.close()
return mpld3.fig_to_html(fig)
if __name__ == "__main__":
app.run()
frontpage.html (HTML template in folder ..templateshtml)
<!DOCTYPE html>
<html>
<head>
<meta charset="UTF-8">
<title>Envic Oy Cloud</title>
</head>
<body>
<div class="headers">
<h1>Web-Application</h1>
<h2> Version 0.0.1 </h2>
</div>
<div class="buttons">
<h3 class="buttonheader">Logger 1</h3>
<a class="templink" href="http://127.0.0.1:5000/temperature/" target="_blank"> Check temperature </a>
</body>
</html>
I use Bash on windows to run the code with
export FLASK_APP=myCloud.py
flask run
I tried to solve the issue for a long time, but couldn't find a solution. It has something to do with Flask/mpld3 compatibility for me. I made the very same web app, but this time using a simple pyramid WSGI. I can now refresh the plot as many times and redirect myself into any view without the server hanging. I'll still leave the post unsolved, because I would still like to use Flask. I will continue my research too. Here is the pyramid version that works for me:
from wsgiref.simple_server import make_server
from pyramid.config import Configurator
from pyramid.response import Response
import numpy as np
import matplotlib.pyplot as plt, mpld3
from datetime import datetime
def hello_world(request):
return Response('<h1>Testing the Pyramid version!</h1><a href="http://localhost:8888/second_view">Check temperature</a>')
def second_view(request):
with open('C:/Email/file.txt') as f:
cpt = 0 # Line amount value
all_lines = # List that has every Nth value
for line in f:
cpt += 1 # Goes through every line and adds 1 to the counter
# How often values are plotted (every Nth value)
if cpt%100 == 0:
all_lines.append(line) # When 30th line is counted, add that line to all_lines list
if cpt == 500: # How many values are plotted (counts from first)
break
dates = [str(line.split(';')[0]) for line in all_lines]
date = [datetime.strptime(x,'%Y.%m.%d_%H:%M') for x in dates]
y = [float(line.split(';')[1]) for line in all_lines]
z = [float(line.split()[2]) for line in all_lines]
plt.figure(figsize=(10,5))
plt.title('Humidity', fontsize=15)
plt.ylabel('Humidity RH', fontsize=15)
fig = plt.figure()
plot = plt.plot_date(date, z, 'b-')
myfig = mpld3.fig_to_html(fig, template_type='simple')
return Response(myfig)
if __name__ == '__main__':
config = Configurator()
config.add_route('hello_world', '/hello_world')
config.add_route('second_view', '/second_view')
config.add_view(hello_world, route_name='hello_world')
config.add_view(second_view, route_name='second_view')
app = config.make_wsgi_app()
server = make_server('', 8888, app)
server.serve_forever()
return
Thanks for the response. Tried that, but unfortunately didn't solve the issue. I also tried to do f.close() before the return statement, but it still remains the same. I am taking the same guess as you, that it is the file handler that doesn't close and the code cannot process anything else.
– DeersAreFriends
Jul 2 at 7:21
The issue seems to be with the plot itself (matplotlib or mpld3), see the edit in my original post.
– DeersAreFriends
Jul 2 at 7:33
Hmmm... I can't seem to replicate your problem. However, while playing with your code, I noticed you have 2
plt.figures. I believe you want to remove the 2nd one and set fig to the 1st statement. I also tried using plt.clf() and plt.cla() right before the return statement to see if pyplot can flush its cache.– Scratch'N'Purr
Jul 2 at 8:44
plt.figure
fig
plt.clf()
plt.cla()
return
Interesting. I now have only one route with only a simple code that creates a plot with matplotlib and prints it on the web page with mpld3, and I still have the issue. It loads once, but after that it becomes unresponsive and the page can't be even refreshed. The loading icon appears on chrome but it just keeps spinning forever.
– DeersAreFriends
Jul 2 at 9:23
1 Answer
1
Here is what I attempted. I didn't get any issues with the server hanging after refreshing or opening a new tab of the link. However, after I shutdown the server (Ctrl + C), the console threw some exceptions that suggests that mpld3 or matplotlib had opened some new threads. Specifically, the exception is RuntimeError: main thread is not in main loop.
Ctrl + C
RuntimeError: main thread is not in main loop
I did some googling and came across this link. The guy suggested using fig_to_dict with json. I tried his solution and I was still getting the exception.
fig_to_dict
json
Now, I'm going to write down both approaches and let you decide which to use. I don't know if either of them will work for you. For my setup, the app runs fine despite the exceptions after closing the server.
I'm also going to use your example that didn't read the txt file. I have debug set to True so I can make sure the GET requests are being processed when I refresh the chart or open a new instance of the link.
True
Approach 1 (fig_to_html)
app.py
from flask import Flask
from flask import render_template
import matplotlib.pyplot as plt
import mpld3
app = Flask(__name__, template_folder='/path/to/templates')
@app.route("/")
@app.route("/index")
def index():
return render_template('frontpage.html')
@app.route('/temperature')
def temperature():
date = ([1, 2, 3, 4])
y = ([1, 2, 3, 4])
fig = plt.figure(figsize=(10, 5))
plt.title('Temperature', fontsize=15)
plt.ylabel('Temperature' + u'u2103', fontsize=15)
plt.plot(date, y, 'b-')
plt.ylim([0, 40])
myfig = mpld3.fig_to_html(fig, template_type='simple')
plt.clf() # clear figure
plt.cla() # clear axes
plt.close('all') # close all figures
# Print as HTML
return myfig
if __name__ == "__main__":
app.run(debug=True) # run on debug mode
Approach 2 (fig_to_dict)
app.py
from flask import Flask
from flask import render_template
import matplotlib.pyplot as plt
import mpld3
import json
app = Flask(__name__, template_folder='/path/to/templates')
@app.route("/")
@app.route("/index")
def index():
return render_template('frontpage.html')
@app.route('/temperature')
def temperature():
date = ([1, 2, 3, 4])
y = ([1, 2, 3, 4])
fig = plt.figure(figsize=(10, 5))
plt.title('Temperature', fontsize=15)
plt.ylabel('Temperature' + u'u2103', fontsize=15)
plt.plot(date, y, 'b-')
plt.ylim([0, 40])
single_chart = dict()
single_chart['id'] = "temp_figure"
single_chart['json'] = json.dumps(mpld3.fig_to_dict(fig))
plt.clf() # clear figure
plt.cla() # clear axes
plt.close('all') # close figure
# Print as HTML
return render_template('temperature.html', single_chart=single_chart)
if __name__ == "__main__":
app.run(debug=True) # run on debug mode
And here are the template files. I noticed you used a static link in your frontpage.html so I replaced that with a placeholder that allows Flask to auto-populate the URL. You also had a div tag that you didn't close.
frontpage.html
div
frontpage.html
<!DOCTYPE html>
<html>
<head>
<meta charset="UTF-8">
<title>Envic Oy Cloud</title>
</head>
<body>
<div class="headers">
<h1>Web-Application</h1>
<h2> Version 0.0.1 </h2>
</div>
<div class="buttons"></div>
<h3 class="buttonheader">Logger 1</h3>
<a class="templink" href="{{ url_for('temperature') }}" target="_blank"> Check temperature </a>
</body>
</html>
temperature.html (for 2nd approach only)
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8" />
<title>Sample Page</title>
<script src="https://cdnjs.cloudflare.com/ajax/libs/require.js/2.1.10/require.min.js"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/jquery/2.0.3/jquery.min.js"></script>
<script type="text/javascript" src="http://d3js.org/d3.v3.min.js"></script>
<script type="text/javascript" src="http://mpld3.github.io/js/mpld3.v0.2.js"></script>
</head>
<body>
<div id="{{single_chart.id}}">
</div>
<script type="text/javascript">
var figureId = "{{single_chart.id}}";
var json01 = {{single_chart.json|safe}};
mpld3.draw_figure(figureId, json01);
</script>
</body>
</html>
BTW, I'm using Python 3.5.4, Flask==0.12.2, matplotlib==2.1.2, and mpld3==0.3. I tested using Chrome Version 67.0.3396.87.
Thanks for the help. I tried out your version, but unfortunately I still have the issue. I click the link, it opens and loads the plot. After I close the plot and try to open it again, it does not process 'GET' requests anymore. I am using python 2.7, could it be that?
– DeersAreFriends
Jul 3 at 5:55
The fig to dict version did not work at all. It loaded up a blank page. After that the GET requests didn't work anymore again. This issue is so frustrating. I might try the code on another PC with python 3.5. Oh BTW, my program couldn't navigate to URLs with the route address @app.route('/temperature'), for me I need to put another forward slash on it to work ('/temperature/'). Otherwise it gives me an error that URL not found.
– DeersAreFriends
Jul 3 at 7:09
I tried with 2.7 and didn't have issues either. I did some more googling and found this repo that might be worth looking at. The guy creates a lock before creating his plot. If this doesn't work, then I can't really help you any further since I'm out of ideas.
– Scratch'N'Purr
Jul 3 at 7:11
Okay. Seems like a really weird issue.. I can't thank you enough for your help, was really kind of you. Would have taken a lot of my time troubleshooting this and thinking that there is something wrong with my code. I'll try that one out, and if it does't work, i'll switch my PC and go with a fresh start. Hopefully I get it working :)
– DeersAreFriends
Jul 3 at 7:37
Not a problem! Hope you get it working! I'd be interested to see the solution if you figure it out. :)
– Scratch'N'Purr
Jul 3 at 8:16
By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.
Try unindenting your
returnstatement. My hunch is that the file handler never closes the file so you probably have a file lock on the file in subsequent reads.– Scratch'N'Purr
Jul 2 at 6:59