Making charts and output them as images to the browser in Django

Simple graph made with matplotlib

Simple graph made with matplotlib

Lets say you are working on a website made in Django. And you want to make some nice looking graphs real time, as images from dynamic data. This can be done by using the python 2D graph library matplotlib. The library can be found in the debian package python-matplotlib. A simple graph showing a sine curve, seen to the right, can be generated in regular python using the following code(taken from this example):

from pylab import *
 
t = arange(0.0, 2.0, 0.01)
s = sin(2*pi*t)
plot(t, s, linewidth=1.0)
 
xlabel('time (s)')
ylabel('voltage (mV)')
title('About as simple as it gets, folks')
grid(True)
show()

Output graph to browser from a Django view

If you want to output this graph as a PNG image to the browser from a view in Django, you can store the image in a string buffer and output this buffer using the HttpReponse class in Django and set the mime type to image/png.

from django.http import HttpResponse
from django.shortcuts import render
from matplotlib import pylab
from pylab import *
import PIL, PIL.Image, StringIO
 
def index(request):
    return render(request, 'yourapp/index.html')
 
def showimage(request):
    # Construct the graph
    t = arange(0.0, 2.0, 0.01)
    s = sin(2*pi*t)
    plot(t, s, linewidth=1.0)
 
    xlabel('time (s)')
    ylabel('voltage (mV)')
    title('About as simple as it gets, folks')
    grid(True)
 
    # Store image in a string buffer
    buffer = StringIO.StringIO()
    canvas = pylab.get_current_fig_manager().canvas
    canvas.draw()
    pilImage = PIL.Image.frombytes("RGB", canvas.get_width_height(), canvas.tostring_rgb())
    pilImage.save(buffer, "PNG")
    pylab.close()
 
    # Send buffer in a http response the the browser with the mime type image/png set
    return HttpResponse(buffer.getvalue(), content_type="image/png")

Let’s say you want to display this image on your index page, you can have something like this on your index.html template:

<h1>Hello world of django + matplotlib</h1>
 
 
 
<img src="/showimage/" alt="">

Assuming you have an urls.py with something like this:

from django.urls import include, path
from . import views
 
urlpatterns = [
    path('', views.index, name='main-view'),
    path('/showimage/', views.showimage, name='showimage'),

27 Responses

  1. Anton says:

    If you use django 1.11+, you can choose django-matplotlib field (http://django_matplotlib.readthedocs.org/). This is virtual field (it doesn’t generate a column in the db) which can be easily integrated to Django Admin.

  2. Wonhee says:

    could see the full view.py and the main html page please?

  3. Anonymous says:

    What would be helpful is an FULL example from a CBV (meaning views.py + template file).

  4. Praveen says:

    Thanks for the Notes.
    But how to render the image to html file ?
    I am unable pass it html file. Please help me

    • gayathri says:

      same as praveen… how to write the html file for this

      • Erik Smistad says:

        Lets say the URL pointing to the view show_image is /show_image/, then you add the following in your HTML file:

        <img src="/show_image/" alt="">
        • sajwal says:

          I still do not understand. My index page is on localhost:8000/. in my urls.py it is on path(‘ ‘,views.index, name =’index’). I have a function def index(request). what do I write in index.html to display this image there?

          • Akshay Nimbalkar says:

            Hi Sajwal, did you resolve this issue? Even. i dont understand what to write in index.html to display the image generated from matplotlib or seaborn.

        • Anonymous says:

          It gives me this error when I followed your logic:

          Terminating app due to uncaught exception ‘NSInternalInconsistencyException’, reason: ‘NSWindow drag regions should only be invalidated on the Main Thread!’

          • Sajwal says:

            f you are running a webserver and using it to save Matplotlib make sure to set the backend to a non-interactive one (matplotlib.use(‘agg’) or matplotlib.pyplot.switch_backend(‘Agg’)) so that your server does not try to create (and then destroy) GUI windows that will never be seen (or if they are will be more of a nuisance).
            after import matplotlib
            matplotlib.use(‘agg’)

        • Sajwal says:

          I understand now. Thank you. Just somethings that worked for me are that I had to change
          buffer = StringIO.StringIO() (we can use just StringsIO() in python 3, imported from io)
          to
          buffer = BytesIO()
          which I had to import from io
          from io import BytesIO

          also for some reason I had to add
          matplotlib.use(‘Agg’)

          so that I did not get that NSInternalInconsistencyException’, reason: ‘NSWindow drag regions should only be invalidated on the Main Thread!’ error

          Anyways, thank you so much. Great post!

  5. ojas says:

    Couple of suggestions:

    As of April 2018:
    from string is replaced with frombytes
    user content_type instead of mimetype

  6. You don’t need io or cStringIO. You can just do:
    canvas = FigureCanvasAgg(f)
    response = HttpResponse(content_type=’image/png’)
    canvas.print_png(response)
    plt.close(f)
    return response

  7. Anton says:

    -pylab.cose()
    +pylab.close()

  8. Derek says:

    Can you show an example that involves creating and sending multiple chart images to the same page – I cannot get this to work.

  9. Satvir says:

    Thanku it was really helpful.

  10. noemi says:

    Hi! Thanks for your website, it is really useful. About the output graph to browser from a Django view, I wonder if instead of using HttpResponse we can use render_to_response and represent the graph in a specific html page. I’ve done something like this:

    # serialize to HTTP response
    response = HttpResponse(buffer.getvalue(), mimetype=”image/png”)
    return render_to_response(‘eQL/dev/showImage.html’, {‘response’:response})

    But I don’t know how to plot it in the html file.

    Thanks

  11. masood says:

    whats the book you prefer to learn python with that

  12. Nathan Moos says:

    you could use cStringIO…its faster

    • Erik Smistad says:

      I tried using cStringIO instead of StringIO and measured execution time using datetime. Both buffer types ran in about 0.25 seconds. I think this is because the code only performs one large write to the buffer. If you have a buffer with several writes and reads, cStringIO will probably be notably faster

Leave a Reply

Your email address will not be published.