Python save matplotlib figure on an PIL Image object

Question

HI, is it possible that I created a image from matplotlib and I save it on an image object I created from PIL? Sounds very hard? Who can help me?

Answer 1

To render Matplotlib images in a webpage in the Django Framework:

• create the matplotlib plot

• save it as a png file

• store this image in a string buffer (using PIL)

• pass this buffer to Django's HttpResponse (set mime type image/png)

• which returns a response object (the rendered plot in this case).
  
  

In other words, all of these steps should be placed in a Django view function, in views.py:

from matplotlib import pyplot as PLT
import numpy as NP
import StringIO
import PIL
from django.http import HttpResponse 


def display_image(request) :
    # next 5 lines just create a matplotlib plot
    t = NP.arange(-1., 1., 100)
    s = NP.sin(NP.pi * x)
    fig = PLT.figure()
    ax1 = fig.add_subplot(111)
    ax1.plot(t, s, 'b.')

    buffer = StringIO.StringIO()
    canvas = PLT.get_current_fig_manager().canvas
    canvas.draw()
    pil_image = PIL.Image.frombytes(
        'RGB', canvas.get_width_height(), canvas.tostring_rgb())
    pil_image.save(buffer, 'PNG')
    PLT.close()
    # Django's HttpResponse reads the buffer and extracts the image
    return HttpResponse(buffer.getvalue(), mimetype='image/png')

Answer 2

I was having the same question and I stumbled upon this answer. Just wanted to add to the above answer that PIL.Image.fromstring has been deprecated and frombytes should be used now instead of fromstring. Hence, we should modify line:

pil_image = PIL.Image.fromstring('RGB', canvas.get_width_height(), 
                 canvas.tostring_rgb())

to

pil_image = PIL.Image.frombytes('RGB', canvas.get_width_height(), 
                 canvas.tostring_rgb())