imaging.py
#

Image generation via Python Image Library

Let's say we want to plot the Bank of England's historical interest rates as a graph.

 

      		
    

      
        

          #
        

        
Load the Bank of England historical interest rates


      		
      
        
import urllib
url = 'https://spreadsheets.google.com/pub?\
key=0AonYZs4MzlZbcGhOdG0zTG1EWkVQTjBYWm9pWHVRWkE&\
output=csv&\
range=B2:B3798'
data = urllib.urlopen(url).read()
data = [float(x) for x in data.split('\n')]

      		
    

      
        

          #
        

        
Load the Python Image Library

      		
      
        
import Image

      		
    

      
        

          #
        

        
Configure the output path

      		
      
        
import os.path
outpath = '../../talk-docs/python-imaging'

      		
    

      
        

          #
        

        
Create a new RGB image of width 316 px (for 316 years of data)
with color #E0E0E0

      		
      
        
img = Image.new('RGB', (316, 220), (224, 224, 224))

      		
    

      
        

          #
        

        
Save the image as a PNG. The extension determines the file type.

      		
      
        
img.save(os.path.join(outpath, 'image-0.png'))

      		
    

      
        

          #
        

        
Here's the result:


Result

      		
      
        

      		
    

      
        

          #
        

        
Now let' draw axes.

      		
      
        

      		
    

      
        

          #
        

        
Import the ImageDraw library that lets us draw on images

      		
      
        
import ImageDraw
draw = ImageDraw.Draw(img)

      		
    

      
        

          #
        

        
Draw a line every 50 pixels. We'll use y = 0 - 200 to plot 0 - 20% rates.


draw.line takes ((x1,y1), (x2,y2)) as a parameter. By default, it draws
a white line.

      		
      
        
draw.line( ((0, 50), (img.size[0],  50)) )
draw.line( ((0,100), (img.size[0], 100)) )
draw.line( ((0,150), (img.size[0], 150)) )
draw.line( ((0,200), (img.size[0], 200)) )
img.save(os.path.join(outpath, 'image-1.png'))

      		
    

      
        

          #
        

        
Here's the result:


Result

      		
      
        

      		
    

      
        

          #
        

        
Now let's draw the actual graph.
draw.line takes a fill= parameter where you can specify the colour.

      		
      
        
for (i, rate) in enumerate(data[::12]):
    draw.line(((i, 200), (i, 200-rate*10)), fill=(108,108,108))
img.save(os.path.join(outpath, 'image-2.png'))

      		
    

      
        

          #
        

        
Here's the result:


Result

      		
      
        

      		
    

      
        

          #
        

        
Now we'll text labels for the years

      		
      
        
import ImageFont
font = ImageFont.truetype("C:/SYSROOT/Fonts/arial.ttf", 10)
for (i, rate) in enumerate(data[::12]):
    year = 1694 + i
    if not year % 50:
        draw.text((i - 12, 200), str(year), font=font, fill=(0,0,0))
img.save(os.path.join(outpath, 'image-3.png'))

      		
    

      
        

          #
        

        
Result

      		
      
        

      		
    

      
        

          #
        

        
Image processing via Python Image Library


PIL lets you manipulate the images as well. This is best shown via demos.

      		
      
        
img = Image.open('sample.jpg')
img.save(os.path.join(outpath, 'sample.png'))

      		
    

      
        

          #
        

        
Result

      		
      
        

      		
    

      
        

          #
        

        
Resize it

      		
      
        
img = img.resize((200,150))
img.save(os.path.join(outpath, 'small.png'))

      		
    

      
        

          #
        

        
Result

      		
      
        

      		
    

      
        

          #
        

        
Brighten and saturate the image

      		
      
        
import ImageEnhance
ImageEnhance.Brightness(img).enhance(2.0).save(os.path.join(outpath, 'bright.png'))
ImageEnhance.Color(img).enhance(2.0).save(os.path.join(outpath, 'color.png'))

      		
    

      
        

          #
        

        
Result
Result

      		
      
        

      		
    

      
        

          #
        

        
For more samples, see
ngimage

      		
      
        

      		
    

      
        

          #
        

        
Emboss it, or trace contours

      		
      
        
import ImageFilter
img.filter(ImageFilter.EMBOSS).save(os.path.join(outpath, 'emboss.png'))
img.filter(ImageFilter.CONTOUR).save(os.path.join(outpath, 'contour.png'))

      		
    

      
        

          #
        

        
Result
Result

      		
      
        

      		
    

      
        

          #
        

        
Image generation via SVG


Another option is to use SVG to render the file.
This makes it look better on print.


Some sample reports created this way are at
Report Bee and
KLP.

      		
      
        

      		
    

      
        

          #
        

        
We'll use tornado templates to render the file

      		
      
        
from tornado import template

      		
    

      
        

          #
        

        
Create a basic template that'll draw a grey box

      		
      
        
svg = '''<svg width="316" height="220" xmlns="http://www.w3.org/2000/svg">
<rect x="0" y="0" width="316" height="220" style="fill:#e0e0e0"/>'''

      		
    

      
        

          #
        

        
Add the axes

      		
      
        
svg += '''
<line x1="0" x2="316" y1="50" y2="50" stroke-width="1" stroke="#fff"/>
<line x1="0" x2="316" y1="100" y2="100" stroke-width="1" stroke="#fff"/>
<line x1="0" x2="316" y1="150" y2="150" stroke-width="1" stroke="#fff"/>
<line x1="0" x2="316" y1="200" y2="200" stroke-width="1" stroke="#fff"/>'''

      		
    

      
        

          #
        

        
Add the graph

      		
      
        
svg += '''
{% for (i, rate) in enumerate(data[::12]) %}
 <rect x="{{ i }}" width="1" y="{{ 200 - rate*10 }}" height="{{ rate*10 }}"
  fill="#6c6c6c" stroke-width="0.1" stroke="#fff" />
{% end %}'''

      		
    

      
        

          #
        

        
Add the labels

      		
      
        
svg += '''
<g style="font-family:Arial;font-size:10px">
{% for (i, rate) in enumerate(data[::12]) %}
  {% set year = 1694 + i %}
  {% if not year % 50 %}
    <text x="{{ i-12 }}" y="200" style="dominant-baseline:hanging">
      {{ year }}
    </text>
  {% end %}
{% end %}
</g></svg>'''

      		
    

      
        

          #
        

        
Generate the template and write the data

      		
      
        
open(os.path.join(outpath, 'graph.svg'), 'w').write(template.Template(svg).generate(data=data))

      		
    

      
        

          #
        

        
Here's the result: