Plausible Accuracy

A few months back when I was just starting to poke around with Python, I saw this XKCD comic come through my RSS feed (my apologies if this clashes with the right hand sidebar; maximizing your window might help):

XKCD
At the time, I thought it was sort of funny, more for the complete nerdiness of creating a pet from an Eee PC and a hamster ball than anything else. The kicker at the end about importing a soul was just icing.

I bring this up because in preparation for the Elsevier Article 2.0 Challenge coming up in September, I wanted to start spending more time learning how to handle XML files. Since Python has become my language of choice (ok, full honesty - it’s the only language I can speak at all really, and even then only in primitive grunts), I wanted to see how hard it would be to work up an XML parser. It’s really easy. You just have to import it.

import xml.etree.ElementTree as ET

I wrote a very very simple and short script just to make sure that it was as easy as I thought it was, and sure enough this is the case.

xmlparser.py

# Read in our XML file
infile = raw_input("Input XML file name > ")
xmltree = ET.ElementTree(file=infile)
rootelem = xmltree.getroot()

print "This should be root_element"
print rootelem.tag

print "This should print two subelement tags"
for subelement in rootelem:
	print subelement.tag

print "This should print out the content of the sub elements"
for subelement in rootelem:
	print subelement.text

And I used a self-generated test file, test.xml:

<root_element>
	<sub_element>This is a sub element</sub_element>
	<sub_element id="2">This is a sub element with the ID set to "2"</sub_element>
</root_element>

and the output pretty much matches what you would guess:

Input XML file name > test.xml
This should be root_element
root_element
This should print two subelement tags
sub_element
sub_element
This should print out the content of the sub elements
This is a sub element
This is a sub element with the ID set to "2"

This took all of about 10 minutes to do… I’m still sort of stunned.  I’m sure the programmers/Python jockies are laughing right now, but c’est la vie I suppose.

I mean, it’s really almost frighteningly simple.  Let’s try playing with Hamlet, available online in XML format of course.  We can write a quick script to count how often Rosencrantz speaks:

#!/usr/bin/python

# Initialization
import xml.etree.ElementTree as ET

# Read in our XML file
infile = raw_input("Input XML file name > ")
xmltree = ET.ElementTree(file=infile)
rootelem = xmltree.getroot()

i = 0
act_list = rootelem.findall('ACT')
for act in act_list:
	scene_list = act.findall('SCENE')
	for scene in scene_list:
		speech_list = scene.findall('SPEECH')
		for speech in speech_list:
			speaker_list = speech.findall('SPEAKER')
			for speaker in speaker_list:
				if speaker.text == "ROSENCRANTZ":
					i = i + 1

print i

It’s 49, in case you are wondering.

I’m pretty excited in my experimentation with ElementTree so far.  As usual I’ve got a ton to learn, but it’s great to know that this powerful tool was lurking inside of python the whole time.

This entry was posted on Tuesday, June 24th, 2008 at 8:27 pm and is filed under linux/OSS. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.