Django snippets: "Zoom in" on rendered HTML that the test client returns

# python
import re
from cStringIO import StringIO
# django
from django.test import TestCase


class BaseTestCase(TestCase):

    def _zoom_html(self, html, css_selector):

        try:
            from lxml.html import parse
            from lxml import etree
            from lxml.cssselect import CSSSelector
        except ImportError:
            return html

        parser = etree.HTMLParser()
        tree = etree.parse(StringIO(html), parser)
        page = tree.getroot()

        if isinstance(css_selector, basestring):
            selector = CSSSelector(css_selector)
        else:
            raise ValueError("css_select must be a string")

        html_chunks = []
        for part in selector(page):
            html_chunks.append(etree.tostring(part, pretty_print=True))

        return ''.join(html_chunks)

    def _strip_html(self, html):
        return re.sub('</?\w.*?>', '', html)

More like this

custom template tag sample by shinyzhu 3 years, 3 months ago
Generic CSV Export by zbyte64 4 years, 11 months ago
Yet another SQL debugging facility by miracle2k 5 years, 9 months ago
Cookie based flash errors and notices (a la Rails) by alexk 4 years, 8 months ago
filter for extracting a number of paragraphs from any HTML code by rafadev 1 year, 11 months ago

If you have this as your base class for all unit tests you can do the following:

class TestViews(BaseTestCase):
    def test_generated_stats(self):
        "test that certain stuff in the response"

        ...create some content for testing or use fixtures...

        response = self.client.get('/some/page/')

        # At this point response.content is a huge string filled with HTML tags and
        # "junk" that you don't need for testing the content thus making it difficult
        # to debug the generated HTML because it so huge. 
        # So we can zoom in on the <div id="stats>...</div> node
        html = self._zoom_html(response.content, '#stats')

        # the variable 'html' will now be something like this:
        """
        <div id="stats">
            <p>
                <strong>2</strong> students<br/>
                <em>1</em> warning.
            </p>
        </div>
        """

        # This makes it easier to debug the response and easier to test
        # against but the HTML might still be in the way so this would fail:
        self.assertTrue('2 students' in html) # will fail

        # To strip away all html use _strip_html()
        content = self._strip_html(html)

        # Now this will work
        self.assertTrue('2 students' in content) # will work

It works for me and I find this very useful so I thought I'd share it.

Author:: peterbe
Posted:: April 21, 2009
Language:: Python
Django Version:: 1.0
Tags:: css test client lxml lxml.html
Score:: 1 (after 1 ratings)

Django snippets

"Zoom in" on rendered HTML that the test client returns

More like this

Tools

Comments

Snippets

About