An ElementTree Builder

Fredrik Lundh | November 2006 | Originally posted to online.effbot.org

Here’s a simple helper class for ElementTree, which lets you use a more convenient syntax to construct XML fragment trees:

import xml.etree.cElementTree as ET
import functools

class _E(object):

    def __call__(self, tag, *children, **attrib):
        elem = ET.Element(tag, attrib)
        for item in children:
            if isinstance(item, dict):
                elem.attrib.update(item)
            elif isinstance(item, basestring):
                if len(elem):
                    elem[-1].tail = (elem[-1].tail or "") + item
                else:
                    elem.text = (elem.text or "") + item
            elif ET.iselement(item):
                elem.append(item)
            else:
                raise TypeError("bad argument: %r" % item)
        return elem

    def __getattr__(self, tag):
        return functools.partial(self, tag)

# create factory object
E = _E()

This started out as a small tweak to the append helper function, but quickly turned into something more like Oren Tirosh’ ~~ElementBuilder~~ (dead link) (which is one of all those tools that I’ve stumbled upon, found interesting, but never gotten around to use).

Anyway, unlike the ordinary Element factory, the E factory allows you to pass in more than just a tag and some optional attributes; you can also pass in text and other elements. The text is added as either text or tail attributes, and elements are inserted at the right spot. Some small examples:

>>> ET.tostring(E("tag"))
'<tag />'
>>> ET.tostring(E("tag", "text"))
'<tag>text</tag>'
>>> ET.tostring(E("tag", "text", key="value"))
'<tag key="value">text</tag>'
>>> ET.tostring(E("tag", E("subtag", "text"), "tail"))
'<tag><subtag>text</subtag>tail</tag>'

For simple tags, the factory also allows you to write “E.tag(…)” instead of “E(‘tag’, …)”:

>>> ET.tostring(E.tag())
'<tag />'
>>> ET.tostring(E.tag("text"))
'<tag>text</tag>'
>>> ET.tostring(E.tag(E.subtag("text"), "tail"))
'<tag><subtag>text</subtag>tail</tag>'

Here’s a somewhat larger example; this shows how to generate HTML documents, using a mix of prepared factory functions for inline elements, nested E.tag calls, and embedded XHTML fragments:

# some common inline elements
A = E.a
I = E.i
B = E.b

def CLASS(*args):
    # class is a reserved word, so we use a helper function
    return {"class": " ".join(args)}

page = (
  E.html(
    E.head(
      E.title("This is a sample document")
    ),
    E.body(
      E.h1("Hello!", CLASS("title")),
      E.p("This is a paragraph with ", B("bold"), " text in it!"),
      E.p("This is another paragraph, with a ",
          A("link", href="http://www.python.org"), "."),
      E.p("Here are some reservered characters: <spam&egg>."),
      ET.XML("<p>And finally, here's an embedded XHTML fragment.</p>"),
    )
  )
)

print ET.tostring(page)

Here’s a prettyprinted version of the output from the above script:

<html>
  <head>
    <title>This is a sample document</title>
  </head>
  <body>
    <h1 class="title">Hello!</h1>
    <p>This is a paragraph with <b>bold</b> text in it!</p>
    <p>This is another paragraph, with a <a href="http://www.python.org">link</a>.</p>
    <p>Here are some reservered characters: &lt;spam&amp;egg&gt;.</p>
    <p>And finally, here's an embedded XHTML fragment.</p>
  </body>
</html>

I’m still experimenting with more mechanisms, including support for callables and nested sequences, but the above makes many ElementTree-based generation tasks a lot simpler.

Addendum

Tweaking, tweaking. This one’s pretty neat, I think:

from datetime import datetime

RSS = ElementBuilder(typemap={
    datetime: lambda e, v: v.strftime("%a, %d %b %Y %H:%M:%S %z")
    })

item = RSS.item(
    RSS.title("E=RSS squared"),
    RSS.link("http://effbot.org"),
    RSS.description("Yet another E-related example"),
    RSS.pubDate(datetime.now(UTC()))
    )

Code will be posted later. Stay tuned.