intertwingly

Hm, the test fails:

[blalor@beaker ...ps/Phonebook/data]> python lazydom.py
Traceback (most recent call last):
  File "lazydom.py", line 102, in ?
  for name in feed[atom.entry][atom.author][atom.name]:
  File "lazydom.py", line 83, in _iter_
  return result._iter_()
AttributeError: 'list' object has no attribute '_iter_'

What version of python are you using?  Iterator types are new in version 2.2.

On some versions of Redhat, this is installed as python2.

I got the same error and I'm using Python 2.2.2 on Red Hat. I fixed it by changing line 83 to "return iter(result)"

Here's my feedback on the clever lazydom.py. You can use yield to implement __iter__: from __future__ import generators # must be first line of python code ... def __iter__(self): for element in self.list: yield self.new(element) You can use the...

Thanks, all, for the feedback.  Particularly Gary.

Here's a new version.  Tested on Python 2.2 and Python 2.3.

One note: I am definitely trying to maintain the XPath-like abstraction whereby attributes are simply specially named elements.  I did, however, add some support to the namespace class to make namespace qualified attributes easier to access.

Look a look at the SF project pywebsvcs; Brian Overhof is doing a great job of adding schema<->python code generation.

Another option for this kind of thing is Aaron's xmltramp: http://www.aaronsw.com/2002/xmltramp/

This is fun.  Python's '/' operator can be used to implement path syntax.  If you add this line of code to the lazydom class:

  __div__ = __getitem__

then you can write this:

  # iteration and traversal
  for name in feed/atom.entry/atom.author/atom.name:
    print name
  print

  # attribute access and comparison
  print feed/atom.generator/'@name'
  print feed/atom.generator/'@name' == 'Blogger'

Egads, Gary, you're evil!  And here I thought overloading '+' to mean both numeric addition and string concatenation was bad! :-)

Gary: deliciously evil.  Playing with it a bit, my one problem is that the precedence order is wrong.

feed/atom.entry[0] will be evaluated as:

• feed/('http://purl.org/atom/ns#','entry')[0]
• feed/'http://purl.org/atom/ns#'

My (initial) conclusion: perhaps div should be considered instead of index, but having both is dangerous.

More evil.  If you add "_floordiv_ = find", then you can write:

  for issued in feed.find//atom.issued:
  print issued

The more I think about this, and read about it, e.g. here, or here, or think about it in this context, the more firmly I believe that Java, or any strongly typed programming language, is a good vehicle for building... [more]

Trackback from Random Stuff

What's with that trackback? I assumed it's autogenerated, but it's
missing a critical "not". Here's from the link target:

"""
The more I think about this, and read about it, e.g. here, or here, or think about it in this context, the more firmly I believe that Java, or any strongly typed programming language, is not a good vehicle for building document-oriented Web services.
"""

Maybe the 1st version of the linked entry was wrong?

Here's the code.  Key new features vs the original:  element[:] returns all children of all elements in the list element[f] will apply function f to all elements in the list, and act either as a  filter or as a  map or even... [more]

Trackback from Sam Ruby