I really like using ElementTree and it’s lxml and xml.etree brethern.  But one of my pet peeves is how it deals with namespaces. I understand the reasoning — it’s just difficult and bulky.  Combining the syntax of {namespace-uri} with an XPath-like search string can be confusing.

I’m muddling over this while I’m writing yet another utility script to help me with some XML.  I stop and decided I’m going to build a help class, wittingly called PathBuilder.

class PathBuilder:

    def __init__(self,namespace_uri):

        self._template=string.Template("{%s}$tag" %namespace_uri)

    def all_tags(self,tag):
        return "//"+self._template.substitute(tag=tag)

    def all_decendents(self,tag):
        return ".//"+self._template.substitute(tag=tag)

    def children(self,tag):

        return self._template.substitute(tag=tag)

My methods above are all that I need at this point, but you can see how to build them all. I can put all my path building knowledge in that class and then just tell it what I want and it will give it to me.