Helper class for dealing with namespaces in elementtree (or any other situation where clark notation is used for qnames).

qnamemap.py

import re

# regular expressions for chars in an NCName
rCHARS = {
    'startchar': u"_a-zA-Z"
        u"\u00C0-\u00D6\u00D8-\u00F6\u00F8-\u02FF\u0370-\u037D\u037F-\u1FFF\u200C-"
        u"\u200D\u2070-\u218F\u2C00-\u2FEF\u3001-\uD7FF\uF900-\uFDCF\uFDF0-"
        u"\uFFFD\u10000-\uEFFFF",
    'char': u"0-9\u00B7\u0300-\u036F\u203F-\u2040.-",
}
rPrefixedName = re.compile(u"(?P<prefix>[{startchar}][{startchar}{char}]*)"
    u":(?P<localname>[{startchar}][{startchar}{char}]*)".format(**rCHARS))


class QNameMap(object):
    """Hold a mapping of prefixes to namespaces
    
    Provides convenience methods to convert back and forth between them,
    especially when using ElementTree.find* methods.

    >>> Q = QNameMap({'ns':'http://example.org'})
    >>> Q('ns:localname')
    '{http://example.org}localname'
    >>> Q.clarkify('.//ns:localname') # best used with et.find(Q.clarify(...))
    './/{http://example.org}localname'
    """
    
    def __init__(self, nsmap):
        self.map = nsmap

    def __call__(self, qn):
        """Return clark notation version of a QName, even with blanks
        
        Unlike QNameMap.qname(), this will also accept blank prefixes or
        localnames.
        
        >>> Q('ns:')
        'http://example.org'
        >>> Q(':name')
        'name'       
        """
        prefix, localname = qn.split(':', 1)
        if not prefix:
            return localname
        elif not localname:
            return self.map[prefix]
        else:
            return self.qname(prefix, localname)

    def qname(self, prefix, localname):
        """Return clark notation version of a QName"""
        return '{{{}}}{}'.format(self.map[prefix], localname)

    def clarkify(self, s):
        """Replace prefixed names with clark-notation fully-qualified names"""
        return rPrefixedName.sub(lambda m: self.qname(**m.groupdict()), s)