Streaming XML generation library
Latest commit d607901 Feb 6, 2016 @isagalaev Version 0.5
Failed to load latest commit information.
.gitignore added tests suite Dec 13, 2012
.travis.yml added python 2.5 to travis conf Jan 11, 2013
AUTHORS Merge Jan 11, 2013
LICENSE.txt Modernized and simplified package structure. Apr 3, 2015 Modernized and simplified package structure. Apr 3, 2015
tox.ini Modernized and simplified package structure. Apr 3, 2015


Build Status

Elementflow is a Python library for generating XML as a stream. Some existing XML producing libraries (like ElementTree, lxml) build a whole XML tree in memory and then serialize it. It might be inefficient for moderately large XML payloads (think of a content-oriented Web service producing lots of XML data output). Python's built-in xml.sax.saxutils.XMLGenerator is very low-level and requires closing elements by hand.

Also, most XML libraries, to be honest, suck when dealing with namespaces.


Basic XML generation:

import elementflow
file = open('text.xml', 'w') # can be any  object with .write() method

with elementflow.xml(file, u'root') as xml:
    xml.element(u'item', attrs={u'key': u'value'}, text=u'text')
    with xml.container(u'container', attrs={u'key': u'value'}):
        xml.element(u'subelement', text=u'subelement text')

Using with is required to properly close container elements. The library expects unicode strings on input and produces utf-8 encoded output (you may omit those "u"s for pure ASCII strings if you want to, Python will convert them automatically).

XML with namespaces:

with elementflow.xml(file, 'root', namespaces={'': 'urn:n', 'n1': 'urn:n1'}) as xml:
    with xml.container('container', namespaces={'n2': 'urn:n2'):

Elements with namespaces are defined using prefixes. You can define namespaces at the root level and for any container. The library will check for namespace prefixes that wasn't defined beforehand and will raise ValueError in that case.

Pretty-printing is also supported:

with elementflow.xml(file, 'root', indent=True):
    # ...

In some cases it's more convenient to have such XML producer as a Python iterator. This is easily done by wrapping XML generation code into a generator function:

def g():
    xml = elementflow.xml(elementflow.Queue(), 'root')
    with xml:
        for item in collection:
            yield xml.file.pop()
    yield xml.file.pop()

elementflow.Queue() is a temporary buffer that accepts data from an XML generator and is cleared upon calling .pop() on it. You also might want to yield data from the iterator only when this buffer reaches a certain size:

if len(xml.file) > BUFSIZE:
    yield xml.file.pop()