ElementTree or cElementTree are good. CElementTree is said to be faster. The most important thing is the method you use. If your sample is the extent of the document, any parser will do. However, if your actual documents are large, you should consider an iterparsing method to conserve memory.
I think the most robust parser in terms of methods is lxml, which supports XSL and XPath along with ElementTrees. However, it requires separate installation. Python ships with ElementTree and cElementTree which are fine for most cases. --
