问题
I need to tidy up HTML pages and convert them to XML in Python; losing some "bad" parts if needed.
I used TagSoup for some time, but it doesn't understand new "article", "footer" tags, and doesn't like "meta" when they are not in the head; making resulting XML almost impossible to process.
I like what html5lib does so far, but my fifth test (very weird tests) failed; when parsing
<div attr="val"">
using html5lib + xml.dom treebuilder, I got the following in the resulting XML string:
<div attr="val" "="">
which is not a good result for well-formed xml.
When I tried html5lib + lxml as a treebuilder, I got that converted to
<div attr="val" U00022="">
which is better, but the problem is that lxml "eats" closing tags/slashes for <link> tags, making them just <link ... > when outputting XML.
What would you recommend to use?
回答1:
You can use method to set an Element to be self-closing or not, something like this:
from lxml import etree
tree = etree.Element('div', attrib={'attr':'val', 'U00022':''})
etree.tostring(tree)
'<div U00022="" attr="val"/>'
# parse as self-closing tag
etree.tostring(tree, method='xml')
'<div U00022="" attr="val"/>'
# parse as normal HTML
etree.tostring(tree, method='html')
'<div U00022="" attr="val"></div>'
Then just do whatever you want from it.
When you're trying to write from the Element, you can also add method too:
tree = html5lib.parse('<div attr="val" U00022="">', treebuilder='lxml', namespaceHTMLElements=False)
tree.write('yourfilename', method='html')
Printout:
<html><head></head><body><div u00022="" attr="val"></div></body></html>
来源：https://stackoverflow.com/questions/26717379/parse-any-html-to-xml-using-html5lib