How can I check the validity of HTML code with Python?
I need closed tags check, and braces in tags params. Such as |a href="xxx'| and other possible validations, which libs I can use for this?
Add a comment
|
2 Answers
Well, this isn't exactly what you're looking for, but to validate the HTML for a web site I work on, I ask the W3C Validator to check it for me, and I just screen scrape the output to get the basic yes/no result. Note there are several validation services on the web as alternatives, but W3C works well enough for me.
#!/usr/bin/python2.6
import re
import urllib
import urllib2
def validate(URL):
validatorURL = "http://validator.w3.org/check?uri=" + \
urllib.quote_plus(URL)
opener = urllib2.urlopen(validatorURL)
output = opener.read()
opener.close()
if re.search("This document was successfully checked as".replace(
" ", r"\s+"), output):
print " VALID: ", URL
else:
print "INVALID: ", URL
1 Comment
Changaco
stackoverflow.com/a/10519634/2729778 is a similar but cleaner answer.
The html5lib module can be used to perform basic HTML validation:
>>> import html5lib
>>> html5parser = html5lib.HTMLParser(strict=True)
>>> html5parser.parse('<html></html>')
Traceback (most recent call last):
...
html5lib.html5parser.ParseError: Unexpected start tag (html). Expected DOCTYPE.
>>> html5parser.parseFragment('<p>Lorem <a href="/foobar">ipsum</a>')
<Element 'DOCUMENT_FRAGMENT' at 0x7f1d4a58fd60>
>>> html5parser.parseFragment('<p>Lorem </a>ipsum<a href="/foobar">')
Traceback (most recent call last):
...
html5lib.html5parser.ParseError: Unexpected end tag (a). Ignored.
>>> html5parser.parseFragment('<p><form></form></p>')
Traceback (most recent call last):
...
html5lib.html5parser.ParseError: Unexpected end tag (p). Ignored.
>>> html5parser.parseFragment('<option value="example" />')
Traceback (most recent call last):
...
html5lib.html5parser.ParseError: Trailing solidus not allowed on element option
1 Comment
Garrett
This seems not to catch the following invalid HTML:
html5parser.parseFragment('<span><div></div></span>').