Pure Python HTML5 validation, using HTML5lib. Designed to integrate easily in django or other web development environments, to ensure templates actually render as intented, without dangling tags, or other nonsense.
The version on PyPI (0.0.2) uses html5lib's basic parsing and linting, here is the development version, which includes actually checking for valid elements and attributes as per the W3C spec.
Still work in progress - it needs to have easy options for specifying extra tags that you want to allow (such as ng- style or v- javascript templating), and also an easy way to disallow anything you don't want, or forbid external URLs, or whatever.
Please feel free to add to this, contributions welcome!
>>> from html5validate import validate
>>> validate('<!doctype html><html><body><h1>Hi World!</h1></body></html>')
None
>>> validate('<!doctype html><html><body><h1>Hi World!</body></html>')
Traceback (most recent call last):
...
html5lib.html5parser.ParseError: Expected closing tag. Unexpected end of file.So. You run 'validate' on some text, and it either returns None, or throws some kind of error at you.
from django.test import TestCase
from html5validate import validate
class TestIndexViewValid(TestCase):
def test_basic_index(self):
validate(self.client.get(reverse('index'), follow=True))Very Early - I just pulled this out of some django view testing code in one of my other projects, as it seemed more sensible to have it as a stand-alone library that could be unit-tested and stuff, and I can use again easily, and then started hacking a bit on it to make it more general.
Initial basic tests are here - but it should really have a LOT of tests written.
- Write a bunch of tests to see how strict it really is.
- Write some extra tree walkers and checkers to look for valid tags, and throw some decent exceptions, and have customisability, specify what tags users want, etc.