Simple Scrapy middleware to process non-well-formed HTML with BeautifulSoup
The package is on PyPI and can be installed with pip:
pip install scrapy-beautifulsoup
Add the middleware to DOWNLOADER_MIDDLEWARES dictionary setting:
DOWNLOADER_MIDDLEWARES = {
'scrapy_beautifulsoup.middleware.BeautifulSoupMiddleware': 400
}
By default, BeautifulSoup would use the built-in html.parser parser. To change it, set the BEAUTIFULSOUP_PARSER setting:
BEAUTIFULSOUP_PARSER = "html5lib" # or BEAUTIFULSOUP_PARSER = "lxml"
html5lib is an extremely lenient parser and, if the target HTML is seriously broken, you might consider being it your first choice.
Note: html5lib has to be installed in this case:
pip install html5lib
BeautifulSoup itself with the help of an underlying parser of choice does a pretty good job of handling non-well-formed or broken HTML.
In some cases, it makes sense to pipe the HTML through BeautifulSoup to "fix" it.