Add support for UL, OL, PRE, other non-P elements

[migurski]

I’m testing Goose and finding that elements other than paragraphs are unavailable in cleaned_text or top_node.

What I did:

url = 'https://github.com/grangier/python-goose'
g = Goose()
a = g.extract(url=url)

I expected to find the single-item list with “Xavier Grangier” and all the code samples in the output, but they were not there. I would be interested to see an additional property in the output, something like source_node that made the non-cleaned element tree of the original content available.

[migurski]

“Q” is another tag that gets dropped in the HTML cleaning process, which should be kept.

[psilva261]

@migurski Right, all tags get dropped. So the only formatters that stay are '\n'.

Another option would be some attribute cleaned_html that only contains basic html.

[migurski]

It’s not just the tags, it’s also the content. In the example above, the content of the lists is not included in the cleaned text or in the top_node. It’s still in the raw_doc but not unambiguously findable based on xpath or whatever.