Then we don't need to change mobileview API (T106143). We would also get better structured HTML so we can be more deliberate in our DOM transformations.
The first problem here is to split the big HTML content into sections.
Another problem is to see if we can get all the metadata we get from mobileview. It would be good to not have to also call mobileview, too.
- (Wikidata) description
- page id
- information about the lead image
- page sections
- (isDisambiguation -- would be nice but doesn't even seem to work with MW API)