Zimcheck doesn't account for <base />

[rgaudin]

In sotoki, we use the <base /> element to deal properly and cleanly with internal links thorough the ZIM.

Unfortunately, zimcheck doesn't account for this base when checking (internal in this case) links.

So I get tons of errors like

  The following links:
- static/js/webp-hero.bundle.js
(C/users/8672/static/js/webp-hero.bundle.js) were not found in article C/users/8672/gamora

This is because we have a proper base and the link static/js/webp-hero.bunde.js is from the root of the Zim (our base) and not from the context-path of that article.

And this goes on and on on most pages for all the resources listed in the HTML source.

Similarly, we get errors like:

  Found 2 empty links in article: C/users/-1/community

Empty links are valid and in this context points to the base, and thus the root of our Zim.

[rgaudin]

FYI, we changed our scraper to not use a <base /> because it was easier than fixing zimcheck and it also caused issues with kiwix-js