`utils.get_pages(item)` function seems incorrect

[tadeoos]

Hey,

Thanks a lot for creating and maintaining this library, it's very useful!

I was looking into the code that generates <nav epub:type="page-list" id="pages" hidden="hidden"> element inside the nav file, since I got a rather random entry there.

The current code picked up this element:

<div id="cop" class="copyright" epub:type="copyright-page">...</div>

And added this:

# nav.xhtml
...
    <nav epub:type="page-list" id="pages" hidden="hidden">
      <h2>Pages</h2>
      <ol>
        <li>
          <a href="xhtml/blabla.xhtml#cop">cop</a>
        </li>
      </ol>
    </nav>

This got me to:

def get_pages(item):
    body = parse_html_string(item.get_body_content())
    pages = []

    for elem in body.iter():
        if 'epub:type' in elem.attrib:
            if elem.get('id') is not None:
                _text = None
                
                if elem.text is not None and elem.text.strip() != '':
                    _text = elem.text.strip()

                if _text is None:
                    _text = elem.get('aria-label')

                if _text is None:
                    _text = get_headers(elem)

                pages.append((item.get_name(), elem.get('id'), _text or elem.get('id')))

    return pages

I can't understand the intention here... The if 'epub:type' in elem.attrib check seems to assume that anything that specifies any epub:type is of type pagebreak?

Seems to me a better heuristic should be to pick a first pagebreak in the chapter file and use that value (or that value - 1?) for the page value. Alternatively maybe we should add an attribute page_list which will work similarly to links or toc where one could specify this more manually...?