Unnecessary <div> elements

[jemrobinson]

In some cases, there are several wrapping levels of <div> in the ReadabiliPy output. These should be flattened.

The rationale behind this, is that the output of ReadabiliPy should (ideally) not depend on how deeply wrapped in <div>s the input is.

If we go down this route, we should think about which types of tag can be safely stripped out (eg. <div> and <span> are fine, but <table> or <figure> probably aren't).

[jemrobinson]

Example:

<div class="wrapper">
    <div class="article">
        <div id="advert"></div>
        <div class="formatted">
            <p>First paragraph contents.</p>
        </div>
    </div>
</div>

Could/should be reduced to:

<div>
    <p>First paragraph contents.</p>
</div>

What do you think @martintoreilly ? Note that if you add another paragraph you get the following structure:

<div class="wrapper">
    <div class="article">
        <div id="advert"></div>
        <div class="formatted">
            <p>First paragraph contents.</p>
        </div>
        <div class="formatted">
            <div class="aside">
                <p>Second paragraph contents.</p>
            </div>
        </div>
    </div>
</div>

would be reduced to

<div>
    <p>First paragraph contents.</p>
    <p>Second paragraph contents.</p>
</div>