Recognize text-based nutrients in HTML

[WPDOrdina]

As requested in #1, this issue calls for recognizing text-based nutrients in html. This is something for the future, but here a start is made. Below is something that occurs on the Aldi website in the UK.

<h2 class="h5 font-weight-bold">Specifications</h2>
<table class="table table-striped">
  <tr>
    <th>Nutrition information</th>
    <td>Per 100g: Energy 1543kJ, 371kcal Fat 25g of which saturates 12g Carbohydrate 28g of which sugars 16g Fibre 1.6g Protein 7.5g Salt 0.21g</td>
  </tr>
</table>

[wvengen]

This would be done by the addition of a plain-text parser.
When this has been developed far enough, we could extend the HTML parser to detect the situation, and hand the parsing over to the plain-text parser.
Note that we already have a small plain-txt corpus to test on.

Let's leave this for later, and get the HTML and data-based parsers ready for use. Marking as wontfix for the time being, we can remove this label when we're ready to tackle this.