Previously, the article intro text was taken
from the CPS `/page/summary` endpoint. That endpoint returns
the first paragraph of an article.

However, after looking more closely into the
summarization research. I realized we need the
first *section* of the article, which can contain
more than one paragraph.

The code has been altered to use the `mwparserfromhtml`
library for determining the first section instead.

Bug: T375364