Optimize search precision

[elisamussumeci]

Currently we're querying in the link_content, which contains the whole html of the page and includes some stuff that we're not interested at, like comments, ads etc. This affects the precision of the search.

We should be searching the query only on the cleaned text of the article. One way of doing this is using the Goose library to clean the link_content and then start making the search on that cleaned text.