relevant_pages in the nlp.py module currently returns pages that contain the query in its entirety, but according to the textbook (p.872) it should do it "by intersecting hit lists of query words".

My understanding is that for query = "green apples" and pages p1 = "oranges and bananas", p2 = "red and green apples", and p3 = "plums and apples" relevant_pages(query) should return p2 and p3, not just p2 as it currently does.

I can make this quick fix and change the corresponding test if someone confirms that it is indeed an intersection that we want.