MATCH (m:Movie)
CALL apoc.load.html(m.url, { reviews: "div.article__body p" }) YIELD value
WITH split(value.reviews[0].text + " " + value.reviews[1].text, " ") as keywords, m
WITH [x IN keywords | apoc.text.replace(x, "[\\.\\,\\(\\)\\']", "")] as keywords, m
SET m.keywords = filter(x IN keywords WHERE size(x) > 4)