jokes.py

    def parse(self, response):
        for joke in response.xpath("//div[@class='jokes']"):
            yield {
                'joke_text': joke.xpath(".//div[@class='joke-text']/p").extract_first()
            }