Why LinkExtractor does not catch the links that wa

Question

I'm crawling a page that generates data with infinite scrolling. I'm using CrawlSpider, and the rules are defined like this:

rules = (
    Rule(LinkExtractor(restrict_xpaths = ('//*some/xpaths')), callback = 'parse_first_itmes', follow = True),
    Rule(LinkExtractor(restrict_xpaths = ('//*some/other/xpaths')), callback = 'parse_second_itmes'),
)

In the parse_item function, I have a Request makes the AJAX requests:

def parse_first_items(self, response):
    l = ItemLoader(item = AmazonCnCustomerItem(), response = response)

    l.add_xpath('field1', '//p[@class="field1")]/text()')
    l.add_xpath('field2', '//p[@class="field2")]/text()')

    r_url = l.get_xpath('//*/url/xpath/@href')

    r = Request(url = req_url,
                    headers = {"Referer": "the/same/page/url",
                        "X-Requested-With": "XMLHttpRequest"},
                    callback = self.parse_first_items)

    return r, l.load_item()

I get the desired data just fine, but the LinkExtractor in the Second Rule does not catch the urls from the data generated by the Request inside the the parse_first_itmes function.

How can I make the LinkExtractor in the Second Rule extract those links and use them to parse the parse_second_itmes function?