Search code examples
pythonscrapyrequestyield

(Scrapy) Generate Request / Generate & execute Request / Execute Request

I am studying scrapy examples at https://www.accordbox.com/blog/how-crawl-infinite-scrolling-pages-using-python/

Regarding to yield Request of the Scrapy solution code in there, I am very confused.

There are three yield Requests. Sometime a Request is just generated, sometime it is generated and executed, sometime it is just executed.

Could you advice me what are the differences between them please?

Thank you!.

def parse_list_page(self, response):
    next_link = response.xpath(
        "//a[@class='page-link next-page']/@href").extract_first()
    if next_link:
        url = response.url
        next_link = url[:url.find('?')] + next_link

        ################################
        # Generate and Execute Request
        ################################

        yield Request(                           
            url=next_link,
            callback=self.parse_list_page
        )

    for req in self.extract_product(response):
        ################################
        #Just Execute Request
        ################################
        yield req       

def extract_product(self, response):
    links = response.xpath("//div[@class='col-lg-8']//div[@class='card']/a/@href").extract()
    for url in links:
        result = parse.urlparse(response.url)
        base_url = parse.urlunparse(
            (result.scheme, result.netloc, "", "", "", "")
        )
        url = parse.urljoin(base_url, url)

        ################################
        #Just Generate Request
        ################################

        yield Request (
            url=url,
            callback=self.parse_product_page
        )

def parse_product_page(self, response):
    logging.info("processing " + response.url)
    yield None
Solution

  • You may find the figure here useful for answering your question.

    The yields from the parse_list_page method are yielding requests back to the "Engine" (step 7 in the figure). The yield in extract_product is yielding back to parse_list_page. parse_list_page then immediately yields them back to the engine.

    Note that all the code in extract_product could also just go into parse_list_page to make one single method. Having two methods just nicely separates the logic.

    scrapy data flow