qureta.py 文件源码

def parse(self, response):
        self.logger.info('parse: %s' % response)
        is_no_update = False

        # Get list of news from the current page
        articles = response.css('div.view-front > div.view-content > div.views-row')

        if not articles:
            raise CloseSpider('article not found')
        for article in articles:
            # Close the spider if we don't find the list of urls
            url_selectors = article.css('span.field-content a::attr(href)')
            if not url_selectors:
                raise CloseSpider('url_selectors not found')
            url = url_selectors.extract()[0]

            # Example '19 Oct 2016'
            info_selectors = article.css('span.field-content::text')
            if not info_selectors:
                raise CloseSpider('info_selectors not found')
            info_time = info_selectors.extract()[1].strip()

            # Parse date information
            try:
                published_at_wib = datetime.strptime(info_time, '%d %b %Y')
            except ValueError as e:
                raise CloseSpider('cannot_parse_date: %s' % e)

            published_at = wib_to_utc(published_at_wib)

            if self.media['last_scraped_at'] >= published_at:
                is_no_update = True
                break

            # For each url we create new scrapy request
            yield Request('http://www.qureta.com' + url, callback=self.parse_news)

        if is_no_update:
            self.logger.info('Media have no update')
            return

        if response.css('li.next'):
            next_page_url = response.css('li.next > a::attr(href)')[0].extract()
            yield Request('http://www.qureta.com' + next_page_url, callback=self.parse)

    # Collect news item