belkin.py 文件源码

def parse_kb(self, response):
        # initial html tokenization to find regions segmented by e.g. "======"
        # or "------"
        filtered = response.xpath(
            "//div[@class='sfdc_richtext']").extract()[0].split("=-")

        for entry in [x and x.strip() for x in filtered]:
            resp = HtmlResponse(url=response.url, body=entry,
                                encoding=response.encoding)

            for link in resp.xpath("//a"):
                href = link.xpath("@href").extract()[0]
                if "cache-www" in href:
                    text = resp.xpath("//text()").extract()
                    text_next = link.xpath("following::text()").extract()

                    item = FirmwareLoader(item=FirmwareImage(),
                                          response=response,
                                          date_fmt=["%b %d, %Y", "%B %d, %Y",
                                                    "%m/%d/%Y"])

                    version = FirmwareLoader.find_version_period(text_next)
                    if not version:
                        version = FirmwareLoader.find_version_period(text)

                    item.add_value("version", version)
                    item.add_value("date", item.find_date(text))
                    item.add_value("url", href)
                    item.add_value("product", response.meta["product"])
                    item.add_value("vendor", self.name)
                    yield item.load_item()