kuaidaili.py 文件源码

def parse_page(self, response):
        pattern = re.compile(
                '<tr>\s.*?<td.*?>(.*?)</td>\s.*?<td.*?>(.*?)</td>\s.*?<td.*?>(.*?)</td>\s.*?<td.*?>('
                '.*?)</td>\s.*?<td.*?>(.*?)</td>\s.*?<td.*?>(.*?)</td>\s.*?<td.*?>(.*?)</td>\s.*?</tr>',
                re.S)
        items = re.findall(pattern, response.body)

        for item in items:
            proxy = Proxy()
            proxy.set_value(
                    ip = item[0],
                    port = item[1],
                    country = item[4],
                    anonymity = item[2],
                    source = self.name,
            )

            self.add_proxy(proxy)