hgg070/hgg070_spider/spiders/roll_zuqiu.py

# -*- coding: utf-8 -*-
import logging
# import lxml
import scrapy
import xmltodict

from ..items import ZuqiuItem


class ZuqiuSpider(scrapy.Spider):
    name = 'roll_zuqiu'
    allowed_domains = ['m.hgg070.com']
    custom_settings = {
        "ITEM_PIPELINES": {
            "hgg070_spider.pipelines.roll_zuqiu.RollPipeline": 200,
        },
        # 'LOG_LEVEL': 'DEBUG',
        # 'LOG_FILE': cpath + "/log/sports_{}_{}_{}.log".format(to_day.year, to_day.month, to_day.day)
    }

    headers = {
        'Accept': '*/*',
        'Accept-Encoding': 'gzip, deflate',
        'Accept-Language': 'zh-CN,zh;q=0.9',
        'Content-Length': '130',
        'Content-type': 'application/x-www-form-urlencoded',
        'Cookie': '_ga=GA1.2.1009358217.1572056223; _gid=GA1.2.97506800.1572056223; _gat=1',
        'Host': 'm.hgg070.com',
        'Origin': 'http://m.hgg070.com',
        'Proxy-Connection': 'keep-alive',
        'Referer': 'http://m.hgg070.com/',
        'User-Agent': 'Mozilla/5.0 (iPhone; CPU iPhone OS 11_0 like Mac OS X) AppleWebKit/604.1.38 (KHTML, like Gecko) Version/11.0 Mobile/15A372 Safari/604.1'
    }

    def start_requests(self):
        url = "http://m.hgg070.com/app/member/get_league_list.php"
        from_data = {
            'uid': '9965a18b03dc6aacf12290bd5b8267fd7e38ec76eadf899b6108e82da5974cdd',
            'langx': 'zh-cn',
            'ltype': '3',
            'gtype': 'FT',
            'showtype': 'RB',
            'sorttype': '',
            'date': '',
            'isP': ''
        }
        yield scrapy.FormRequest(url=url, formdata=from_data, callback=self.parse, headers=self.headers, dont_filter=True)

    def parse(self, response):
        leagues = response.xpath('//serverresponse/game/league')
        url = 'http://m.hgg070.com/app/member/get_game_list.php'
        if leagues:
            self.headers['Content-Length'] = '141'
            for league in leagues:
                lid = league.xpath('.//league_id/text()').extract_first()
                from_data = {
                    'uid': '9965a18b03dc6aacf12290bd5b8267fd7e38ec76eadf899b6108e82da5974cdd',
                    'langx': 'zh-cn',
                    'ltype': '3',
                    'gtype': 'FT',
                    'showtype': 'RB',
                    'lid': lid,
                    'sorttype': '',
                    'date': '',
                    'isP': ''
                }
                yield scrapy.FormRequest(url=url, formdata=from_data, callback=self.parse_match, headers=self.headers, dont_filter=True)
        else:
            print('未获取到联赛id')
            return

    def parse_match(self, response):
        url = 'http://m.hgg070.com/app/member/get_game_more.php'
        self.headers['Content-Length'] = '132'
        gids = response.xpath('//serverresponse/game/gid/text()').extract()
        tags = response.xpath('//serverresponse/game/more_count/text()').extract()
        if gids:
            for i, gid in enumerate(gids):
                from_data = {
                    'uid': '9965a18b03dc6aacf12290bd5b8267fd7e38ec76eadf899b6108e82da5974cdd',
                    'langx': 'zh-cn',
                    'ltype': '3',
                    'gtype': 'FT',
                    'showtype': 'RB',
                    'date': '',
                    'isP': '',
                    'gid': gid,
                }
                tag = tags[i]
                yield scrapy.FormRequest(url=url, formdata=from_data, callback=self.parse_odds, headers=self.headers,
                                         meta={'tag': tag}, dont_filter=True)

    def parse_odds(self, response):
        logger = logging.getLogger(__name__)
        tag = response.meta['tag']
        game = xmltodict.parse(response.text)
        try:
            game_odds = game['serverresponse']['game'][0]
        except:
            game_odds = game['serverresponse']['game']
        if game_odds['gopen'] == 'Y':
            item = ZuqiuItem()
            item['data'] = game_odds
            item['tag'] = tag
            yield item
        else:
            logger.info('gopen == "N", 详细赔率盘口未开启')
            return