hgg070/hgg070_spider/spiders/zuqiu.py

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

import scrapy

from items import ZuqiuItem


class ZuqiuSpider(scrapy.Spider):
    name = 'zuqiu'
    allowed_domains = ['m.hgg070.com']
    custom_settings = {
        "ITEM_PIPELINES": {
            "hgg070_spider.pipelines.zuqiu.ZuqiuPipeline": 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"
        h_types = [('FT', '', '130'), ('FU', 'P', '131'), ('FU', "", '130'), ('FU', 'P', '131')]
        for h_type in h_types:
            # show_type, isp, length = h_type
            show_type, isp, length = h_types[3]
            self.headers['Content-Length'] = length
            from_data = {
                'uid': '013dc3a00cbd488238236010f78ab4a41af7e6ff05ceb96bc0854b60807a42eb',
                'langx': 'zh-cn',
                'ltype': '3',
                'gtype': 'FT',
                'showtype': show_type,
                'sorttype': '',
                'date': '',
                'isP': isp
            }
            yield scrapy.FormRequest(url=url, formdata=from_data, callback=self.parse, headers=self.headers,
                                     meta={'showtype': show_type, 'isp': isp}, dont_filter=True)

    def parse(self, response):
        if response.status == 400:
            print(response.status)
            print('parse', response.url)
        leagues = response.xpath('//serverresponse/game/league')
        url = 'http://m.hgg070.com/app/member/get_game_list.php'
        if leagues:
            showtype = response.meta['showtype']
            isp = response.meta['isp']
            if showtype == 'FT' and isp == '':
                date = ''
                self.headers['Content-Length'] = '147'
            elif showtype == 'FU' and isp == 'P':
                date = 'all'
                self.headers['Content-Length'] = '151'
            elif showtype == 'FU' and isp == '':
                date = 'all'
                self.headers['Content-Length'] = '150'
            else:
                date = 'all'
                self.headers['Content-Length'] = '151'
            for league in leagues:
                lid = league.xpath('.//league_id/text()').extract_first()
                from_data = {
                    'uid': '013dc3a00cbd488238236010f78ab4a41af7e6ff05ceb96bc0854b60807a42eb',
                    'langx': 'zh-cn',
                    'ltype': '3',
                    'gtype': 'FT',
                    'showtype': showtype,
                    'lid': lid,
                    'sorttype': 'league',
                    'date': date,
                    'isP': isp
                }
                yield scrapy.FormRequest(url=url, formdata=from_data, callback=self.parse_match, headers=self.headers,
                                         meta={'showtype': showtype, 'isp': isp}, dont_filter=True)
        else:
            print('未获取到联赛id')
            return

    def parse_match(self, response):
        if response.status == 400:
            print(response.status)
            print('parse_match', response.url)
        url = 'http://m.hgg070.com/app/member/get_game_more.php'
        showtype = response.meta['showtype']
        isp = response.meta['isp']
        if showtype == 'FT' and isp == '':
            date = ''
            self.headers['Content-Length'] = '132'
        elif showtype == 'FU' and isp == 'P':
            date = 'all'
            self.headers['Content-Length'] = '136'
        elif showtype == 'FU' and isp == '':
            date = ''
            self.headers['Content-Length'] = '132'
        else:
            date = 'all'
            self.headers['Content-Length'] = '136'
        gids = response.xpath('//serverresponse/game/gid/text()').extract()
        if gids:
            for gid in gids:
                from_data = {
                    'uid': '013dc3a00cbd488238236010f78ab4a41af7e6ff05ceb96bc0854b60807a42eb',
                    'langx': 'zh-cn',
                    'ltype': '3',
                    'gtype': 'FT',
                    'showtype': showtype,
                    'date': date,
                    'isP': isp,
                    'gid': gid,
                }
                yield scrapy.FormRequest(url=url, formdata=from_data, callback=self.parse_odds, headers=self.headers,
                                         meta={'showtype': showtype, 'isp': isp}, dont_filter=True)

    def parse_odds(self, response):
        # print(response.text)
        # game_lists = []
        if response.status == 400:
            print(response.status)
            print('parse_odds', response.url)
        game = response.xpath('//serverresponse/game')[0]
        logger = logging.getLogger(__name__)
        if game:
            game_odds = {}
            gopen = game.xpath('//game/gopen/text()').extract_first()
            if gopen == 'Y':
                game = lxml.etree.fromstring(game.extract())
                for i in game.getchildren():
                    if i.text == None:
                        game_odds[i.tag] = ""
                    else:
                        game_odds[i.tag] = i.text
            else:
                logger.info('gopen == "N", 详细赔率盘口未开启')
            item = ZuqiuItem()
            item['all'] = game_odds
            yield item