斗鱼直播实时数据爬取 标签： django数据爬虫 2017-05-30 23:46 229人阅读 评论(0)

思路

1, 解析URL

2, 利用爬虫神器 bs4 和 正则表达式得到想要的信息;
3, 进库和本地保存

DJango后台展示和本地CSV(卖相太差，不发了)

* 存储本地的CSV 直接运行 DySpyder().summary_data180() 即可*

直接上代码

# -*- coding: utf-8 -*-
import os
import re
import django
import urllib.request as ur

class DySpyder():

    def __init__(self):
        pass

    def open_url(self, url):
        headers = {'User-Agent': 'Mozilla/5.0 (Windows NT 6.1; WOW64; rv:23.0) Gecko/20100101 Firefox/23.0'}
        req = ur.Request(url=url, headers=headers)  # python2，urllib.request()
        response = ur.urlopen(req)  # python2，urllib2.urlopen()
        return response.read().decode('utf-8')

    def tv_spyder(self):
        url = "https://www.douyu.com/directory/all"
        data = self.open_url(url)
        from bs4 import BeautifulSoup
        cate = ['', 'serach_lastli', 'last','lastserach_lastli']  # - - s- l - ll 6loop
        soup1 = BeautifulSoup(data, 'html.parser')
        soup = soup1.find("ul", id='live-list-contentbox')
        res = []
        for c in cate:
            tmp = soup.findAll('li', c)
            res.extend(tmp)
        return res

    import datetime
    def set_data(self, x):
        import datetime
        res = {}
        # title.__init__
        title0 = str(x.find("h3").next_element)
        spans = x.findAll(["span"])
        # basic info to the link
        tag, dy_name, dy_num = tuple([s.next_element for s in spans][2:5])
        parterb = r'''.*<img data-original=(.*?) height="163" src=(.*?) width="283"/>.*'''
        # the urls of img and gif
        img, gif = re.findall(parterb, repr(x))[0]
        p2 = r'''.*a class="play-list-link" (.*?)=(.*?) (.*?)=(.*?) (.*?)=(.*?) (.*?)=(.*?) (.*?)=(.*?) (.*?)=(.*?) (.*?)=(.*?) (.*?)=(.*?)>.*'''
        t1 = [x for x in re.findall(p2, repr(x))][0]
        # the head of link-info
        for i in range(int(len(t1)/2 - 1)):
            res.setdefault(t1[2*i], t1[2*i+1])
        res.setdefault("dt", datetime.datetime.today())
        res.setdefault('tag', tag)
        res.setdefault('dy_name', dy_name)
        res.setdefault('dy_num', dy_num)
        res.setdefault('title0', title0)
        res.setdefault('img', img)
        res.setdefault('gif', gif)

        return res

    def summary_data180(self):
        l = [self.set_data(x) for x in self.tv_spyder()]
        import pandas as pd
        df_tmp = pd.DataFrame(l)
        df_tmp.to_csv("C:\\Users\\lenovo\\Desktop\\dy180.csv")
        return df_tmp

    #print(summary_data180())

    def main(self):
        os.environ.setdefault("DJANGO_SETTINGS_MODULE", "minicms.settings")
        django.setup()

        from tv.models import Info
        from django.utils import timezone

        df = self.summary_data180()
        print(df.columns)
        import numpy as np
        array2 = np.array(df)
        for i in range(len(df)):
            Info.objects.create(data_rid=array2[i][0],
                                data_rpos=array2[i][1],
                                data_sid=array2[i][2],
                                data_sub_rt=array2[i][3],
                                data_tid=array2[i][4],
                                dt=timezone.now(), ##修改了时间
                                dy_name=array2[i][6],
                                dy_num=array2[i][7],
                                gif=array2[i][8],
                                href=array2[i][9],
                                img=array2[i][10],
                                tag=array2[i][11],
                                target=array2[i][12],
                                title0=array2[i][13]
                                )
        print("执行完毕")


dyspyder = DySpyder()
#dyspyder.main()

后续

• 随着时间更新， 每隔 10min 自动一次到数据库——可以获取 Tag 或者用户的规律
• 增加 虎牙-战旗-龙珠 的数据
• 增加图片进库和自己定义的页面实时优化; 实现多直播平台的归一化推荐

Django 存库模板

from django.db import models

# Create your models here.

class Info(models.Model):
    data_rid = models.CharField("房间ID", max_length=20)
    data_rpos = models.CharField("", max_length=20)
    data_sid= models.CharField("", max_length=20)
    data_sub_rt = models.CharField("", max_length=20)
    data_tid = models.CharField("", max_length=20)
    dt = models.DateTimeField("时间")
    dy_name = models.CharField("账号名字", max_length=50)
    dy_num = models.CharField("观看数", max_length=20)
    gif= models.CharField("GIF", max_length=120)
    href = models.CharField("房间url", max_length=20)
    img = models.CharField("IMG_url", max_length=120)
    tag = models.CharField("标签", max_length=120)
    target = models.CharField("目标", max_length=20)
    title0 = models.CharField("标题", max_length=120)

    def __str__(self):
        return self.dy_name + "_" + self.title0

    class Meta:
        verbose_name = '斗鱼时间信息'
        verbose_name_plural = '斗鱼时间信息180条'

class ImgTools(models.Model):
    img_url = models.URLField(verbose_name="线上路径")
    dt = models.DateTimeField("时间")
    data_rid = models.CharField("房间ID", max_length=20)
    upload_to = models.URLField(verbose_name="本地路径")

TXT 爬取更新

def find_min(nums):
    for i in range(len(nums)):
        if nums[i+1] > nums[i]:
            return i, nums[i]

def set_urls(book_id):
    url = "http://www.biqudu.com/" + book_id + "/"
    partern = r".*<dd> <a href=(.*?)>(.*?)</a></dd>.*"
    import pandas as pd
    import numpy as np
    ## 本方法不能分卷， 后续补上相关的事情优化
    df1 = pd.DataFrame(np.array(re.findall(partern, open_url(url))), columns=["url", "title"])
    df1["num"] = [int(list(re.findall(r".*/(.*?).html", x))[0]) for x in df1["url"]]
    ####### find all-span
    start_index = find_min(df1["num"])[0]
    return df1[start_index: len(df1)]


# 为单独一个小说页面爬取;txt 文档; content;
def detail():
    url = "http://www.biqudu.com/21_21470/1394112.html"
    data = open_url(url)
    from bs4 import BeautifulSoup
    soup = BeautifulSoup(data, 'html.parser')
    content = soup.findAll('div', id="content")[0]
    return content

# print(detail())


def test(request):
    content = detail()
    return render(request, "base_test.html", {"content": content})

* 近期会花精力弄微信小程序， 爬虫放置一段时间。 ==== END ====*