爬虫:新浪微博爬虫的最简单办法

Posted 2020-12-12 chengrongkai

前言：本文主要内容是介绍如何用最简单的办法去采集新浪微博的数据，主要是采集指定微博用户发布的微博以及微博收到的回复等内容，可以通过配置项来调整爬取的微博用户列表以及其他属性。

既然说是最简单的办法，那么我们就得先分析微博爬虫可能选择的几个目标网址，首先肯定是最常见的web网站了

还有就是m站，也就是移动端网页

以及一个无法旧版本的访问入口了，首先可以排除web站了，这个是最麻烦的，它的请求是被js加密过，处理起来很麻烦

那我们为何不退而求其次呢，我们观察下这个m站的请求

可以发现在某个请求里可以发现我们需要的微博信息，既然这样就好办了，我们就可以着手我们的代码了

首先是获取用户信息，通过用户信息获取用户的微博总数，这样就可以知道总共多少页的数据了，代码如下所示

    def get_json(self, params):
        """获取网页中json数据"""
        url = ‘https://m.weibo.cn/api/container/getIndex?‘
        r = requests.get(url, params=params, cookies=self.cookie)
        return r.json()
    def get_page_count(self):
        """获取微博页数"""
        try:
            weibo_count = self.user[‘statuses_count‘]
            page_count = int(math.ceil(weibo_count / 10.0))
            return page_count
        except KeyError:
            sys.exit(u‘程序出错‘)
    def get_user_info(self):
        """获取用户信息"""
        params = {‘containerid‘: ‘100505‘ + str(weibo_config[‘user_id‘])}
        js = self.get_json(params)
        if js[‘ok‘]:
            info = js[‘data‘][‘userInfo‘]
            user_info = {}
            user_info[‘id‘] = weibo_config[‘user_id‘]
            user_info[‘screen_name‘] = info.get(‘screen_name‘, ‘‘)
            user_info[‘gender‘] = info.get(‘gender‘, ‘‘)
            user_info[‘statuses_count‘] = info.get(‘statuses_count‘, 0)
            user_info[‘followers_count‘] = info.get(‘followers_count‘, 0)
            user_info[‘follow_count‘] = info.get(‘follow_count‘, 0)
            user_info[‘description‘] = info.get(‘description‘, ‘‘)
            user_info[‘profile_url‘] = info.get(‘profile_url‘, ‘‘)
            user_info[‘profile_image_url‘] = info.get(‘profile_image_url‘, ‘‘)
            user_info[‘avatar_hd‘] = info.get(‘avatar_hd‘, ‘‘)
            user_info[‘urank‘] = info.get(‘urank‘, 0)
            user_info[‘mbrank‘] = info.get(‘mbrank‘, 0)
            user_info[‘verified‘] = info.get(‘verified‘, False)
            user_info[‘verified_type‘] = info.get(‘verified_type‘, 0)
            user_info[‘verified_reason‘] = info.get(‘verified_reason‘, ‘‘)
            user = self.standardize_info(user_info)
            self.user = user

分页采集数据

        page1 = 0
        random_pages = random.randint(1, 5)
        self.start_date = datetime.now().strftime(‘%Y-%m-%d‘)
        for page in tqdm(range(1, page_count + 1), desc=‘Progress‘):
            is_end = self.get_one_page(page)
            if is_end:
                break
            if page % 20 == 0:  # 每爬20页写入一次文件
                self.weibo_to_mysql(wrote_count)
                wrote_count = self.got_count

            # 通过加入随机等待避免被限制。爬虫速度过快容易被系统限制(一段时间后限
            # 制会自动解除)，加入随机等待模拟人的操作，可降低被系统限制的风险。默
            # 认是每爬取1到5页随机等待6到10秒，如果仍然被限，可适当增加sleep时间
            if (page - page1) % random_pages == 0 and page < page_count:
                sleep(random.randint(6, 10))
                page1 = page
                random_pages = random.randint(1, 5)

        self.weibo_to_mysql(wrote_count)  # 将剩余不足20页的微博写入文件
        print(u‘微博爬取完成，共爬取%d条微博‘ % self.got_count)

具体采集单页微博代码如下

    def get_one_page(self, page):
        """获取一页的全部微博"""
        try:
            js = self.get_weibo_json(page)
            if js[‘ok‘]:
                weibos = js[‘data‘][‘cards‘]
                for w in weibos:
                    if w[‘card_type‘] == 9:
                        wb = self.get_one_weibo(w)
                        if wb:
                            if wb[‘id‘] in self.weibo_id_list:
                                continue
                            created_at = datetime.strptime(
                                wb[‘created_at‘], ‘%Y-%m-%d‘)
                            since_date = datetime.strptime(
                                self.since_date, ‘%Y-%m-%d‘)
                            if created_at < since_date:
                                if self.is_pinned_weibo(w):
                                    continue
                                else:
                                    print(u‘{}已获取{}({})的第{}页微博{}‘.format(
                                        ‘-‘ * 30, self.user[‘screen_name‘],
                                        self.user[‘id‘], page, ‘-‘ * 30))
                                    return True
                            if (‘retweet‘ not in wb.keys()):
                                self.weibo.append(wb)
                                self.weibo_id_list.append(wb[‘id‘])
                                self.got_count += 1
            print(u‘{}已获取{}({})的第{}页微博{}‘.format(‘-‘ * 30,
                                                    self.user[‘screen_name‘],
                                                    self.user[‘id‘], page,
                                                    ‘-‘ * 30))
        except Exception as e:
            print("Error: ", e)
            traceback.print_exc()

获取具体微博信息的代码

    def get_one_weibo(self, info):
        """获取一条微博的全部信息"""
        try:
            weibo_info = info[‘mblog‘]
            weibo_id = weibo_info[‘id‘]
            retweeted_status = weibo_info.get(‘retweeted_status‘)
            is_long = weibo_info.get(‘isLongText‘)
            if retweeted_status:  # 转发
                retweet_id = retweeted_status.get(‘id‘)
                is_long_retweet = retweeted_status.get(‘isLongText‘)
                if is_long:
                    weibo = self.get_long_weibo(weibo_id)
                    if not weibo:
                        weibo = self.parse_weibo(weibo_info)
                else:
                    weibo = self.parse_weibo(weibo_info)
                if is_long_retweet:
                    retweet = self.get_long_weibo(retweet_id)
                    if not retweet:
                        retweet = self.parse_weibo(retweeted_status)
                else:
                    retweet = self.parse_weibo(retweeted_status)
                retweet[‘created_at‘] = self.standardize_date(
                    retweeted_status[‘created_at‘])
                weibo[‘retweet‘] = retweet
            else:  # 原创
                if is_long:
                    weibo = self.get_long_weibo(weibo_id)
                    if not weibo:
                        weibo = self.parse_weibo(weibo_info)
                else:
                    weibo = self.parse_weibo(weibo_info)
            weibo[‘created_at‘] = self.standardize_date(
                weibo_info[‘created_at‘])
            return weibo
        except Exception as e:
            print("Error: ", e)
            traceback.print_exc()

    def get_long_weibo(self, id):
        """获取长微博"""
        for i in range(5):
            url = ‘https://m.weibo.cn/detail/%s‘ % id
            html = requests.get(url, cookies=self.cookie).text
            html = html[html.find(‘"status":‘):]
            html = html[:html.rfind(‘"hotScheme"‘)]
            html = html[:html.rfind(‘,‘)]
            html = ‘{‘ + html + ‘}‘
            js = json.loads(html, strict=False)
            weibo_info = js.get(‘status‘)
            if weibo_info:
                weibo = self.parse_weibo(weibo_info)
                return weibo
            sleep(random.randint(6, 10))

以上就是核心的微博信息采集代码了，除了微博信息，我们还需要采集微博评论信息，原理是一样的，找到数据来源

有了微博信息采集的经验，我们很容易就可以找到我们想要的那个接口

具体代码如下

def add_comments_json(self,jsondata):
        for data in jsondata:
            item = dict()
            item[‘id‘] = data.get(‘id‘)
            item[‘mid‘] = data.get(‘mid‘)
            item[‘like_count‘] = data.get("like_count")
            item[‘source‘] = data.get("source")
            item[‘floor_number‘] = data.get("floor_number")
            item[‘screen_name‘] = data.get("user").get("screen_name")
            # 性别
            item[‘gender‘] = data.get("user").get("gender")
            if(item[‘gender‘] == ‘m‘):
                item[‘gender‘] = ‘男‘
            elif(item[‘gender‘] == ‘f‘):
                item[‘gender‘] = ‘女‘
            item[‘rootid‘] = data.get("rootid")
            item[‘create_time‘] = data.get("created_at")
            import time
            item[‘create_time‘] = time.strptime(item[‘create_time‘], ‘%a %b %d %H:%M:%S %z %Y‘)
            item[‘create_time‘] = time.strftime(‘%Y-%m-%d‘,item[‘create_time‘])
            item[‘comment‘] = data.get("text")
            item[‘comment‘] = BeautifulSoup(item[‘comment‘], ‘html.parser‘).get_text()
            item[‘comment‘] = self.clear_character_chinese(item[‘comment‘])
            print(‘当前楼层{},评论{}‘.format(item[‘floor_number‘],item[‘comment‘]))
            # 评论这条评论的信息
            comments = data.get("comments")
            if(comments):
                self.add_comments_json(comments)
            # print jsondata.dumps(comment, encoding="UTF-8", ensure_ascii=False)
            self.comments.append(item)
            
    def get_comments_page(self,max_id, id_type,mid):
        from get_weibo_cookie import get_cookie
        params = {
            ‘max_id‘: max_id,
            ‘max_id_type‘: id_type
            }
        try:
            url = ‘https://m.weibo.cn/comments/hotflow?id={id}&mid={mid}&max_id=‘
            headers = {
            ‘Cookie‘: ‘T_WM=96849642965; __guid=52195957.2500582256236055600.1583058027995.9556; WEIBOCN_FROM=1110006030; SCF=Aimq85D9meHNU4Ip0PFUjYBTDjXFB0VtQr3EKoS8DHQDobRNUO3lDIufAcUg69h4J7BQWqryxQpuU3ReIHHxvQ4.; SUBP=0033WrSXqPxfM725Ws9jqgMF55529P9D9W5H0p180lDMiCjNvXD_-uOh5JpX5KzhUgL.FoM0S0n0eo-0Sh.2dJLoI0qLxKqL1KMLBK-LxK-LBonLBonLxKMLB.-L12-LxK-LBK-LBoeLxK-L1hnL1hqLxKBLB.2LB-zt; XSRF-TOKEN=ca0a29; SUB=_2A25zWlwFDeRhGeFN7FoS8ivPzzWIHXVQpWRNrDV6PUJbkdANLW_9kW1NQ8CH90H5f8j5r1NA4GNPvu6__ERL-Jat; SUHB=0vJIkXXtLIIaZO; SSOLoginState=1583230037; MLOGIN=1; M_WEIBOCN_PARAMS=oid%3D4474164293517551%26luicode%3D20000174%26lfid%3D102803%26uicode%3D20000174; monitor_count=45‘,
            ‘User-Agent‘: ‘Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/70.0.3538.110 Safari/537.36‘,
            ‘X-Requested-With‘: ‘XMLHttpRequest‘
        }
            r = requests.get(url.format(id=mid,mid=mid), params=params,headers=headers)
            print(r.url)
            if r.status_code == 200:
                return r.json()
        except requests.ConnectionError as e:
            print(‘error‘, e.args)    
    
    def add_comments(self,jsondata):
        datas = jsondata.get(‘data‘).get(‘data‘)
        for data in datas:
            item = dict()
            item[‘id‘] = data.get(‘id‘)
            item[‘mid‘] = data.get(‘mid‘)
            item[‘like_count‘] = data.get("like_count")
            item[‘source‘] = data.get("source")
            item[‘floor_number‘] = data.get("floor_number")
            item[‘screen_name‘] = data.get("user").get("screen_name")
            # 性别
            item[‘gender‘] = data.get("user").get("gender")
            if(item[‘gender‘] == ‘m‘):
                item[‘gender‘] = ‘男‘
            elif(item[‘gender‘] == ‘f‘):
                item[‘gender‘] = ‘女‘
            item[‘created_at‘] = self.standardize_date(
                data.get([‘created_at‘]))
            import time
            item[‘create_time‘] = time.strptime(item[‘create_time‘], ‘%a %b %d %H:%M:%S %z %Y‘)
            item[‘create_time‘] = time.strftime(‘%Y-%m-%d‘,item[‘create_time‘])
            item[‘rootid‘] = data.get("rootid")
            
            item[‘comment‘] = data.get("text")
            item[‘comment‘] = BeautifulSoup(item[‘comment‘], ‘html.parser‘).get_text()
            item[‘comment‘] = self.clear_character_chinese(item[‘comment‘])
            print(‘当前楼层{},评论{}‘.format(item[‘floor_number‘],item[‘comment‘]))
            # 评论这条评论的信息
            comments = data.get("comments")

            # print jsondata.dumps(comment, encoding="UTF-8", ensure_ascii=False)
            self.comments.append(item)

我们可以查看下采集到的数据，如下所示

完整代码可以去我的开源项目中查看或者下载，欢迎star，或者留言与我进行交流。

https://gitee.com/chengrongkai/OpenSpiders

本文收发于https://www.bizhibihui.com/blog/article/44