爬虫实践---新浪微博爬取+json+csv
来源:互联网 发布:千千自动发卡源码 编辑:程序博客网 时间:2024/05/17 18:45
首先,奉上女神微博词云一张:
在此之前,我一直以为新浪微博的爬取,需要模拟登录等等
偶然之间,在小歪哥那里得知,有一个网站可以免登录爬取:https://m.weibo.cn/u/+oid,这个oid可以从普通新浪微博那里得到。
点击一个关注用户首页,查看其网页源码,源码页搜索用户名,就会看到如下的内容:
<script type="text/javascript">
var $CONFIG = {};
$CONFIG['islogin']='1';
$CONFIG['oid']='1280761142';
$CONFIG['page_id']='1003061280761142';
$CONFIG['onick']='刘雯';
这个oid就是url里面需要的那个,根据url打开微博链接:
打开开发者工具,这个类似于今日头条的页面,具体可参见今日头条的写作过程。
发现请求网页实际是:
打开上述链接,会发现网页内容并非常见的格式,需要来解析一下,推荐一个json格式在线解析网站:http://www.json.cn/,利用这个就可以看到json格式的网页内容:
import json,用来解析网页内容。接下来可以访问一个页面,看看我们的思路是否正确。
打开URl
https://m.weibo.cn/api/container/getIndex?type=uid&value=1280761142&containerid=1005051280761142
在XHR里面发现:
通过对该网页的json格式转换找到微博总数量和微博名称:
#!/usr/bin/env python# coding=utf-8import requestsimport bs4import jsonimport reimport randomimport csvheaders = {'User-Agent':'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/57.0.2987.133 Safari/537.36'}total_number = 0#https://m.weibo.cn/api/container/getIndex?type=uid&value=1280761142&containerid=1076031280761142#url = 'http://m.weibo.cn/u/1280761142'num = 0ip_list = []save_list =[]url_list = []oid = '1280761142'weibo_url = 'https://m.weibo.cn/api/container/getIndex?type=uid&value='+oidweibo_list = 'https://m.weibo.cn/api/container/getIndex?type=uid&value='+oid+'&containerid=1076031280761142'def get_html(url): try: r= requests.get(url,timeout=30,headers=headers) #如果状态码不是200,发出HTTOERROR r.raise_for_status() #设置正确的编码方式 r.encoding = r.apparent_encoding#r.encoding = 'utf-8' return r.json() except: return "Something Wrong!" def get_content(url): resp_data = get_html(url) userInfo = resp_data['userInfo'] total_number=userInfo['statuses_count'] name = userInfo['screen_name'] print(total_number) print(name) def main(): get_content(weibo_url)if __name__ == '__main__': main()
$ python 123sina\(副本\).py
3329
刘雯
然后,对于微博博文页进行访问,
https://m.weibo.cn/api/container/getIndex?type=uid&value=1280761142&containerid=1076031280761142&page=3
观察发现,其中多了一串数字containerid,然后我们改变page数字,查看是否能够获得微博列表。
答案是可以的,
然后改变页码,遍历微博博文:
texts_url = [] max_num = (int)(total_number/10) for i in range(1,max_num+1): temp = weibo_list+'&page=' + str(i) texts_url.append(temp)获取到每一个微博列表后,对每一个微博:'发表时间','博文','点赞数量','评论数量','转发数量','博文链接'
进行获取,由于存在有的链接并非博文,非博文的text_type不是9,进行判断后再进行获取;存在博文中@对象以及相关网页链接的存在,进行正则匹配后剔除:
text_cards = text_data['cards'] for text_info in text_cards: text={} #print(text_info) text_type = text_info["card_type"] if text_type != 9: continue text['博文链接'] = text_info['scheme'] text_mblog = text_info['mblog'] #print(text_mblog) num=num+1 #处理博文 text_dirty = text_mblog['text'] try: clean1 = re.sub(r"<span.*?</span>",'',text_dirty) text_dirty = re.sub(r'<a.*?</a>','',clean1) except: #print("无span链接") text_dirty = re.sub(r'<a.?*</a>','',text_dirty) pat=re.compile(r'[^\u4e00-\u9fa5]')#删除非中文字符 括号之类的 text_dirty = pat.sub('',text_dirty) text['博文'] = text_dirty.strip() # 获取时间 text['发表时间'] = text_mblog['created_at'] text['转发'] = text_mblog['reposts_count'] text['评论'] = text_mblog['comments_count'] text['点赞'] = text_mblog['attitudes_count']
由于重复频率的获取博文,避免IP被封,通过对西刺代理的获取进行,随机更换IP进行访问:
def get_ip_list(url): global ip_list web_data = requests.get(url, headers=headers) soup = bs4.BeautifulSoup(web_data.text, 'lxml') ips = soup.find_all('tr') for i in range(1, len(ips)): ip_info = ips[i] tds = ip_info.find_all('td') ip_list.append(tds[1].text + ':' + tds[2].text) def get_random_ip(): proxy_list = [] for ip in ip_list: proxy_list.append('http://' + ip) proxy_ip = random.choice(proxy_list) proxies = {'http': proxy_ip} return proxies
将获取到的信息,进行csv存储:
def save_weibo(text): title_csv = ['发表时间','博文','点赞数量','评论数量','转发数量','博文链接'] with open("weibo_liuwen.csv","w+") as csvfile: weibo_csv = csv.writer(csvfile) weibo_csv.writerow(title_csv) for weibo in save_list: text=[] text.append(weibo['发表时间']) text.append(weibo['博文']) text.append(weibo['点赞']) text.append(weibo['评论']) text.append(weibo['转发']) text.append(weibo['博文链接']) weibo_csv.writerow(text)
也可以对微博内容进行词云分析。
完整代码如下:
#!/usr/bin/env python# coding=utf-8import requestsimport bs4import jsonimport reimport randomimport csv# 设置等待时间,避免爬取太快import time# 用于在超时的时候抛出异常,便于捕获重连import socketheaders = {'User-Agent':'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/57.0.2987.133 Safari/537.36'}total_number = 0#https://m.weibo.cn/api/container/getIndex?type=uid&value=1280761142&containerid=1076031280761142#url = 'http://m.weibo.cn/u/1280761142'num = 0ip_list = []save_list =[]url_list = []oid = '1280761142'weibo_url = 'https://m.weibo.cn/api/container/getIndex?type=uid&value='+oidweibo_list = 'https://m.weibo.cn/api/container/getIndex?type=uid&value='+oid+'&containerid=1076031280761142'def get_ip_list(url): global ip_list web_data = requests.get(url, headers=headers) soup = bs4.BeautifulSoup(web_data.text, 'lxml') ips = soup.find_all('tr') for i in range(1, len(ips)): ip_info = ips[i] tds = ip_info.find_all('td') ip_list.append(tds[1].text + ':' + tds[2].text) def get_random_ip(): proxy_list = [] for ip in ip_list: proxy_list.append('http://' + ip) proxy_ip = random.choice(proxy_list) proxies = {'http': proxy_ip} return proxies# seleniumdef get_html(url): try: r= requests.get(url,headers=headers) #如果状态码不是200,发出HTTOERROR r.raise_for_status() #设置正确的编码方式 r.encoding = r.apparent_encoding#r.encoding = 'utf-8' return r.json() except: return "Something Wrong!"def save_weibo(text): title_csv = ['发表时间','博文','点赞数量','评论数量','转发数量','博文链接'] with open("weibo_liuwen.csv","w+") as csvfile: weibo_csv = csv.writer(csvfile) weibo_csv.writerow(title_csv) for weibo in save_list: text=[] text.append(weibo['发表时间']) text.append(weibo['博文']) text.append(weibo['点赞']) text.append(weibo['评论']) text.append(weibo['转发']) text.append(weibo['博文链接']) weibo_csv.writerow(text) ''' with open('刘雯weibo.txt','a') as f: for txt in text: f.write(txt) f.close() ''' def get_content(url): global total_number,num,save_list resp_data = get_html(url) userInfo = resp_data['userInfo'] total_number=userInfo['statuses_count'] name = userInfo['screen_name'] print(total_number) print(name) texts_url = [] max_num = (int)(total_number/10) for i in range(1,max_num+1): temp = weibo_list+'&page=' + str(i) texts_url.append(temp) data = { 'Referer':'https://m.weibo.cn/u/'+oid, } for text_url in texts_url: print(text_url) proxies = get_random_ip() print(proxies) # 超时重连 state = False timeout = 3 socket.setdefaulttimeout(timeout) while not state: try: r = requests.get(text_url,headers=headers,data=data,proxies=proxies) state = True except socket.timeout: print("超时重连") state = False proxies = get_random_ip() print(proxies) text_data = r.json() text_cards = text_data['cards'] for text_info in text_cards: text={} #print(text_info) text_type = text_info["card_type"] if text_type != 9: continue text['博文链接'] = text_info['scheme'] text_mblog = text_info['mblog'] #print(text_mblog) num=num+1 #处理博文 text_dirty = text_mblog['text'] try: clean1 = re.sub(r"<span.*?</span>",'',text_dirty) text_dirty = re.sub(r'<a.*?</a>','',clean1) except: #print("无span链接") text_dirty = re.sub(r'<a.?*</a>','',text_dirty) pat=re.compile(r'[^\u4e00-\u9fa5]')#删除非中文字符 括号之类的 text_dirty = pat.sub('',text_dirty) text['博文'] = text_dirty.strip() # 获取时间 text['发表时间'] = text_mblog['created_at'] text['转发'] = text_mblog['reposts_count'] text['评论'] = text_mblog['comments_count'] text['点赞'] = text_mblog['attitudes_count'] # 博文是否有图片 print(text['博文'],end='\t') print(text['发表时间']) save_list.append(text) save_weibo(text['博文']) #save_list.append(text['博文']) print("获取"+str(num)+"条博文...") # print(text_data)def main(): get_ip_list('http://www.xicidaili.com/nn/') #print(proxies) get_content(weibo_url) save_weibo(text)if __name__ == '__main__': main()
- 爬虫实践---新浪微博爬取+json+csv
- Python爬取新浪英超曼联文章内页--bs4,json,txt和csv以及编码
- 新浪微博小爬虫
- 新浪微博小爬虫
- 新浪新闻爬虫程序
- 新浪微博爬虫
- 新浪网页爬虫
- json 爬虫
- csv(txt)转json
- Python JSON、CSV
- Python爬虫小实践:寻找失踪人口,爬取失踪儿童信息并写成csv文件,方便存入数据库
- Python爬虫小实践:寻找失踪人口,爬取失踪儿童信息并写成csv文件,方便存入数据库
- 爬虫系列9CSV模块
- 新浪博客推广实践
- python开发新浪博客爬虫
- python调用新浪API爬虫
- 新浪微博爬虫策略
- Python新浪微博爬虫
- 操作系统高级教程系列_1
- 网络安全思维导图(全套11张)
- pycharm-professional-2017.2.3的激活方式
- Java 平台的三个版本
- C++ STL之deque
- 爬虫实践---新浪微博爬取+json+csv
- 黑暗之魂3 四大薪王的背景
- 补码、取反
- Vector和List的使用
- 【C51】单片机独立按键与矩阵按键
- Java垃圾回收
- SpringBoot+Maven项目实战(1):新建Maven项目
- ICPC北京网络赛-Minimum
- 文件对话框