基于百度AI的自然语言处理文字分类

前言：

需要在百度AI平台注册登录并创建项目。

爬虫代码

 1 import scrapy
 2 from BaiDuAi.items import  BaiduaiItem
 3 
 4 class AiSpider(scrapy.Spider):
 5     name = 'ai'
 6     # allowed_domains = ['www.xxx.com']
 7     #人民网url
 8     start_urls = ['http://politics.people.com.cn/n1/2018/1217/c1001-30470023.html']
 9 
10     def parse(self, response):
11         title=response.xpath('/html/body/div[4]/h1/text()').extract_first()
12         content=response.xpath('//*[@id="rwb_zw"]//text()').extract()
13         content=''.join(content).strip('\n \t')
14         item=BaiduaiItem()
15         item['title']=title
16         item['content']=content
17 
18         yield  item
爬虫代码

管道代码

 1 from aip import AipNlp
 2 
 3 """ 你的 APPID AK SK """
 4 APP_ID = '15198150'
 5 API_KEY = 'jaObSr6rmSmqsjWfKGGpmwxB'
 6 SECRET_KEY = '808Eiz4FPkfMwS2ajClXYhKrcFMN1YUN'
 7 
 8 client = AipNlp(APP_ID, API_KEY, SECRET_KEY)
 9 
10 class BaiduaiPipeline(object):
11     keys=[]
12     def process_item(self, item, spider):
13         title=item['title'].replace('\xa0','')
14         content=item['content'].replace('\xa0','')
15         keys_dict=client.keyword(title,content)
16         for dic in keys_dict['items']:
17             self.keys.append(dic['tag'])
18 
19 
20         keys="/".join(self.keys)
21         typec_dic=client.topic(title,content)
22         news_type=typec_dic['item']['lv1_tag_list'][0]['tag']
23 
24         with open('./xinwen.html','w',encoding='utf-8')as fp:
25             fp.write(title+'\n\n'+content+'\n\n'+keys+'\n\n'+news_type)
26         return item
管道

 

posted @ 2018-12-18 17:48  北伽  阅读(933)  评论(0)  编辑  收藏  举报