爬虫库
使用简单的requests库,这是一个阻塞的库,速度比较慢。
解析使用XPATH表达式
总体采用类的形式
多线程
使用concurrent.future并发模块,建立线程池,把future对象扔进去执行即可实现并发爬取效果
数据存储
使用Python ORM sqlalchemy保存到数据库,也可以使用自带的csv模块存在CSV中。
API接口
因为API接口存在数据保护情况,一个电影的每一个分类只能抓取前25页,全部评论、好评、中评、差评所有分类能爬100页,每页有20个数据,即最多为两千条数据。
因为时效性原因,不保证代码能爬到数据,只是给大家一个参考思路,上代码:
from datetime import datetime
import random
import csv
from concurrent.futures import ThreadPoolExecutor, as_completed
from lxml import etree
import pymysql
import requests
from models import create_session, Comments
#随机UA
USERAGENT = [
'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/535.1 (KHTML, like Gecko) Chrome/14.0.835.163 Safari/535.1',
'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/73.0.3683.103 Safari/537.36',
'Mozilla/5.0 (Windows NT 6.1; WOW64; rv:6.0) Gecko/20100101 Firefox/6.0',
'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/534.50 (KHTML, like Gecko) Version/5.1 Safari/534.50',
'Opera/9.80 (Windows NT 6.1; U; zh-cn) Presto/2.9.168 Version/11.50',
'Mozilla/5.0 (Windows; U; Windows NT 6.1; ) AppleWebKit/534.12 (KHTML, like Gecko) Maxthon/3.0 Safari/534.12'
]
class CommentFetcher:
headers = {'User-Agent': ''}
cookie = ''
cookies = {'cookie': cookie}
# cookie为登录后的cookie,需要自行复制
base_node = '//div[@class="comment-item"]'
def __init__(self, movie_id, start, type=''):
'''
:type: 全部评论:'', 好评:h 中评:m 差评:l
:movie_id: 影片的ID号
:start: 开始的记录数,0-480
'''
self.movie_id = movie_id
self.start = start
self.type = type
self.url = 'https://movie.douban.com/subject/{id}/comments?start={start}&limit=20&sort=new_score\&status=P&percent_type={type}&comments_only=1'.format(
id=str(self.movie_id),
start=str(self.start),
type=self.type
)
#创建数据库连接
self.session = create_session()
#随机useragent
def _random_UA(self):
self.headers['User-Agent'] = random.choice(USERAGENT)
#获取api接口,使用get方法,返回的数据为json数据,需要提取里面的HTML
def _get(self):
self._random_UA()
res = ''
try:
res = requests.get(self.url, cookies=self.cookies, headers=self.headers)
res = res.json()['html']
except Exception as e:
print('IP被封,请使用 * ')
print('正在获取{} 开始的记录'.format(self.start))
return res
def _parse(self):
res = self._get()
dom = etree.HTML(res)
#id号
self.id = dom.xpath(self.base_node + '/@data-cid')
#用户名
self.username = dom.xpath(self.base_node + '/div[@class="avatar"]/a/@title')
#用户连接
self.user_center = dom.xpath(self.base_node + '/div[@class="avatar"]/a/@href')
#点赞数
self.vote = dom.xpath(self.base_node + '//span[@class="votes"]/text()')
#星级
self.star = dom.xpath(self.base_node + '//span[contains(@class,"rating")]/@title')
#发表时间
self.time = dom.xpath(self.base_node + '//span[@class="comment-time "]/@title')
#评论内容 所有span标签class名为short的节点文本
self.content = dom.xpath(self.base_node + '//span[@class="short"]/text()')
#保存到数据库
def save_to_database(self):
self._parse()
for i in range(len(self.id)):
try:
comment = Comments(
id=int(self.id[i]),
username=self.username[i],
user_center=self.user_center[i],
vote=int(self.vote[i]),
star=self.star[i],
time=datetime.strptime(self.time[i], '%Y-%m-%d %H:%M:%S'),
content=self.content[i]
)
self.session.add(comment)
self.session.commit()
return 'finish'
except pymysql.err.IntegrityError as e:
print('数据重复,不做任何处理')
except Exception as e:
#数据添加错误,回滚
self.session.rollback()
finally:
#关闭数据库连接
self.session.close()
#保存到csv
def save_to_csv(self):
self._parse()
f = open('comment.csv', 'w', encoding='utf-8')
csv_in = csv.writer(f, dialect='excel')
for i in range(len(self.id)):
csv_in.writerow([
int(self.id[i]),
self.username[i],
self.user_center[i],
int(self.vote[i]),
self.time[i],
self.content[i]
])
f.close()
if __name__ == '__main__':
with ThreadPoolExecutor(max_workers=4) as executor:
futures = []
for i in ['', 'h', 'm', 'l']:
for j in range(25):
fetcher = CommentFetcher(movie_id=26266893, start=j * 20, type=i)
futures.append(executor.submit(fetcher.save_to_csv))
for f in as_completed(futures):
try:
res = f.done()
if res:
ret_data = f.result()
if ret_data == 'finish':
print('{} 成功保存数据'.format(str(f)))
except Exception as e:
f.cancel()
来源:https://www.cnblogs.com/qingdeng123/p/11678245.html
标签:python,多,线程,爬取,豆瓣影评,api,接口