python3scrapy模块爬取国家粮油信息中心的政策法规和产业信息标题、文章内容等信息到数据库
1、创建Scrapy项目
scrapy startproject Grain
2.进入项目目录,使用命令genspider创建Spider
scrapy genspider grain grainoil.com.cn
3、定义要抓取的数据(处理items.py文件)
# -*- coding: utf-8 -*-
import scrapy
class GrainItem(scrapy.Item):
# 文章的目录
news_path = scrapy.Field()
# 文章的分类
news_cate = scrapy.Field()
# 文章标题
news_title = scrapy.Field()
# 文章发布日期
news_date = scrapy.Field()
# 文章来源
news_source = scrapy.Field()
# 文章导读
news_guide = scrapy.Field()
# 文章内容
news_content = scrapy.Field()
# 文章链接
news_url = scrapy.Field()
4、编写提取item数据的Spider(在spiders文件夹下:grain.py)
# -*- coding: utf-8 -*-
# 爬取国家粮油信息中心的政策法规和产业信息标题、文章内容等信息
import scrapy
import re
from Grain.items import GrainItem
import requests
import os
from bs4 import BeautifulSoup
import datetime
class GrainSpider(scrapy.Spider):
name = 'grain'
allowed_domains = ['grainoil.com.cn']
start_urls = ['http://www.grainoil.com.cn/newsListHome/1471.jspx',
'http://www.grainoil.com.cn/newsListChannel/18.jspx'
]
def parse(self, response):
items = []
# 获取下一页链接,location.href=encodeURI('18_1.jspx');
next_url_list = response.xpath('//div[@class="container"]/div[@class="m_listpage"]/div[2]/div[@class="page-in"]/a/@onclick').extract()
# 利用正则匹配出18_1.jspx中的 18_1,最后再进行组合,有18.jspx下划线匹配一次或0次
patt = re.compile(r"location\.href=encodeURI\('(\d+_?\d+)\.jspx'\);")
for i in range(len(next_url_list)):
next_url = ''
# result结果为'18_1'
result = patt.search(next_url_list[i]).group(1)
# 如果是18开头的那么属于政策法规的下一页
if result[:2] == '18':
next_url = "http://www.grainoil.com.cn/newsListChannel/" + result + ".jspx"
# 如果是1471开头的属于产业信息的下一页
elif result[:4] == "1471":
next_url = "http://www.grainoil.com.cn/newsListHome/" + result + ".jspx"
yield scrapy.Request(url = next_url,callback=self.parse)
# 获取每个文章的链接地址
news_url = response.xpath('//div[@class="container"]/div[@class="m_listpage"]/div[2]/ol/li/a/@href').extract()
# 因为获取的文章链接地址中有:80需要去掉,不去也可以
for i in range(len(news_url)):
item = GrainItem()
# 这个是去掉:80的情况
# item['news_url'] = 'http://www.grainoil.com.cn' + news_url[i][29:]
item['news_url'] = news_url[i]
items.append(item)
for item in items:
yield scrapy.Request(url = item['news_url'],meta={'meta_1':item},callback=self.parse_news)
def parse_news(self,response):
item = GrainItem()
# 提取每次Response的meta数据
meta_1 = response.meta['meta_1']
try:
# 获取文章的标题,有空格
news_title = response.xpath('//div[@class="container"]/div[@class="m_details"]/div[@class="m_tit"]/h2/text()').extract()[0].strip()
item['news_title'] = news_title
# 获取文章导读
html = requests.get(meta_1['news_url'])
soup = BeautifulSoup(html.text,'lxml')
# 存储获取p标签的文章内容
content_list = []
for i in soup.select("p"):
content_list.append(i.get_text())
# content_list前两个是发布时间和字号大小,不需要
if len(content_list[2])!=0:
# content_list_2存储去掉每个p标签内容首尾空格后的文章内容
content_list_2 = []
# 去掉每段话首尾空格http://www.grainoil.com.cn/ChannelPolicies/90735.jhtml
# 每个列表后面六个是版权信息无用,为了避免文章内容字少导致导读里抓取到,直接去掉
for i in range(len(content_list[2:-6])):
content_list_2.append(content_list[2:][i].strip())
# 如果文章开头就是图片,会以#TRS_AUTOADD开头,直接导读是标题
if content_list_2[0][:12] != '#TRS_AUTOADD':
# 如果文章长度不超过100,考虑到列表中的每个内容首尾可能有空格、换行符、空白符,需要去掉
content_join = "".join(content_list_2).replace('\n','').replace(" ","").replace('\u3000','').strip()
if len(content_join) >= 100:
news_guide = content_join[:100] + "......"
else:
news_guide = content_join + "......"
else:
news_guide = news_title
else:
if len(content_list[3]) != 0:
if content_list[3][:11] != '中国粮食信息网版权所有':
content_list_3 = []
# 每个列表后面六个是版权信息无用,为了避免文章内容字少导致导读里抓取到,直接去掉
for i in range(len(content_list[3:-6])):
content_list_3.append(content_list[3:][i].strip())
# 如果文章长度不超过100
content_join = "".join(content_list_3).replace('\n','').replace(" ","").replace('\u3000','').strip()
if len(content_join) >= 100:
news_guide = content_join[:100] + "......"
else:
news_guide = content_join + "......"
# 还有文章内容可能是表格的情况,并不是p标签,直接获取标题为导读,获取到的p标签是版权信息
# http://www.grainoil.com.cn/ChannelPolicies/90729.jhtml
else:
news_guide = news_title
# content_list[2]和[3]为0可能为只有一张图片,导读直接是文章标题
# http://www.grainoil.com.cn/ChannelPolicies/90799.jhtml
else:
news_guide = news_title
item['news_guide'] = news_guide
# 获取文章的来源
news_source = response.xpath('//div[@class="container"]/div[@class="m_details"]/div[@class="m_tit"]/p[1]/a/text()').extract()[0].strip()
if len(news_source) != 0:
if news_source != '本站原创':
item['news_source'] = news_source
else:
item['news_source'] = '国家粮油信息中心'
else:
item['news_source'] = '国家粮油信息中心'
# 获取文章发布日期,需要调整 >>发布时间:2018-06-27 09:26:35 来源:
news_date_list = response.xpath('//div[@class="container"]/div[@class="m_details"]/div[@class="m_tit"]/p[1]/text()').extract()[0]
news_date = news_date_list.split(":")[1].split(" ")[0]
item['news_date'] = news_date
# 利用正则匹配文章的内容
html = requests.get(meta_1['news_url'])
# patt = re.compile(r'<div class="m_detailsbox" id="size">[\s\S]*?</div>')
patt = re.compile(r'(<div class="m_detailsbox" id="size">[\s\S]*?)<div class="m_detailsbox" style="padding-top')
result = patt.findall(html.text)
news_content = ''
for i in range(len(result)):
news_content += result[i]
# 发现文章中有个'oldsrc='有的是大写'OLDSRC=',会导致图片不显示,替换调
# http://www.grainoil.com.cn/ChannelPolicies/90284.jhtml
# 发现文章中有的图片路径地址不全都是src="/NewsDoc开头,需补全替换
# http://www.grainoil.com.cn/ChannelPolicies/13743.jhtml
# 有的文章结尾有符号']]>',去掉
# http://www.grainoil.com.cn/ChannelPolicies/15469.jhtml
item['news_content'] = news_content.replace('src="/NewsDoc','src="http://www.grainoil.com.cn/NewsDoc').replace('OLDSRC=','').replace('oldsrc=','').replace("]]>","").replace('href="javascript:void(0)"','').replace('宋体', '微软雅黑').replace('仿宋','微软雅黑').replace('Courier New','微软雅黑')
item['news_url'] = meta_1['news_url']
# 处理分类
cate = response.xpath('//div[@class="container"]/div[@class="m_crumbs"]/text()|//div[@class="container"]/div[@class="m_crumbs "]/text()').extract()[1].strip()
if cate == '政策法规':
item['news_cate'] = cate
# 创建目录的时候,不能包含一些特殊字符要转义
char_list = ['*','|',':','?','/','<','>','"','\\']
news_title_result = news_title
for i in char_list:
if i in news_title :
news_title_result = news_title.replace(i,"_")
news_path = './Data/'+ cate +"/" + news_date + "/" + news_title_result
# 如果目录不存在则创建
if (not os.path.exists(news_path)):
os.makedirs(news_path)
item['news_path'] = news_path
print("处理数据:%s" % news_path[7:])
elif cate == '生产气象'or cate == '综合信息':
item['news_cate'] = '产业信息'
# 创建目录的时候,不能包含一些特殊字符要转义
char_list = ['*', '|', ':', '?', '/', '<', '>', '"', '\\']
news_title_result = news_title
for i in char_list:
if i in news_title:
news_title_result = news_title.replace(i, "_")
news_path = './Data/产业信息/' + news_date + "/" + news_title_result
# 如果目录不存在则创建
if (not os.path.exists(news_path)):
os.makedirs(news_path)
item['news_path'] = news_path
print("处理数据:%s" % news_path[7:])
# d1 = datetime.datetime.strptime('2018-08-01', '%Y-%m-%d')
# d2 = datetime.datetime.strptime(item['news_date'], '%Y-%m-%d')
# if d2 > d1:
# print("%s日期大于2018-08-01数据处理中..." % item['news_date'])
yield item
# 发现这个网站有些文章是会员文章,无法查看
# http://www.grainoil.com.cn/newsListChannel/18_63.jspx以后
except:
print("会员文章无法查看!",meta_1['news_url'])
5.处理pipelines管道文件保存数据,可将结果保存到文件中(pipelines.py)
# -*- coding: utf-8 -*-
import json
# 转码操作,继承json.JSONEncoder的子类
class MyEncoder(json.JSONEncoder):
def default(self, o):
if isinstance(o, bytes):
return str(o, encoding='utf-8')
return json.JSONEncoder.default(self, o)
class GrainPipeline(object):
def process_item(self, item, spider):
try:
file_name = item['news_title']
file_name_result = file_name
# 创建文件名字的时候,不能包含一些特殊字符要转义
char_list = ['*', '|', ':', '?', '/', '<', '>', '"', '\\']
for i in char_list:
if i in file_name:
file_name_result = file_name.replace(i, "_")
with open(item['news_path'] + "/" + file_name_result + ".html", "w+")as f:
f.write(item['news_content'])
except Exception as e:
print(e)
print("[ERROR]%s文件保存失败,请注意!" % item['news_title'])
return item
def close_spider(self, spider):
print("数据保存本地处理完毕,谢谢使用!")
6.增加Grainpipelines.py文件,同时将数据保存到mysql数据库中
# -*- coding: utf-8 -*-
import json
import pymysql
# 转码操作,继承json.JSONEncoder的子类
class MyEncoder(json.JSONEncoder):
def default(self, o):
if isinstance(o, bytes):
return str(o, encoding='utf-8')
return json.JSONEncoder.default(self, o)
class DBPipeline(object):
@classmethod
def from_crawler(cls,crawler):
cls.MYSQL_HOST = crawler.settings.get('MYSQL_HOST')
cls.MYSQL_DBNAME = crawler.settings.get('MYSQL_DBNAME')
cls.MYSQL_USER = crawler.settings.get('MYSQL_USER')
cls.MYSQL_PASSWD = crawler.settings.get('MYSQL_PASSWD')
cls.MYSQL_PORT = crawler.settings.get('MYSQL_PORT')
cls.MYSQL_CHARSET = crawler.settings.get('MYSQL_CHARSET')
# 计数器,统计写入了多少
cls.count = 0
def process_item(self, item, spider):
try:
self.connect = pymysql.connect(host=self.MYSQL_HOST, db=self.MYSQL_DBNAME, user=self.MYSQL_USER, passwd=self.MYSQL_PASSWD,
port=self.MYSQL_PORT, charset=self.MYSQL_CHARSET)
# 通过cursor执行增删查改
self.cursor = self.connect.cursor()
# 查重处理
self.cursor.execute(
"""SELECT news_url FROM grainoil WHERE news_url = %s""", item['news_url'])
# 是否有重复数据
repetition = self.cursor.fetchone()
# 重复
if repetition:
print("数据库已有此条数据,不再添加,URL地址为:", repetition[0])
else:
print("写入数据库中...")
# 插入数据
self.cursor.execute(
"""INSERT INTO grainoil(news_cate,news_title, news_date, news_source, news_guide ,
news_content, news_url)VALUES(%s,%s, %s, %s, %s, %s, %s)""",
(item['news_cate'], item['news_title'], item['news_date'], item['news_source'],
item['news_guide'], item['news_content'], item['news_url']))
except Exception as error:
# 出现错误时打印错误日志
print("存入数据库错误为:", error)
self.connect.rollback()
# 如果没有捕获到异常,执行else中的语句
else:
self.count += 1
# 提交sql语句
self.connect.commit()
return item
def close_spider(self, spider):
self.cursor.close()
self.connect.close()
print("数据库处理完毕,本次共计增加%d条数据,谢谢使用!" % self.count)
7.配置settings文件(settings.py,这里通过设置调用数据库)
# Obey robots.txt rules,具体含义参照:https://www.cdsy.xyz/computer/programme/Python/241210/cd64912.html
ROBOTSTXT_OBEY = False
# 将数据保存在mysql
MYSQL_HOST = 'localhost'
MYSQL_DBNAME = 'python3'
MYSQL_USER = 'root'
MYSQL_PASSWD = '123456'
MYSQL_PORT = 3306
MYSQL_CHARSET = 'utf8'
# 下载延迟
DOWNLOAD_DELAY = 4
# Override the default request headers:添加User-Agent信息
DEFAULT_REQUEST_HEADERS = {
'User-Agent': 'Mozilla/5.0 (compatible; MSIE 9.0; Windows NT 6.1; Trident/5.0);',
# 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
# 'Accept-Language': 'en',
}
# Configure item pipelines去掉下面注释,打开管道文件
ITEM_PIPELINES = {
'Grain.pipelines.GrainPipeline': 100,
'Grain.Grainpipelines.DBPipeline': 300,
}
# 还可以将日志存到本地文件中(可选添加设置)
LOG_FILE = "grain.log"
LOG_LEVEL = "DEBUG"
# 包含打印信息也一起写进日志里
LOG_STDOUT = True
8.记得提前打开mysql数据库,并且建立好相应的表
# 创建国家粮油信息中心文章的数据库表
CREATE TABLE grainoil(id int PRIMARY KEY auto_increment not null,news_cate varchar(6),news_title varchar(100),news_date date,
news_source varchar(30),news_guide VARCHAR(150),news_content MEDIUMTEXT,news_url VARCHAR(90));
9.以上设置完毕,进行爬取:执行项目命令crawl,启动Spider:
scrapy crawl grain