1. Scrapy创建爬虫项目

1.1 Scrapy框架介绍

Scrapy是一个基于Twisted的异步处理框架，是纯Python实现的爬虫框架，其模块间的耦合程度低，可扩展性极强，可灵活完成各种需求。

1.1.1 架构介绍

Scrapy架构图
分为以下几个部分：

Engine
item
Scheduler
Downloader
Spiders
Item Pipeline
Downkoader Middlewares
Spider Middlewares

1.1.2 数据流

1.1.3 项目结构

Scrapy框架是通过命令行来创建项目，用IDE(Pycharm等)编写代码。项目创建之后，项目文件结构如下：

scrapy.cfg  # 项目配置文件，定义配置文件路径、部署相关信息
project/
	__init__.py
	items.py   # 定义Item的数据结构
	pipelines.py  # 数据的存储
	settings.py   # 项目的全局配置
	middlewares.py  # 中间件，一般不用修改
	spiders/
		__init__.py
		***.py   # 爬虫主程序

1.2 创建项目

在之前，已经在anaconda虚拟环境下安装好了Scrapy库、pymysql库，MySQL、MongoDB数据库。

1.2.1 创建项目

直接用Scrapy命令生成项目文件：
在命令行中cd目标文件夹，scrapy startproject sina

1.2.2 创建Spider

spider是自己定义的类。Scrapy用它来从网页里抓取内容，并解析抓取的结果。此类必须继承Scrapy提供的Spider类scrapy.Spider，还要定义Spider的名称和起始要求，以及怎样处理爬取后的结果的方法。

cd sina
scrapy genspider sina sina.com.cn

在这里插入图片描述

1.2.3 创建Item

Item是保存数据的容器，使用方法和字典类似。

创建Item需要继承scrapy.Item类，并且定义类型为scrapy.Field的字段。打开新浪新闻网站，看一看，可以获取到的内容有title、times、context、type等。

import scrapy

class SinaItem(scrapy.Item):
    # define the fields for your item here like:
    # name = scrapy.Field()
    pass

class GuoneiItem(scrapy.Item):
    # define the fields for your item here like:
    title = scrapy.Field()
    desc = scrapy.Field()
    times = scrapy.Field()

class ZongyiItem(scrapy.Item):
    # define the fields for your item here like:
    title = scrapy.Field()
    desc = scrapy.Field()
    times = scrapy.Field()

class DataItem(scrapy.Item):
    # define the fields for your item here like:
    title = scrapy.Field()
    desc = scrapy.Field()
    times = scrapy.Field()
    type = scrapy.Field()

2. 爬取数据

解析Response

import scrapy
from scrapy.http import Request
from scrapy.selector import Selector
from selenium import webdriver
import datetime
import re
from sina.items import GuoneiItem
from sina.items import DataItem

class Sina2Spider(scrapy.Spider):
    name = 'sina2'
    allowed_domains = ['sina.com.cn']

    def __init__(self, *args, **kwargs):
        super(Sina2Spider, self).__init__(*args, **kwargs)
        self.pages = 2
        self.flag = 0
        self.start_urls = ['https://news.sina.com.cn/china/','https://ent.sina.com.cn/zongyi/','https://ent.sina.com.cn/film/']
        self.option = webdriver.ChromeOptions()
        self.option.add_argument('headless')
        self.option.add_argument('no=sandbox')
        self.option.add_argument('--blink-setting=imagesEnable=false')

    def start_requests(self):
        for url in self.start_urls:
            yield Request(url=url, callback=self.parse)

    def parse(self, response):
        driver = webdriver.Chrome(chrome_options=self.option)
        driver.set_page_load_timeout(30)
        driver.get(response.url)
        for i in range(self.pages):
            while not driver.find_element_by_xpath("//div[@class='feed-card-page']").text:
                driver.execute_script("window.scrollTo(0, document.body.scrollHeight);")
            title = driver.find_elements_by_xpath("//h2[@class='undefined']/a[@target='_blank']")
            time = driver.find_elements_by_xpath("//h2[@class='undefined']/../div[@class='feed-card-a feed-card-clearfix']/div[@class='feed-card-time']")

            for i in range(len(title)):
                item = DataItem()
                if response.url == "https://news.sina.com.cn/china/":
                    item['type'] = 'news'

                if response.url == "https://ent.sina.com.cn/zongyi/":
                    item['type'] = 'zongyi'
                
                if response.url == "https://ent.sina.com.cn/film/":
                    item['type'] = 'film'


                eachtitle = title[i].text
                item['title'] = eachtitle
                item['desc'] = ''
                eachtime = time[i].text
                #print(eachtitle)
                href = title[i].get_attribute('href')
                #对网页时间进行处理
                today = datetime.datetime.now()
                eachtime = eachtime.replace('今天', str(today.month) + '月' + str(today.day) + '日')
                #print(eachtime)
                if '分钟前' in eachtime:
                    minute = int(eachtime.split('分钟前')[0])
                    t = datetime.datetime.now() - datetime.timedelta(minutes=minute)
                    t2 = datetime.datetime(year=t.year, month=t.month, day=t.day, hour=t.hour, minute=t.minute)
                    print(str(t2))
                else:
                    if '年' not in eachtime:
                        eachtime = str(today.year) + '年' + eachtime
                    t1 = re.split('[年月日:]', eachtime)
                    t2 = datetime.datetime(year=int(t1[0]), month=int(t1[1]), day=int(t1[2]), hour=int(t1[3]), minute=int(t1[4]))

                item['times'] = t2

                yield Request(url=response.urljoin(href), meta={'name': item}, callback=self.parse_namedetail)

            try:
                driver.find_element_by_xpath("//div[@class='feed-card-page']/span[@class='pagebox_next']/a").click()
            except:
                break

    def parse_namedetail(self, response):
        selector = Selector(response)
        desc = selector.xpath("//div[@class='article']/p/text()").extract()
        item = response.meta['name']
        desc = list(map(str.strip, desc))
        item['desc'] = ''.join(desc)
        print(item)
        yield item

3. 保存到MySQL

pipelines

from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy import Column, create_engine, Text, DateTime, String, Integer
from sqlalchemy.orm import sessionmaker
import pymysql

Base = declarative_base()

class Data(Base):
    __tablename__ = 'data'
    id = Column(Integer(), primary_key=True)
    times = Column(DateTime)
    title = Column(Text())
    content = Column(Text())
    type = Column(Text())

class SinaPipeline:
    def __init__(self):
        self.engine = create_engine('mysql+pymysql://root:123456@localhost:3306/sina2', encoding='utf-8')
        Base.metadata.create_all(self.engine)
        self.DBSession = sessionmaker(bind=self.engine)

    def process_item(self, item, spider):
        new = Data()
        new.title = item['title']
        new.times = item['times']
        new.content = item['desc']
        new.type = item['type']
        session = self.DBSession()
        session.add(new)
        session.commit()
        return item