Python爬虫：Scrapy从脚本运行爬虫的5种方式

测试环境

# 环境一
Python 3.6.5
Scrapy==1.5.0

# 环境二
Python 2.7.5
Scrapy==1.1.2

一、命令行运行爬虫

1、编写爬虫文件 ​​baidu.py​​

# -*- coding: utf-8 -*-

from scrapy import Spider

class BaiduSpider(Spider):
    name = 'baidu'

    start_urls = ['http://baidu.com/']

    def parse(self, response):
        self.log("run baidu")

2、运行爬虫（2种方式）

# 运行爬虫
$ scrapy crawl baidu 

# 在没有创建项目的情况下运行爬虫
$ scrapy runspider baidu.py

二、文件中运行爬虫

1、​​cmdline​​方式运行爬虫

# -*- coding: utf-8 -*-

from scrapy import cmdline, Spider


class BaiduSpider(Spider):
    name = 'baidu'

    start_urls = ['http://baidu.com/']

    def parse(self, response):
        self.log("run baidu")


if __name__ == '__main__':
    cmdline.execute("scrapy crawl baidu".split())

2、​​CrawlerProcess​​方式运行爬虫

# -*- coding: utf-8 -*-

from scrapy import Spider
from scrapy.crawler import CrawlerProcess
from scrapy.utils.project import get_project_settings

class BaiduSpider(Spider):
    name = 'baidu'

    start_urls = ['http://baidu.com/']

    def parse(self, response):
        self.log("run baidu")


if __name__ == '__main__':
    # 通过方法 get_project_settings() 获取配置信息
    process = CrawlerProcess(get_project_settings())
    process.crawl(BaiduSpider)
    process.start()

3、通过​​CrawlerRunner​​ 运行爬虫

# -*- coding: utf-8 -*-

from scrapy import Spider
from scrapy.crawler import CrawlerRunner
from scrapy.utils.log import configure_logging
from twisted.internet import reactor


class BaiduSpider(Spider):
    name = 'baidu'

    start_urls = ['http://baidu.com/']

    def parse(self, response):
        self.log("run baidu")


if __name__ == '__main__':
    # 直接运行控制台没有日志
    configure_logging(
        {
            'LOG_FORMAT': '%(message)s'
        }
    )

    runner = CrawlerRunner()

    d = runner.crawl(BaiduSpider)
    d.addBoth(lambda _: reactor.stop())
    reactor.run()

三、文件中运行多个爬虫

项目中新建一个爬虫 ​​SinaSpider​​

# -*- coding: utf-8 -*-

from scrapy import Spider


class SinaSpider(Spider):
    name = 'sina'

    start_urls = ['https://www.sina.com.cn/']

    def parse(self, response):
        self.log("run sina")

1、​​cmdline​​方式不可以运行多个爬虫

如果将两个语句放在一起，第一个语句执行完后程序就退出了，执行到不到第二句

# -*- coding: utf-8 -*-

from scrapy import cmdline

cmdline.execute("scrapy crawl baidu".split())
cmdline.execute("scrapy crawl sina".split())

记得之前我还写过一个使用 ​​cmdline​​运行多个爬虫的脚本

文章：​​Python爬虫：scrapy定时运行的脚本​​

不过有了以下两个方法来替代，就更优雅了

2、​​CrawlerProcess​​方式运行多个爬虫

备注：爬虫项目文件为：

scrapy_demo/spiders/baidu.py

scrapy_demo/spiders/sina.py

# -*- coding: utf-8 -*-

from scrapy.crawler import CrawlerProcess

from scrapy_demo.spiders.baidu import BaiduSpider
from scrapy_demo.spiders.sina import SinaSpider

process = CrawlerProcess()
process.crawl(BaiduSpider)
process.crawl(SinaSpider)
process.start()

此方式运行，发现日志中中间件只启动了一次，而且发送请求基本是同时的，说明这两个爬虫运行不是独立的，可能会相互干扰

3、通过​​CrawlerRunner​​ 运行多个爬虫

# -*- coding: utf-8 -*-

from scrapy.crawler import CrawlerRunner
from scrapy.utils.log import configure_logging
from twisted.internet import reactor

from scrapy_demo.spiders.baidu import BaiduSpider
from scrapy_demo.spiders.sina import SinaSpider


configure_logging()
runner = CrawlerRunner()
runner.crawl(BaiduSpider)
runner.crawl(SinaSpider)
d = runner.join()
d.addBoth(lambda _: reactor.stop())

reactor.run()

此方式也只加载一次中间件，不过是逐个运行的，会减少干扰，官方文档也推荐使用此方法来运行多个爬虫

总结

​​cmdline.execute​​ 运行单个爬虫文件的配置最简单，一次配置，多次运行

参考：
​​Scrapy Common Practices​​