Python爬虫：Scrapy中间件Middleware和Pipeline

Scrapy提供了可自定义2种中间件，1个数据处理器

名称	作用	用户设置
数据收集器（Item-Pipeline)	处理item	覆盖
下载中间件（Downloader-Middleware）	处理request/response	合并
爬虫中间件（Spider-Middleware）	处理item/response/request	合并

解释：

用户设置：是指​​custom_settings​​

可是它们继承的父类竟然是Object…，每次都要查文档。

正常来说应该提供一个抽象函数作为接口，给使用者实现自己的具体功能，不知道为啥这么设计

通过几段代码及注释，简要说明三个中间件的功能

1、Spider

​​baidu_spider.py​​

from scrapy import Spider, cmdline

class BaiduSpider(Spider):
    name = "baidu_spider"

    start_urls = [
        "https://www.baidu.com/"
    ]

    custom_settings = {
        "SPIDER_DATA": "this is spider data",
        "DOWNLOADER_MIDDLEWARES": {
                "scrapys.mymiddleware.MyMiddleware": 100,
            },
        "ITEM_PIPELINES": {
            "scrapys.mypipeline.MyPipeline": 100,
        },
        "SPIDER_MIDDLEWARES":{
            "scrapys.myspidermiddleware.MySpiderMiddleware": 100,
        }
    }

    def parse(self, response):
        pass


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

2、Pipeline

​​mypipeline.py​​

class MyPipeline(object):
    def __init__(self, spider_data):
        self.spider_data = spider_data

    @classmethod
    def from_crawler(cls, crawler):
        """
        获取spider的settings参数,返回Pipeline实例对象
        """
        spider_data = crawler.settings.get("SPIDER_DATA")
        print("### pipeline get spider_data: {}".format(spider_data))

        return cls(spider_data)

    def process_item(self, item, spider):
        """
        return Item 继续处理
        raise DropItem 丢弃
        """
        print("### call process_item")

        return item

    def open_spider(self, spider):
        """
        spider开启时调用
        """
        print("### spdier open {}".format(spider.name))


    def close_spider(self, spider):
        """
        spider关闭时调用
        """
        print("### spdier close {}".format(spider.name))

3、Downloader-Middleware

​​mymiddleware.py​​

class MyMiddleware(object):
    def __init__(self, spider_data):
        self.spider_data = spider_data

    @classmethod
    def from_crawler(cls, crawler):
        """
        获取spider的settings参数,返回中间件实例对象
        """
        spider_data = crawler.settings.get("SPIDER_DATA")
        print("### middleware get spider_data: {}".format(spider_data))

        return cls(spider_data)

    def process_request(self, request, spider):
        """
        return
            None: 继续处理Request
            Response: 返回Response
            Request: 重新调度
        raise IgnoreRequest:  process_exception -> Request.errback
        """
        print("### call process_request")

    def process_response(self, request, response, spider):
        """
        return
            Response: 继续处理Response
            Request: 重新调度
        raise IgnoreRequest: Request.errback
        """
        print("### call process_response")
        return response

    def process_exception(self, request, exception, spider):
        """
        return
            None: 继续处理异常
            Response: 返回Response
            Request: 重新调用
        """
        pass

4、Spider-Middleware

​​myspidermiddleware.py​​

class MySpiderMiddleware(object):
    def __init__(self, spider_data):
        self.spider_data = spider_data

    @classmethod
    def from_crawler(cls, crawler):
        """
        获取spider的settings参数,返回中间件实例对象
        """
        spider_data = crawler.settings.get("SPIDER_DATA")
        print("### spider middleware get spider_data: {}".format(spider_data))

        return cls(spider_data)

    def process_spider_input(self, response, spider):
        """
        response通过时调用
        return None  继续处理response
        raise Exception
        """

        print("### call process_spider_input")

    def process_spider_output(self, response, result, spider):
        """
        response返回result时调用
        return
            iterable of Request、dict or Item
        """
        print("### call process_spider_output")

        for i in result:
            yield i

    def process_spider_exception(self, response, exception, spider):
        """
        return
            None
            iterable of Response, dict, or Item
        """
        pass

运行爬虫后，查看日志

### middleware get spider_data: this is spider data
### spider middleware get spider_data: this is spider data
### pipeline get spider_data: this is spider data

### spdier open baidu_spider
### call process_request
### call process_response
### call process_spider_input
### call process_spider_output
### spdier close baidu_spider

根据日志输出信息，看到大致流程是和Scrapy数据流向图保持一致的

中间件启动顺序

download middleware spider middleware pipeline

处理函数调用顺序

spdier open process_request process_response process_spider_input process_spider_output spdier close

参考

1. 1. 
      
2. ​​Item Pipeline​​
3. ​​下载器中间件(Downloader Middleware)​​
4. ​​Spider中间件(Middleware)​​
5. ​​Scrapy 1.5 documentation​​
   1.