使用Scrapy对新闻进行爬虫（一）

Scrapy Item Pipeline学习笔记

Item Pipeline 主要用于从网页抓取（spider）后对数据Item进行收集，写入数据库或文件中。

执行方式

spider 在获得item后，会传递给item pipeline，进行后续数据收集工作。
在setting中对item pipeline类路径进行配置，scrapy框架会调用该item pipeline类，为了正确调用，
item pipeline类必须按照框架要求实现一些方法。使用者只需关注实现这些方法即可。

例子

下面文件实现了一个简单的item pipeline类，对抓取的新闻数据进行进一步处理并写入文件中。这些方法的功能见注释。
1. 文件：pipelines.py

注意事项：
1. 初始化函数实现非常自由，无需限定参数，只需保证from_crawler类方法能够调用该初始化函数生成相应的实例及可。
2. 框架所使用方法声明参数固定。（保证框架能够正确调用）

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

# Define your item pipelines here
#
# Don't forget to add your pipeline to the ITEM_PIPELINES setting
# See: http://doc.scrapy.org/en/latest/topics/item-pipeline.html


class News2FileFor163Pipeline(object):
    """
    pipeline: process items given by spider
    """

    def __init__(self, filepath, filename):
        """
        init for the pipeline class
        """
        self.fullname = filepath + '/' + filename
        self.id = 0
        return

    def process_item(self, item, spider):
        """
        process each items from the spider.
        example: check if item is ok or raise DropItem exception.
        example: do some process before writing into database.
        example: check if item is exist and drop.
        """
        for element in ("url","source","title","editor","time","content"):
            if item[element] is None:
                raise DropItem("invalid items url: %s" % str(item["url"]))
        self.fs.write("news id: %s" % self.id)
        self.fs.write("\n")
        self.id += 1
        self.fs.write("url: %s" % item["url"][0].strip().encode('UTF-8'))
        self.fs.write("\n")
        self.fs.write("source: %s" % item["source"][0].strip().encode('UTF-8'))
        self.fs.write("\n")
        self.fs.write("title: %s" % item["title"][0].strip().encode('UTF-8'))
        self.fs.write("\n")
        self.fs.write("editor: %s" % item["editor"][0].strip().
                      encode('UTF-8').split('：')[1])
        self.fs.write("\n")
        time_string = item["time"][0].strip().split()
        datetime = time_string[0] + ' ' + time_string[1]
        self.fs.write("time: %s" % datetime.encode('UTF-8'))
        self.fs.write("\n")
        content = ""
        for para in item["content"]:
            content += para.strip().replace('\n', '').replace('\t', '')
        self.fs.write("content: %s" % content.encode('UTF-8'))
        self.fs.write("\n")
        return item

    def open_spider(self, spider):
        """
        called when spider is opened.
        do something before pipeline is processing items.
        example: do settings or create connection to the database.
        """
        self.fs = open(self.fullname, 'w+')
        return

    def close_spider(self, spider):
        """
        called when spider is closed.
        do something after pipeline processing all items.
        example: close the database.
        """
        self.fs.flush()
        self.fs.close()
        return

    @classmethod
    def from_crawler(cls, crawler):
        """
        return an pipeline instance.
        example: initialize pipeline object by crawler's setting and components.
        """
        return cls(crawler.settings.get('ITEM_FILE_PATH'),
                   crawler.settings.get('ITEM_FILE_NAME'))

1. 抓取project中settting.py的相关配置代码

# Configure item pipelines
# See http://scrapy.readthedocs.org/en/latest/topics/item-pipeline.html
ITEM_PIPELINES = {
    'NewsSpiderMan.pipelines.News2FileFor163Pipeline': 300,
}

进一步需求

如果抓取数据的内容非常多，使用item pipeline 对数据处理并写入数据库中乃王道。