A brief introduction to the crawler framework (talonspider) in Python

This article introduces to you a simple introduction and usage method of the crawler developed using pythonframeworktalonspider. Friends who need it can refer to it

1.Why write this?

There is no need to use a relatively large framework to crawl some simple pages. It would be troublesome to write them by hand.

So I wrote talonspider for this requirement:

•1. For single page item extraction - click here for a detailed introduction
•2.Spider module-click here for a detailed introduction

2.Introduction &&Usage

2.1.item

This module can be used independently. For some websites with relatively simple requests (such as only get requests), you can quickly write what you want by using this module alone. Crawler, for example (the following uses python3, see the examples directory for python2):

2.1.1. Single page single target

For example, to obtain this URL http://book.qidian.com/ The book information, cover and other information of info/1004608738 can be written directly like this:

import time
from talonspider import Item, TextField, AttrField
from pprint import pprint

class TestSpider(Item):
  title = TextField(css_select='.book-info>h1>em')
  author = TextField(css_select='a.writer')
  cover = AttrField(css_select='a#bookImg>img', attr='src')

  def tal_title(self, title):
    return title

  def tal_cover(self, cover):
    return 'http:' + cover

if name == 'main':
  item_data = TestSpider.get_item(url='http://book.qidian.com/info/1004608738')
  pprint(item_data)

For details, see qidian_details_by_item.py

2.1.1. Multiple targets on a single page

For example, to obtain the 25 movies displayed on the Douban 250 movie homepage, this page has 25 targets. You can write directly like this:

from talonspider import Item, TextField, AttrField
from pprint import pprint

# 定义继承自item的爬虫类
class DoubanSpider(Item):
  target_item = TextField(css_select='p.item')
  title = TextField(css_select='span.title')
  cover = AttrField(css_select='p.pic>a>img', attr='src')
  abstract = TextField(css_select='span.inq')

  def tal_title(self, title):
    if isinstance(title, str):
      return title
    else:
      return ''.join([i.text.strip().replace('\xa0', '') for i in title])

if name == 'main':
  items_data = DoubanSpider.get_items(url='movie.douban.com/top250')
  result = []
  for item in items_data:
    result.append({
      'title': item.title,
      'cover': item.cover,
      'abstract': item.abstract,
    })
  pprint(result)

For details, see douban_page_by_item.py

2.2.spider

When you need to crawl hierarchical pages, such as crawling all Douban 250 movies, then the spider part comes in handy Usage:

# !/usr/bin/env python
from talonspider import Spider, Item, TextField, AttrField, Request
from talonspider.utils import get_random_user_agent


# 定义继承自item的爬虫类
class DoubanItem(Item):
  target_item = TextField(css_select='p.item')
  title = TextField(css_select='span.title')
  cover = AttrField(css_select='p.pic>a>img', attr='src')
  abstract = TextField(css_select='span.inq')

  def tal_title(self, title):
    if isinstance(title, str):
      return title
    else:
      return ''.join([i.text.strip().replace('\xa0', '') for i in title])


class DoubanSpider(Spider):
  # 定义起始url，必须
  start_urls = ['https://movie.douban.com/top250']
  # requests配置
  request_config = {
    'RETRIES': 3,
    'DELAY': 0,
    'TIMEOUT': 20
  }
  # 解析函数 必须有
  def parse(self, html):
    # 将html转化为etree
    etree = self.e_html(html)
    # 提取目标值生成新的url
    pages = [i.get('href') for i in etree.cssselect('.paginator>a')]
    pages.insert(0, '?start=0&filter=')
    headers = {
      "User-Agent": get_random_user_agent()
    }
    for page in pages:
      url = self.start_urls[0] + page
      yield Request(url, request_config=self.request_config, headers=headers, callback=self.parse_item)

  def parse_item(self, html):
    items_data = DoubanItem.get_items(html=html)
    # result = []
    for item in items_data:
      # result.append({
      #   'title': item.title,
      #   'cover': item.cover,
      #   'abstract': item.abstract,
      # })
      # 保存
      with open('douban250.txt', 'a+') as f:
        f.writelines(item.title + '\n')


if name == 'main':
  DoubanSpider.start()

Console:

/Users/howie/anaconda3/envs/work3/bin/python /Users/howie/Documents/programming/python/git/talonspider/examples/douban_page_by_spider.py
2017-06-07 23:17:30,346 - talonspider - INFO: talonspider started
2017-06-07 23:17:30,693 - talonspider_requests - INFO: GET a url: https://movie.douban.com/top250
2017-06-07 23:17:31,074 - talonspider_requests - INFO: GET a url: https://movie.douban.com/top250?start=25&filter=
2017-06-07 23:17:31,416 - talonspider_requests - INFO: GET a url: https://movie.douban.com/top250?start=50&filter=
2017-06-07 23:17:31,853 - talonspider_requests - INFO: GET a url: https://movie.douban.com/top250?start=75&filter=
2017-06-07 23:17:32,523 - talonspider_requests - INFO: GET a url: https://movie.douban.com/top250?start=100&filter=
2017-06-07 23:17:33,032 - talonspider_requests - INFO: GET a url: https://movie.douban.com/top250?start=125&filter=
2017-06-07 23:17:33,537 - talonspider_requests - INFO: GET a url: https://movie.douban.com/top250?start=150&filter=
2017-06-07 23:17:33,990 - talonspider_requests - INFO: GET a url: https://movie.douban.com/top250?start=175&filter=
2017-06-07 23:17:34,406 - talonspider_requests - INFO: GET a url: https://movie.douban.com/top250?start=200&filter=
2017-06-07 23:17:34,787 - talonspider_requests - INFO: GET a url: https://movie.douban.com/top250?start=225&filter=
2017-06-07 23:17:34,809 - talonspider - INFO: Time usage：0:00:04.462108

Process finished with exit code 0

At this time, douban250 will be generated in the current directory .txt, see douban_page_by_spider.py for details.

3. Description

This is a work of study. There are still many areas to be improved. Comments are welcome. The project address is talonspider.

The above is the detailed content of A brief introduction to the crawler framework (talonspider) in Python. For more information, please follow other related articles on the PHP Chinese website!