前言

上一篇中我們在維基百科的內(nèi)部網(wǎng)站上隨機(jī)跳轉(zhuǎn)進(jìn)入文章類網(wǎng)頁，而忽視外部網(wǎng)站鏈接。本篇文章將處理網(wǎng)站的外部鏈接并試圖收集一些網(wǎng)站數(shù)據(jù)。和單個域名網(wǎng)站爬取不同，不同域名的網(wǎng)站結(jié)構(gòu)千差萬別，這就意味我們的代碼需要更加的靈活以適應(yīng)不同的網(wǎng)站結(jié)構(gòu)。
因此，我們將代碼寫成一組函數(shù)，這些函數(shù)組合起來就可以應(yīng)用在不同類型的網(wǎng)絡(luò)爬蟲需求。

隨機(jī)跳轉(zhuǎn)外部鏈接

利用函數(shù)組，我們可以在50行左右滿足爬取外部網(wǎng)站的需求。
示例代碼：

from urllib.request import urlopen
from bs4 import BeautifulSoup
import re
import datetime
import random
from urllib.parse import quote

pages = set()
random.seed(datetime.datetime.now())


''' 獲取一個網(wǎng)頁的所有互聯(lián)網(wǎng)鏈接'''

# 獲取網(wǎng)頁所有內(nèi)部鏈接
def get_internal_links(soup, include_url):
    internal_links = []
    # find all links that befin with a '/'
    print(include_url)
    for link in soup.find_all('a',
                              href=re.compile(r'^((/|.)*' + include_url + ')')):
        if link.attrs['href'] is not None:
            if link.attrs['href'] not in internal_links:
                internal_links.append(link.attrs['href'])
    return internal_links


# retrieves a list of all external links found on a page
#獲取網(wǎng)頁上所有外部鏈接
def get_external_links(soup, exclude_url):
    external_links = []
    # Finds all links that starts with 'http' or 'www' that do not contain the
    # current URL
    for link in soup.find_all('a',
                              href=re.compile(r'^(http|www)((?!' + exclude_url + ').)*$')):
        if link.attrs['href'] is not None:
            if link.attrs['href'] not in external_links:
                external_links.append(link.attrs['href'])
    return external_links


#拆分網(wǎng)址獲取主域名
def split_address(address):
    address_parts = address.replace('http://', '').split('/')
    return address_parts


#隨機(jī)外部鏈接跳轉(zhuǎn)
def get_random_external_link(starting_page):
    html = urlopen(starting_page)

    soup = BeautifulSoup(html, 'lxml')
    external_links = get_external_links(
        soup, split_address(starting_page)[0])  # find the domain URL
    if len(external_links) == 0:
        internal_links = get_internal_links(soup, starting_page)
        print(len(internal_links))
        return get_external_links(soup,
                                  internal_links[random.randint(0, len(internal_links) - 1)])
    else:
        return external_links[random.randint(0, len(external_links) - 1)]



hop_count = set()

#只跳轉(zhuǎn)外部鏈接，設(shè)置跳轉(zhuǎn)次數(shù)loop, 默認(rèn)跳轉(zhuǎn)5次
def follow_external_only(starting_site, loop=5):
    global hop_count
    external_link = get_random_external_link(
        quote(starting_site, safe='/:?='))
    print('Random external link is: ' + external_link)
    while len(hop_count) < loop:
        hop_count.add(external_link)
        print(len(hop_count))
        follow_external_only(external_link)



follow_external_only("http://www.baidu.com")

由于代碼沒有異常處理和反反爬蟲處理，因此一定會報錯。由于跳轉(zhuǎn)是隨機(jī)的，可以多運(yùn)行幾次，有興趣的可以根據(jù)每次的報錯原因完善代碼。
輸出結(jié)果：

Random external link is: http://v.baidu.com/v?ct=301989888&rn=20&pn=0&db=0&s=25&ie=utf-8&word=
1
Random external link is: http://baishi.baidu.com/watch/6388818335201070269.html
2
Random external link is: http://v.baidu.com/tv/
3
Random external link is: http://player.baidu.com/yingyin.html
4
Random external link is: http://help.baidu.com/question?prod_en=player
5
Random external link is: http://home.baidu.com
[Finished in 6.3s]

抓取網(wǎng)頁上所有外部鏈接

把代碼寫成函數(shù)的好處是可以簡單地修改或者添加以滿足不同的需求而不會破壞代碼。比如：
目的：爬取整個網(wǎng)頁所有外部鏈接并對每個鏈接標(biāo)記
我們可以添加如下函數(shù)：

# Collects a list of all external URLs found on the site
all_ext_links = set()
all_int_links = set()


def get_all_external_links(site_url):
    html = urlopen(site_url)
    soup = BeautifulSoup(html, 'lxml')
    print(split_address(site_url)[0])
    int
    internal_links = get_internal_links(soup, split_address(site_url)[0])
    external_links = get_external_links(soup, split_address(site_url)[0])
    for link in external_links:
        if link not in all_ext_links:
            all_ext_links.add(link)
            print(link)
    for link in internal_links:
        if link not in all_int_links:
            print('About to get link: ' + link)
            all_int_links.add(link)
            get_all_external_links(link)

# follow_external_only("http://www.baidu.com")
get_all_external_links('http://oreilly.com')

輸出結(jié)果如下：

oreilly.com
oreilly.com
https://cdn.oreillystatic.com/pdf/oreilly_high_performance_organizations_whitepaper.pdf
http://twitter.com/oreillymedia
http://fb.co/OReilly
https://www.linkedin.com/company/oreilly-media
https://www.youtube.com/user/OreillyMedia
About to get link: https://www.oreilly.com
https:
https:
https://www.oreilly.com
http://www.oreilly.com/ideas
https://www.safaribooksonline.com/?utm_medium=content&utm_source=oreilly.com&utm_campaign=lgen&utm_content=20170601+nav
http://www.oreilly.com/conferences/
http://shop.oreilly.com/
http://members.oreilly.com
https://www.oreilly.com/topics
https://www.safaribooksonline.com/?utm_medium=content&utm_source=oreilly.com&utm_campaign=lgen&utm_content=20170505+homepage+get+started+now
https://www.safaribooksonline.com/accounts/login/?utm_medium=content&utm_source=oreilly.com&utm_campaign=lgen&utm_content=20170203+homepage+sign+in
https://www.safaribooksonline.com/live-training/?utm_medium=content&utm_source=oreilly.com&utm_campaign=lgen&utm_content=20170201+homepage+take+a+live+online+course
https://www.safaribooksonline.com/learning-paths/?utm_medium=content&utm_source=oreilly.com&utm_campaign=lgen&utm_content=20170201+homepage+follow+a+path
https://www.safaribooksonline.com/?utm_medium=content&utm_source=oreilly.com&utm_campaign=lgen&utm_content=20170505+homepage+unlimited+access
http://www.oreilly.com/live-training/?view=grid
https://www.safaribooksonline.com/your-experience/?utm_medium=content&utm_source=oreilly.com&utm_campaign=lgen&utm_content=20170201+homepage+safari+platform
https://www.oreilly.com/ideas/8-data-trends-on-our-radar-for-2017?utm_medium=referral&utm_source=oreilly.com&utm_campaign=lgen&utm_content=link+2017+trends
https://www.oreilly.com/ideas?utm_medium=referral&utm_source=oreilly.com&utm_campaign=lgen&utm_content=link+read+latest+articles
http://www.oreilly.com/about/
http://www.oreilly.com/work-with-us.html
http://www.oreilly.com/careers/
http://shop.oreilly.com/category/customer-service.do
http://www.oreilly.com/about/contact.html
http://www.oreilly.com/emails/newsletters/
http://www.oreilly.com/terms/
http://www.oreilly.com/privacy.html
http://www.oreilly.com/about/editorial_independence.html
About to get link: https://www.safaribooksonline.com/?utm_medium=content&utm_source=oreilly.com&utm_campaign=lgen&utm_content=20170601+nav
https:
https:
https://www.oreilly.com/
About to get link: https://www.oreilly.com/
https:
https:
About to get link: https://www.oreilly.com/topics
......

程序會一直循環(huán)下去直到達(dá)到python默認(rèn)的循環(huán)極限， 有興趣的朋友可以像上面的代碼一樣添加默認(rèn)循環(huán)限制loop=5。