获取某个github组织的所有库

github_spider

# ! python3
"""
获取某个用户的所有库
安装依赖库：
  pip install requests,bs4
运行：
  python3 github_spider.py nasa
  
  会打印出每个项目的git地址
  比如下载该组织的的所有库
   python3 github_spider.py nasa | xargs -n1 git clone
TODO: 扩展到用户 
  
Author: imakar@qq.com
"""

import sys
import requests
from bs4 import BeautifulSoup

def get_org_pages_count(org):
    homepage_url="https://github.com/{name}".format(name=org)
    homepage=requests.get(homepage_url)
    soup0=BeautifulSoup(homepage.content,"html.parser")
    page_count=int(soup0.select_one("#org-repositories > div > div > div > div > em.current")['data-total-pages'])
    return page_count

def get_org_repos_in_page(org,page):
    url="https://github.com/{name}?page={page}".format(name=org,page=page)
    soup0=BeautifulSoup(requests.get(url).content,"html.parser")
    alist=soup0.select("#org-repositories > div > div.org-repos.repo-list > li > div.d-inline-block.mb-1 > h3 > a")
    return [each.text.strip() for each in alist]

def get_user_type(user):
    return "org"

def print_repos(user):
    if get_user_type(sys.argv[1])=="org":    
        for i in range(1,get_org_pages_count(user)):
            for each in get_org_repos_in_page(user,i):
                print("https://github.com/{user}/{repo}.git".format(user=user,repo=each))


if __name__=="__main__":
    if len(sys.argv)!=2:
        raise ValueError("arguments count must be 2")
    user=sys.argv[1]
    print_repos(user)