latest_blame.py

from bs4 import BeautifulSoup
from datetime import datetime

def get_most_recent_git_change_for_file(url):
    """
    Parameters:
    ----------
            
        url: str, the url to where the github file is stored
    
    Example:
    --------
    
        get_most_recent_git_change_for_file('https://github.com/ghandic/confluenceapi/blob/master/README.md')
    """
    
    assert isinstance(url, str), 'url should be the path to where a github file is stored'
    
    url_list = url.split('/')
    url_list[5] = 'blame'
    new_url = '/'.join(url_list)
    
    r = requests.get(url=new_url)
    
    soup = BeautifulSoup(r.text)
    
    all_blame_times = [datetime.strptime(t.get('datetime'), "%Y-%m-%dT%H:%M:%SZ") for t in soup.find_all('time-ago')]
    
    return max(all_blame_times).strftime('%d-%m-%Y %H:%M:%S')

# Or get by using the API


import requests
import json
import dateutil.parser

def get_last_update_time(url):
    
    info = {
        'repo':url.split('/')[4],
        'owner':url.split('/')[3],
        'path': '/'.join(url.split('/')[7:] if isinstance(url.split('/')[7:], list) else url.split('/')[7]),
        'branch':url.split('/')[6]
    }
    
    r = requests.get('https://api.github.com/repos/{owner}/{repo}/commits/{branch}?path={path}&page=1'.format(**info))
    
    if r.status_code == 200:
        last_updated = dateutil.parser.parse(json.loads(r.text)['commit']['committer']['date'])
        return last_updated.strftime('%d/%m/%Y %H:%M:%S')
    else:
        return 'Failed to load'
    
get_last_update_time('https://github.com/plotly/dash-recipes/blob/master/downloads/NYC-download.xlsx')