update

#!/usr/bin/env python3
import argparse
import numpy as np
import os
import pandas as pd
import requests
import sys


GITHUB_TOKEN = os.getenv('GITHUB_TOKEN')
HEADERS = {'Authorization': 'Bearer %s' % (GITHUB_TOKEN)}


def gh_query(query: str):
    request = requests.post('https://api.github.com/graphql', json={'query': query}, headers=HEADERS)
    if request.status_code != 200:
        raise Exception(f'Query failed with status code {request.status_code}: {query}')
    return request.json()


def parse_github_url(url: str):
    if not url.startswith('https://github.com'):
        return None
    
    tokens = url.split('/')
    if len(tokens) != 5:
        return None
    
    return tokens[-2:]


def update(x: pd.Series) -> pd.Series:
    # check whether url is a github url
    gh_repo = parse_github_url(x.url)

    # update github stars
    if gh_repo is not None:
        owner, name = gh_repo
        query = '''
            query {
                repository(owner: "%s", name: "%s") {
                    stargazerCount
                }
            }
            ''' % (owner, name)
        result = gh_query(query)
        x.github_stars = result['data']['repository']['stargazerCount']
    else:
        x.github_stars = np.nan

    return x


def update_safe(x: pd.Series) -> pd.Series:
    try:
        sys.stderr.write('Updating \'%s\'...\n' % (x['name']))
        return update(x)
    except Exception as e:
        sys.stderr.write('Failed to update \'%s\': %s\n' % (x['name'], e))
        return x


def main():
    # parse command-line args
    parser = argparse.ArgumentParser()
    parser.add_argument('input', help='Input dataset (CSV)')

    args = parser.parse_args()

    # load input file
    df = pd.read_csv(args.input)

    # fetch github metadata
    df = df.apply(lambda x: update_safe(x), axis='columns')

    # save output file
    df.to_csv(args.input, index=False, float_format='%.0f')


if __name__ == '__main__':
    main()