Fetch GitHub org members with role and verified domain email

fetch-verified-emails.py

import csv
import os
import requests
import sys

# Verify the org name variable has been passed
sys_args = sys.argv
if len(sys_args)==1:
    print("Please provide the GitHub org name")
    print("{} <github-org>".format(sys_args[0]))
    exit()
else:
    github_org = "slackhq"

# Your github admin token, store this in an environment variable
# or pass it in when you invoke the script
# GITHUB_TOKEN="<your OWNER token>" python3 fetch-verified-emails.py

github_api_token = os.getenv('GITHUB_TOKEN', '')

headers = {"Authorization": "token {}".format(github_api_token)}

# Create the members_list array and add the column headers
member_list = [["name", "username", "role", "Verified emails"]]


def run_query(query):
    # Query the GraphQL API
    request = requests.post('https://api.github.com/graphql', json={'query': query}, headers=headers)
    if request.status_code == 200:
        return request.json()
    else:
        raise Exception("Query failed to run by returning code of {}. {}".format(request.status_code, query))


def fetch_user_data(org_name, cursor="null"):
    # Get a paginated list of GitHub org members, including membership role and verified domain emails

    # The GraphQL query, defined as a multi-line string.
    query_template = """
    {
      organization(login: "%s") {
        membersWithRole(first: 100, after: %s) {
          edges {
            cursor
            node {
              login
              name
              organizationVerifiedDomainEmails(login:"%s")
            }
            role
          }
        }
      }
      rateLimit {
        limit
        cost
        remaining
        resetAt
      }
    }
    """

    # Add org name and cursor vars to the query template and run the query
    query = query_template%(org_name, cursor, org_name)
    result = run_query(query)

    # If the results contain errors, log out the results
    if "errors" in result:
        print("⚠️ QUERY RETURNED WITH ERRORS")
        print(result)

    # Typical success and error response:
    # For errors, membersWithRole will be None
    # For success, 'errors' won't be present
    # {
    #     'data': {
    #         'organization': {
    #             'membersWithRole': {
    #                 [{
    #                     'cursor': 'Y3Vyc29yOnYyOpHNESc=',
    #                     'node': {
    #                         'login': 'roach',
    #                         'name': 'Jason Roche',
    #                         'organizationVerifiedDomainEmails': [
    #                             '[email protected]',
    #                             '[email protected]'
    #                         ]
    #                     },
    #                     'role': 'OWNER'}]
    #             }
    #         },
    #         'rateLimit': {
    #             'limit': 5000, 'cost': 1, 'remaining': 4992, 'resetAt': '2022-01-26T00:14:27Z'
    #         }
    #     },
    #     'errors': [
    #         {
    #             'type': 'INVALID_CURSOR_ARGUMENTS',
    #             'path': ['organization', 'membersWithRole', 'edges'],
    #             'locations': [{'line': 5, 'column': 11}],
    #             'message': '`null` does not appear to be a valid cursor.'
    #         }
    #     ]
    # }

    # Print the remaining rate limit count and reset timestamp
    # rate_limit = result["data"]["rateLimit"]
    # print("Remaining rate limit: {}, resets at: {}".format(rate_limit["remaining"],rate_limit["resetAt"]))

    # Iterate through the org members
    org_members = result["data"]["organization"]["membersWithRole"]["edges"]
    for member_node in org_members:
        member_info = member_node["node"]

        # Flatten the organizationVerifiedDomainEmails list
        member_emails = ""
        if member_info["organizationVerifiedDomainEmails"]:
            # print("Verified domain emails: {}".format(" ".join(member_info["organizationVerifiedDomainEmails"])))
            member_emails = ", ".join(member_info["organizationVerifiedDomainEmails"])

        member_list.append([member_info["login"], member_info["name"], member_node["role"], member_emails])

    if org_members and "cursor" in org_members[-1]:
        current_cursor = '"{}"'.format(org_members[-1]["cursor"])
        # print("LOADING MORE MEMBERS")
        fetch_user_data(org_name, current_cursor)

# Fetch the users and export to CSV
fetch_user_data(github_org)

# Verify that the member list contains more than just the header row and export to CSV
if len(member_list) > 1:
    print("Found {} org members".format(len(member_list)-1))
    # Export to CSV with the org name in the filename, ex: github_org_members_slackhq.csv
    with open("github_org_members_{}.csv".format(github_org),"w+") as my_csv:
        csvWriter = csv.writer(my_csv,delimiter=',')
        csvWriter.writerows(member_list)
        print("Exported CSV")
else:
    print("Member list is empty.")