zulip/tools/check-issue-labels

#!/usr/bin/env python3
import argparse
import configparser
import os
import re
import sys
from typing import Any

import requests

# Scans zulip repository for issues that don't have any `area` labels.
# GitHub API token is required as GitHub limits unauthenticated
# requests to 60/hour. There is a good chance that this limit is
# bypassed in consecutive attempts.
# The API token can be generated here
# https://github.com/settings/tokens/new?description=Zulip%20Issue%20Label%20Checker
#
# Copy conf.ini-template to conf.ini and populate with your API token.
#
# usage: python check-issue-labels
# Pass --force as an argument to run without a token.


def get_config() -> configparser.ConfigParser:
    config = configparser.ConfigParser()
    config.read(os.path.join(os.path.dirname(os.path.abspath(__file__)), "conf.ini"))
    return config


def area_labeled(issue: dict[str, Any]) -> bool:
    for label in issue["labels"]:
        label_name = str(label["name"])
        if "area:" in label_name:
            return True
    return False


def is_issue(item: dict[str, Any]) -> bool:
    return "issues" in item["html_url"]


def get_next_page_url(link_header: str) -> str | None:
    matches = re.findall(r"\<(\S+)\>; rel=\"next\"", link_header)
    try:
        return matches[0]
    except IndexError:
        return None


def check_issue_labels() -> None:
    parser = argparse.ArgumentParser()
    parser.add_argument("--force", action="store_true")
    args = parser.parse_args()

    if not args.force:
        config = get_config()
        try:
            token = config.get("github", "api_token")
        except configparser.Error:
            print(
                "Error fetching GitHub API token. Copy conf.ini-template to conf.ini and populate with "
                "your API token. If you want to continue without using a token use --force."
            )
            sys.exit(1)

    next_page_url: str | None = "https://api.github.com/repos/zulip/zulip/issues"
    unlabeled_issue_urls: list[str] = []
    while next_page_url:
        try:
            if args.force:
                response = requests.get(next_page_url)
            else:
                response = requests.get(next_page_url, headers={"Authorization": f"token {token}"})
            if response.status_code == 401:
                sys.exit("Error. Please check the token.")
            if response.status_code == 403:
                sys.exit(
                    "403 Error. This is generally caused when API limit is exceeded. You use an API "
                    "token to overcome this limit."
                )
        except requests.exceptions.RequestException as e:
            print(e)
            sys.exit(1)

        next_page_url = get_next_page_url(response.headers["Link"])
        unlabeled_issue_urls.extend(
            item["html_url"]
            for item in response.json()
            if is_issue(item) and not area_labeled(item)
        )

    if len(unlabeled_issue_urls):
        print("The following issues don't have any area labels associated with it")
        print("\n".join(unlabeled_issue_urls))
    else:
        print("No GitHub issues found with missing area labels.")


if __name__ == "__main__":
    check_issue_labels()