Function process_git_stats

scanner.py:496–592

(git_url_titles_preemptions)

            clone_or_pull_git_repository(url)

    def process_git_stats(git_url_titles_preemptions):
        GITHUB_STATS_CACHE_FILENAME = 'github-stats-cache.json'
        GITHUB_STATS_FILENAME = 'github-stats.json'

        github_stats = {}
        try:
            with open(GITHUB_STATS_CACHE_FILENAME, 'r', encoding='utf-8') as file:
                github_stats = json.load(file)
        except FileNotFoundError:
            pass

        def is_rate_limit_exceeded():
            return g.rate_limiting[0] <= 20

        if is_rate_limit_exceeded():
            print(f"GitHub API Rate Limit Exceeded: remained - {(g.rate_limiting_resettime - datetime.datetime.now().timestamp())/60:.2f} min")
        else:
            def renew_stat(url):
                if is_rate_limit_exceeded():
                    return

                if 'github.com' not in url:
                    return None

                print('.', end="")
                sys.stdout.flush()
                try:
                    # Parsing the URL
                    parsed_url = urlparse(url)
                    domain = parsed_url.netloc
                    path = parsed_url.path
                    path_parts = path.strip("/").split("/")
                    if len(path_parts) >= 2 and domain == "github.com":
                        owner_repo = "/".join(path_parts[-2:])
                        repo = g.get_repo(owner_repo)
                        owner = repo.owner
                        now = datetime.datetime.now(datetime.timezone.utc)
                        author_time_diff = now - owner.created_at
                        
                        last_update = repo.pushed_at.strftime("%Y-%m-%d %H:%M:%S") if repo.pushed_at else 'N/A'
                        item = {
                            "stars": repo.stargazers_count,
                            "last_update": last_update,
                            "cached_time": now.timestamp(),
                            "author_account_age_days": author_time_diff.days,
                        }
                        return url, item
                    else:
                        print(f"\nInvalid URL format for GitHub repository: {url}\n")
                except Exception as e:
                    print(f"\nERROR on {url}\n{e}")

                return None

            # resolve unresolved urls
            with concurrent.futures.ThreadPoolExecutor(11) as executor:
                futures = []
                for url, title, preemptions, node_pattern in git_url_titles_preemptions: