Cleanup GitHub Actions artifacts

cleanup_artifacts.py

# Requirements: Python 3.9+
# pip install PyGithub
# Get token here https://github.com/settings/tokens/new (need all repo permissions)
# Run the script: TOKEN='<token>' REPOSITORY=<org>/<repo> python cleanerV2.py <start_page>

import time
import sys
import datetime
import os

from github import Github
from github.GithubException import RateLimitExceededException


TOKEN = os.environ.get("TOKEN")
REPOSITORY = os.environ.get("REPOSITORY")
CUTOFF_DATE = datetime.datetime.now() - datetime.timedelta(days=7)  # 1 week ago
CUTOFF_SIZE = 1  # 1 B



def _delete_artifacts():
    start_page = int(sys.argv[1]) if len(sys.argv) > 1 else 0

    g = Github(
        login_or_token=TOKEN,
        per_page=100,
    )

    candidates_to_delete = 0
    candidates_size = 0
    keep_size = 0

    try:
        repo = g.get_repo(REPOSITORY)

        repo.get_artifacts()
        artifacts_response = repo.get_artifacts()

        current_page = start_page
        for i in range(1000):
            print("CURRENT PAGE:", current_page, "RATE LIMIT", g.rate_limiting, " --------------------------------")
            print("CANDIDATES COUNT:", candidates_to_delete, "CANDIDATE SIZE", candidates_size, "KEEP SIZE", keep_size)
            artifacts = artifacts_response.get_page(current_page)

            if len(artifacts) == 0:
                print("No artifacts found, aborting")
                exit(0)

            for artifact in artifacts:
                to_be_deleted = artifact.created_at < CUTOFF_DATE and artifact.size_in_bytes > CUTOFF_SIZE
                print(artifact.created_at, artifact.name, artifact.expires_at, artifact.size_in_bytes, to_be_deleted)
                if to_be_deleted:
                    candidates_size += artifact.size_in_bytes
                    candidates_to_delete += 1
                    artifact.delete()
                else:
                    keep_size += artifact.size_in_bytes
            current_page += 1

    except RateLimitExceededException:
        reset_time = datetime.datetime.fromtimestamp(g.rate_limiting_resettime)
        delta = reset_time - datetime.datetime.now()
        print("Rate limit exceeded, reset in:", delta.seconds, "seconds")

        time.sleep(60)

        _delete_artifacts()


if __name__ == "__main__":
    _delete_artifacts()