drop.chapril.org-tools/cleaning/remove_expired_files.py

126 lines
3.2 KiB
Python

import logging
import argparse
import os
import re
import datetime
import redis
BACKUP_START_TIME = datetime.time(1, 30, 0)
BACKUP_END_TIME = datetime.time(3, 30, 0)
def parse_args():
parser = argparse.ArgumentParser()
parser.add_argument("-v", "--verbose", action="store_true",
help="increase output verbosity")
parser.add_argument("-r", "--redis-host", default="127.0.0.1",
help="specify redis host")
parser.add_argument("-f", "--file-dir", default="/var/drop/files",
help="specify dir where files are stored")
parser.add_argument("-d", "--dry", action="store_true",
help="Run in dry mode")
return parser.parse_args()
def convert_bytes_to_string(bytes_or_string):
if isinstance(bytes_or_string, str):
return bytes_or_string
if isinstance(bytes_or_string, bytes):
return bytes_or_string.decode('UTF-8')
return None
def configure_logging(args):
log_level = logging.INFO
if args.verbose:
log_level = logging.DEBUG
logging.basicConfig(level=log_level)
def create_redis_client(args):
logging.debug('Connecting to redist host %s...', args.redis_host)
redis_client = redis.StrictRedis(host=args.redis_host)
logging.debug('Subscribing to events...')
pubsub = redis_client.pubsub()
pubsub.psubscribe("*")
return redis_client, pubsub
def locate_file_with_id(args, file_id):
for file in os.listdir(args.file_dir):
if re.match(r'^[0-9]+-' + file_id + r'$', file):
return os.path.join(args.file_dir, file)
return None
def delete_files(args, file_ids):
for file_id in file_ids:
delete_file(args, file_id)
def delete_file(args, file_id):
logging.info('File with id %s will be deleted', file_id)
if not re.match(r'^[a-f0-9]{16}$', file_id):
logging.warning('file id %s is not valid', file_id)
return
file_path = locate_file_with_id(args, file_id)
if not file_path:
logging.warning('file with id %s was no found', file_id)
return
logging.info('Removing file with id %s on disk', file_id)
if not args.dry:
os.remove(file_path)
def is_in_backup_period():
now = datetime.datetime.now().time()
if BACKUP_START_TIME <= BACKUP_END_TIME:
return BACKUP_START_TIME <= now <= BACKUP_END_TIME
else:
return BACKUP_START_TIME <= now or now <= BACKUP_END_TIME
def main():
args = parse_args()
file_ids_to_delete = []
configure_logging(args)
redis_client, pubsub = create_redis_client(args)
logging.debug('Listen to messages...')
for msg in pubsub.listen():
msg_type = msg.get('type', None)
channel = convert_bytes_to_string(msg.get('channel', None))
data = convert_bytes_to_string(msg.get('data', None))
if msg_type != 'pmessage':
logging.debug('Message of type %s ignored', msg_type)
continue
if channel != '__keyevent@0__:expired':
logging.debug('Event of type %s ignored', channel)
continue
file_ids_to_delete.append(data)
if not is_in_backup_period():
delete_files(args, file_ids_to_delete)
if __name__ == "__main__":
main()