hash and save files that change on your filesystem in realtime

cwatch.py

#!/usr/bin/env python3

import argparse
import datetime
import re
import os
import hashlib
import logging
import sys
import subprocess
import shutil
import shlex
import signal
import sys
import time
import traceback

parser = argparse.ArgumentParser()

parser.add_argument('--savedir', required=True)
parser.add_argument('--logdir', required=True)
parser.add_argument('--watchdir', required=True)
parser.add_argument('--interval', type=int, default=1000)
args, _ = parser.parse_known_args()

INTERVAL = args.interval
SAVE_DIR = args.savedir
LOG_DIR = args.logdir
WATCH_DIR = args.watchdir
NOTIFY_CMD = 'notifywait'
PATH_RE = re.compile(b'^Change\s*\d+\s*in\s+(\/.*)$')
LOG_FILENAME = 'gsdepot.log'

os.makedirs(SAVE_DIR, exist_ok=True)
os.makedirs(LOG_DIR, exist_ok=True)

def setup_logging():
    logFormatter = logging.Formatter("%(asctime)s [%(threadName)-12.12s] [%(levelname)-5.5s]  %(message)s")
    rootLogger = logging.getLogger()
    fileHandler = logging.FileHandler("{}/{}".format(LOG_DIR, LOG_FILENAME))
    fileHandler.setFormatter(logFormatter)
    rootLogger.addHandler(fileHandler)
    consoleHandler = logging.StreamHandler()
    consoleHandler.setFormatter(logFormatter)
    rootLogger.addHandler(consoleHandler)
    rootLogger.setLevel(logging.DEBUG)

setup_logging()

def signal_handler(sig, frame):
        print('GANGSTALKER SALISBURY DRIP DRY')
        sys.exit(0)

signal.signal(signal.SIGINT, signal_handler)

while True:
    try:
        p = subprocess.Popen([NOTIFY_CMD, WATCH_DIR],
                             stdout=subprocess.PIPE,
                             stderr=subprocess.PIPE)
        so, se = p.communicate()

        for line in so.split(b'\n'):
            print(line)
            path_line = line.split(b',')[0].strip()
            m = PATH_RE.match(path_line.strip())
            if m:
                path = m.group(1)
                logging.info("SAVING_PATH: {}".format(path))
                filename = os.path.basename(path)
                m = hashlib.sha256()
                m.update(open(path, 'rb').read())
                m.hexdigest()
                iso_time = datetime.datetime.now().strftime("%Y-%m-%d")
                os.makedirs("{}/{}".format(SAVE_DIR, iso_time), exist_ok=True)
                shutil.copy(path, 
                            "{}/{}/{}-{}".format(SAVE_DIR, 
                                                 iso_time, 
                                                 m.hexdigest(), 
                            filename.decode('utf-8')))
                os.chmod(path, 0000)
    except:
        traceback.print_exc()

    time.sleep(float(INTERVAL)/1000)

sys.exit(0)