Simple backup script in Python

Since rsync lacks compression and I'm not a big fan of rdiff, I thought I'd try my own little backup script. Since I'm very inexperienced with file management in Python, I'm sure there is lot of room for improvement.

The script creates the entire folder structure within the backup folder and then creates syncs recursively for all files within the source folders. Every file above a certain threshold gets gziped on the way.

A possible command would be:


python3 sync.py -target MY_BACKUPFOLDER -source IMPORTANT_1 IMPORTANT_2

Please tell me what you think of the source:

```
""" Simple backup script which just creates the root structure in an other
folder and syncs everything which recursevely lies within one of the source
folders. For files bigger than a threshold they are first gziped."""

import argparse
import gzip
import os
import shutil
import sys

def parse_input():
parser = argparse.ArgumentParser()
parser.add_argument('-target', nargs=1, required=True,
help='Target Backup folder')
parser.add_argument('-source', nargs='+', required=True,
help='Source Files to be added')
parser.add_argument('-compress', nargs=1, type=int,
help='Gzip threshold in bytes', default=[100000])

# no input means show me the help
if len(sys.argv) == 1:
parser.print_help()
sys.exit()

return parser.parse_args()

def size_if_newer(source, target):
""" If newer it returns size, otherwise it returns False """

src_stat = os.stat(source)
try:
target_ts = os.stat(target).st_mtime
except FileNotFoundError:
try:
target_ts = os.stat(target + '.gz').st_mtime
except FileNotFoundError:
target_ts = 0

# The time difference of one second is necessary since subsecond accuracy
# of os.st_mtime is striped by copy2
return src_stat.st_size if (src_stat.st_mtime - target_ts > 1) else False

def sync_file(

This program is really a pleasure to read. It accomplishes the desired outcome neatly and succinctly. The only thing I could think of that might be an improvement is threading. Other than that this program looks fantastic.

I can see where making this program threaded can be a great benefit when dealing with smaller files. I'm not sure, but I fear with larger files or smaller buffering, having this program run copy routines in parallel might bring the system to a halt. I wrote up a minor modification to your excellent source code to illustrate my idea:

""" Simple backup script which just creates the root structure in an other
folder and syncs everything which recursevely lies within one of the source
folders. For files bigger than a threshold they are first gziped."""

import argparse
import gzip
import os
import shutil
import sys
import threading

def parse_input():
    parser = argparse.ArgumentParser()
    parser.add_argument('-target', nargs=1, required=True,
                        help='Target Backup folder')
    parser.add_argument('-source', nargs='+', required=True,
                        help='Source Files to be added')
    parser.add_argument('-compress', nargs=1,  type=int,
                        help='Gzip threshold in bytes', default=[100000])

    # no input means show me the help
    if len(sys.argv) == 1:
        parser.print_help()
        sys.exit()

    return parser.parse_args()

def size_if_newer(source, target):
    """ If newer it returns size, otherwise it returns False """

    src_stat = os.stat(source)
    try:
        target_ts = os.stat(target).st_mtime
    except FileNotFoundError:
        try:
            target_ts = os.stat(target + '.gz').st_mtime
        except FileNotFoundError:
            target_ts = 0

    # The time difference of one second is necessary since subsecond accuracy
    # of os.st_mtime is striped by copy2
    return src_stat.st_size if (src_stat.st_mtime - target_ts > 1) else False

def threaded_sync_file(source, target, compress):
    size = size_if_newer(source, target)

    if size:
        thread = threading.Thread(target=transfer_file, 
                                  args=(source, target, size > compress))
        thread.start()
        return thread

def sync_file(source, target, compress):
    size = size_if_newer(source, target)

    if size:
        transfer_file(source, target, size > compress)

def transfer_file(source, target, compress):
    """ Either copy or compress and copies the file """

    try:
        if compress:
            with gzip.open(target + '.gz', 'wb') as target_fid:
                with open(source, 'rb') as source_fid:
                    target_fid.writelines(source_fid)
            print('Compress {}'.format(source))
        else:
            shutil.copy2(source, target)
            print('Copy {}'.format(source))
    except FileNotFoundError:
        os.makedirs(os.path.dirname(target))
        transfer_file(source, target, compress)

def sync_root(root, arg):
    target = arg.target[0]
    compress = arg.compress[0]
    threads = []

    for path, _, files in os.walk(root):
        for source in files:
            source = path + '/' + source
            threads.append(threaded_sync_file(source, 
                           target + source, compress))
#            sync_file(source, target + source, compress)
    for thread in threads:
        thread.join()

if __name__ == '__main__':
    arg = parse_input()
    print('### Start copy ####')
    for root in arg.source:
        sync_root(root, arg)
    print('### Done ###')

StackExchange Code Review Q#101616, answer score: 15