pysync

#! /usr/bin/env python

# Wrapper for rsync that adds several features:
# 1. preserves atime on the source,
#    so e.g. you can back up your mutt folder directory without
#    messing up mutt's idea of which folders have new mail).
#    TODO: preserve it on the dest end too.
# 2. Print the list of what will be copied, give the user a chance
#    to confirm, then do it.
# 3. Preserve password for remote copies, so you only have to type
#    it once, not several times.
#    Rsync has two ways to do that, os.environ["RSYNC_PASSWORD"] and
#    --password-file=FILE, but they only work with rsyncd, not with ssh.
#    So use pexpect instead.

# Copyright 2011 by Akkana Peck. Share and enjoy under the GPL v.2 or later.

import sys, os, getpass, pexpect

check_cmd = "rsync -avn --delete %s %s"
sync_cmd = "rsync -av --delete %s %s"

def Usage() :
    print "Usage: %s source dest" % sys.argv
    sys.exit(1)

# atime and mtime for every file under src:
utimes = {}

def find_all_atimes(src) :
    print "Saving all atimes"
    for root, dirs, files in os.walk(src) :
        for file in files :
            path = os.path.join(root, file)
            #atimes[file] = os.path.getatime(path)
            filestat = os.stat(path)
            utimes[path] = ( filestat.st_atime, filestat.st_mtime )

def find_atimes(files) :
    print "Saving all atimes"
    for file in files :
        #atimes[file] = os.path.getatime(os.path.join(root, file))
        filestat = os.stat(file)
        utimes[file] = ( filestat.st_atime, filestat.st_mtime )

def restore_all_atimes(src) :
    print "Restoring all atimes and mtimes: ",
    for file in utimes.keys() :
        print "\rRestoring", file,    #'\tto', utimes[file],
        os.utime(file, utimes[file])
    print "\nRestored all atimes and mtimes"

def sync(src, dst, confirm=True) :
    # Find all atimes BEFORE calling rsync, since rsync will destroy them:
    find_all_atimes(src)

    passwd = None
    if confirm :
        # use rsync to get the list of files needing sync
        print "Getting list of files that need to be transfered"
        deletefiles = []
        copyfiles = []

        child = pexpect.spawn(check_cmd % (src, dst))

        # rsync will send as its first line either a password: prompt,
        # or , if it needs no passwd, "sending incremental file list".
        # If we expect either one we can detect whether a password
        # is needed.
        got = child.expect(['password:', 'sending incremental file list'])
        if got == 0 :    # a password prompt
            passwd = getpass.getpass("Password for %s: " % dst)
            child.sendline (passwd)
            print "Sent password"
        else :           # sent file list without needing password
            print "No password needed"

        while True :
            line = child.readline()
            if not line : break
            line = line.strip()
            if not line or line.startswith("building file list") \
               or line.startswith("sending ") or line.startswith("created ") \
               or line.startswith("sent ") or line.startswith("total ") \
               or line == './' :
                continue
            elif line.startswith("deleting ") :
                deletefiles.append(line[9:])
            else :
                copyfiles.append(line)
        child.close()

        print "Files to delete:"
        for f in deletefiles :
            print "-", f
        print "\nFiles to copy:"
        for f in copyfiles :
            print "+", f

        # We could probably replace utimes list with the list
        # of files actually copied -- does rsync change atime
        # on any non-copied files? Test this.

        # Verify that the user wants to proceed:
        ans = raw_input("Proceed? (y/n) ")
        if ans and ans[0].lower() != 'y' :
            print "\nNot syncing"
            sys.exit(0)

    print "Syncing ..."
    child = pexpect.spawn(sync_cmd % (src, dst))
    if passwd :
        child.expect('password:')
        child.sendline (passwd)
        print "Sent password again"
    #child.interact()
    while child.isalive() :
        print child.readline(),
    child.close()

    restore_all_atimes(src)

if __name__ == "__main__" :
    if len(sys.argv) != 3 :
        Usage()

    src = sys.argv[1]
    dst = sys.argv[2]

    sync(src, dst, True)