Photo management script. This script will copy photos from "~/Pictures/iPhone Incoming" into a tree the script creates, with folders representing month and years, and photo names timestamped.Completely based on the work of the amazing Dr. Drang; see here: http://www.leancrew.com/all-this/2013/10/photo-management-via-the-finder/You can see more a…

organize-photos.py

#!/usr/bin/python

import sys
import os
import subprocess
import shlex
import os.path
from datetime import datetime
from dateutil import tz
from dateutil.parser import parse
import argparse

######################## Functions #########################

fileTypeGroupings = {'photos': ['jpg', 'jpeg', 'jpe', 'png', 'bmp', 'raw'],
                     'movies': ['.3g2','.3gp','.asf','.asx','.avi','.flv','.m4v','.mov','.mp4','.mpg',
                                '.rm','.srt','.swf','.vob','.wmv','.aepx','.ale','.avp','.avs','.bdm',
                                '.bik','.bin','.bsf','.camproj','.cpi','.dash','.divx','.dmsm','.dream',
                                '.dvdmedia','.dvr-ms','.dzm','.dzp','.edl','.f4v','.fbr','.fcproject',
                                '.hdmov','.imovieproj','.ism','.ismv','.m2p','.mkv','.mod','.moi',
                                '.mpeg','.mts','.mxf','.ogv','.otrkey','.pds','.prproj','.psh','.r3d',
                                '.rcproject','.rmvb','.scm','.smil','.snagproj','.sqz','.stx','.swi','.tix',
                                '.trp','.ts','.veg','.vf','.vro','.webm','.wlmp','.wtv','.xvid','.yuv'],
                     'audio': ['mp3', 'wav']}

inverted = {}
for group in fileTypeGroupings:
    for extension in fileTypeGroupings[group]:
        if inverted.setdefault(extension, group) != group:
            raise Exception('duplicate shit for %s' % extension)

def fileDate(f):
  "Return the date/time on which the given file was created"
  try:
#  cDate = subprocess.check_output(['sips', '-g', 'creation', f])
#  cDate = cDate.split('\n')[1].lstrip().split(': ')[1]
#  cDate = subprocess.check_output(['mdls', f, '|', 'grep', 'kMDItemFSCreationDate'])
#  proc1 = subprocess.Popen(['mdls', '-name', 'kMDItemContentCreationDate', f], stdout=subprocess.PIPE,stderr=subprocess.PIPE)
    proc1 = subprocess.Popen(['exiftool', '-CreateDate', f], stdout=subprocess.PIPE,stderr=subprocess.PIPE)
    #
#  proc2 = subprocess.Popen(['grep','kMDItemFSCreationDate'],stdin=proc1.stdout,
#                           stdout=subprocess.PIPE,stderr=subprocess.PIPE)
#  proc1.stdout.close()
    out,err = proc1.communicate()
    print '"%s"' % out
    cDate = out.split('\n')[0].lstrip().split(': ')[1]
    theDate = datetime.strptime(cDate, "%Y:%m:%d %H:%M:%S") #parse(cDate).astimezone(tz.tzlocal())
  except Exception as e:
# fall back to sips
    print 'failed to use exif tool... falling back to sips'
    try:
        cDate = subprocess.check_output(['sips', '-g', 'creation', f])
        cDate = cDate.split('\n')[1].lstrip().split(': ')[1]
        theDate = datetime.strptime(cDate, "%Y:%m:%d %H:%M:%S")
    except Exception as ee:
        # worst case use mdls
        print 'failed to use sips... falling back to mdls'
        try:
            proc1 = subprocess.Popen(['mdls', '-name', 'kMDItemContentCreationDate', f], stdout=subprocess.PIPE,stderr=subprocess.PIPE)
            out,err = proc1.communicate()
            print '"%s"' % out
            cDate = out.split('\n')[0].lstrip().split('= ')[1]
            theDate = parse(cDate).astimezone(tz.tzlocal())
        except Exception as eee:
            print 'failed to use mdls... raising exception'
            # fallback to modified time???
            # return datetime.fromtimestamp(os.path.getmtime(f))
            raise
            

  return theDate


def filesHaveDifferentMd5s(file, duplicate):
    md5orig = getMd5(file)
    md5dup = getMd5(duplicate)
    
    return md5orig != md5dup

md5cache = {}

def getMd5(file):
    if not file in md5cache:
        proc1 = subprocess.Popen(['md5', file], stdout=subprocess.PIPE,stderr=subprocess.PIPE)
        out,err = proc1.communicate()
        md5 = out.split('\n')[0].lstrip().split('= ')[1]
        md5cache[file] = md5
    
    return md5cache[file]

def make_parser():
    """ Creates an ArgumentParser to parse the command line options. """
    parser = argparse.ArgumentParser(description='Rename photos to names based on when they were taken.')
    
    parser.add_argument('--srcdir', help='source directory.', default=os.environ['HOME'] + '/Pictures/iPhone Incoming')
    parser.add_argument('--destdir', help='destination directory.', default=os.environ['HOME'] + '/Pictures/iPhone')
    parser.add_argument('--format', help='The format for the new file names', default='%Y-%m-%d %H-%M-%S')
    
    return parser



###################### Main program ########################


parser = make_parser()
arguments = parser.parse_args()

sourceDir = arguments.srcdir
destDir = arguments.destdir
errorDir = destDir + '/Unsorted/'
fmt = arguments.format


# Where the files are and where they're going.
#sourceDir = os.environ['HOME'] + '/Desktop/synology/Unsorted'
#sourceDir = '/Volumes/media/Pictures/'
#sourceDir = '/Users/lindner/Desktop/synology'
#sourceDir = '/Volumes/Macintosh HD/Users/lindner/Pictures/Photos Library.photoslibrary/Originals/'
#destDir = os.environ['HOME'] + '/Desktop/synology/renamed'
#destDir = '/Users/lindner/Desktop/photos renamed exif'
#destDir = '/Volumes/media/iPhoto/copy_from_script'
#sourceDir = os.environ['HOME'] + '/foo/bar'
#destDir = os.environ['HOME'] + '/zed/blah'
#errorDir = destDir + '/Unsorted2/'

# The format for the new file names.
#fmt = "%Y-%m-%d %H-%M-%S"

# The problem files.
problems = []
md5duplicates = []

# Get all the files in the source folder - include sub directories
file_list = []
for root, dirs, files in os.walk(sourceDir, topdown=False):
    for name in files:
        if not name.endswith('.DS_Store'): # should I exclude other unknown extensions? or just organize everything?
            file_list.append(os.path.join(root, name))

# Prepare to output as processing occurs
lastMonth = 0
lastYear = 0

# Create the destination folder if necessary
if not os.path.exists(destDir):
  os.makedirs(destDir)
if not os.path.exists(errorDir):
  os.makedirs(errorDir)

# Copy files into year and month subfolders. Name the copies according to
# their timestamps. If more than one file has the same timestamp, add
# suffixes 'a', 'b', etc. to the names. 
for file in file_list:
#  print "Processing %s..." % file
  original = file
  filename, extension_with_dot = os.path.splitext(original)
  lowercase_ext = extension_with_dot.lower()
  no_dot_ext = lowercase_ext.rpartition(".")[-1]
  if no_dot_ext == '':
      no_dot_ext = 'no_extension'
  suffix = 'a'
  try:
    fDate = fileDate(original)
    yr = fDate.year
    mo = fDate.month

    if not lastYear == yr or not lastMonth == mo:
      sys.stdout.write('\nProcessing %04d-%02d...' % (yr, mo))
      lastMonth = mo
      lastYear = yr
    else:
      sys.stdout.write('.')
    
    newname = fDate.strftime(fmt)
    if no_dot_ext in inverted:
       group = inverted[no_dot_ext]
    else:
       group = no_dot_ext
    thisDestDir = destDir + '/%s/%04d/%02d' % (group, yr, mo)
    if not os.path.exists(thisDestDir):
      os.makedirs(thisDestDir)

    duplicate = thisDestDir + '/%s%s' % (newname,extension_with_dot)
    md5dup = False
    while os.path.exists(duplicate):
      if not filesHaveDifferentMd5s(file, duplicate):
          md5dup = True
          break;
      newname = fDate.strftime(fmt) + suffix
      duplicate = destDir + '/%s/%04d/%02d/%s%s' % (group,yr, mo, newname,extension_with_dot)
      suffix = chr(ord(suffix) + 1)
    if not md5dup:
        if subprocess.call(['cp', '-p', original, duplicate]) != 0:
            raise Exception("failed to copy files")
    else:
        print "MD5 DUP!!!"
        md5duplicates.append(file)
  except Exception as e:
    print '%s --> %s' % (e, file)
    final_dir = '%s/%s/' % (errorDir, no_dot_ext)
    if not os.path.exists(final_dir):
        os.makedirs(final_dir)
    subprocess.call(['cp', '-p', original, final_dir])
    problems.append(file)
  except:
    sys.exit("Execution stopped.")

# Report the problem files, if any.
if len(problems) > 0:
  print "\nProblem files: %s" % len(problems)
  print "\n".join(problems)
  print "These can be found in: %s" % errorDir
if len(md5duplicates) > 0:
    print "\nMD5 Duplicate files: %s" % len(md5duplicates)
    print "\n".join(md5duplicates)
    print "These files were not copied"