Getting data about files on many Hard Drives

filetype.py

#!/bin/python

# Move to /usr/bin/filetype
# might require a mime module.
# TODO: find which pip module install was successful.

import sys
fileName = sys.argv[1]

# os.path.exists(media.encode('utf-8')

import os
stat = os.stat(fileName)
fileSize = stat.st_size

import hashlib

class FileHasher(object):
    '''
    Hash file contents
    '''
    def __init__(self, path, alg='sha1'):
        self.path = path
        if alg == 'sha1':
            self.hasher = hashlib.md5()
        elif alg == 'md5':
            self.hasher = hashlib.sha1()
        else:
            raise ValueError('Unsupported Hashing algorithm')
        with open(path, 'rb') as File:
            while True:
                data = File.read(0x100000)
                if not data:
                    break
                self.hasher.update(data)
    
    def __str__(self):
        return '{0}'.format(self.hasher.hexdigest())


fileHash = FileHasher(fileName, 'sha1')

## e.g.
#print stat
#posix.stat_result(st_mode=33279, st_ino=12602, st_dev=16657, st_nlink=1, st_uid=1024, st_gid=100, st_size=4109788, st_atime=1440119440, st_mtime=1287784330, st_ctime=1400614047)

from mimetypes import MimeTypes
mime = MimeTypes()
fileType = mime.guess_type(fileName)

print '"{}","{}","{}","{}"'.format(fileName, fileType[0], fileSize, fileHash)

index.sh

#!/bin/bash

set -e

TIMESTAMP=$(date +%Y%m%d%H%M%S)
LOGFILEPATH="/volumeUSB3/Somewhere"

declare -a DIRECTORIES=(\
              "volumeUSB1" \
              "volumeUSB2" \
              "volume1" \
            )

declare -a EXT_MUSIC=(\
              "mp3" \
              "wma" \
              "flac" \
              "ogg" \
              "ogg" \
              "m4a" \
              "oga" \
            )

declare -a EXT_VIDEOS=(\
              "mp4" \
              "mkv" \
              "flv" \
              "avi" \
              "wmv" \
              "rm" \
              "asf" \
              "m4v" \
              "mpg" \
              "mpeg" \
              "m2v" \
            )

function join_by { local d=$1; shift; echo -n "$1"; shift; printf "%s" "${@/#/$d}"; }

#find /volumeUSB2 -type f -iregex ".*\.\($(join_by \\\| ${EXT_MUSIC[@]})\)" -print
#find /volumeUSB2 -type f -iregex '.*\.\(mp3\|wma\|m4a\)$'

for d in "${DIRECTORIES[@]}"; do
  FILE="${LOGFILEPATH}music_${d}.${TIMESTAMP}.txt"
  echo "At ${d} looking for ${EXT_MUSIC[@]}, logging to ${FILE}"
  find "/${d}" -type f -iregex ".*\.\($(join_by \\\| ${EXT_MUSIC[@]})\)" -exec /usr/bin/filetype "{}" + >> "${FILE}"
  FILE="${LOGFILEPATH}video_${d}.${TIMESTAMP}.txt"
  echo "At ${d} looking for ${EXT_VIDEOS[@]}, logging to ${FILE}"
  find "/${d}" -type f -iregex ".*\.\($(join_by \\\| ${EXT_VIDEOS[@]})\)" -exec /usr/bin/filetype "{}" + >> "${FILE}"
done