Script to split files via mencoder or encode to avi from mpg4/mkv.

mencoder_convert.py

#!/usr/bin/python

'''
author: nochn_kellerkind

some magic from mencoder done in this script. 
Examples of whats possible: http://www.wiki.csoft.at/index.php/MEncoder_Scripts
'''

import sys
import os
import subprocess
import fcntl
import time
import re 
from fnmatch import fnmatch
import getopt
from decimal import *

''' estimated filesize '''
DEF_EST_SIZE=2048

estimated_size = None
path	 = None
files    = []
dry_run	 = False
convert	 = False
ext_opts = None

default_tasks = [ 'convert', 'split' ]

def get_diskspace_avail(path):

  s = os.statvfs(path)
  return s.f_bsize * s.f_bavail / 1024 / 1024

def get_length(path,file):

  proc = subprocess.Popen( ['mplayer', '-frames', '0', '-identify', os.path.join(path,file)], shell=False, stdout=subprocess.PIPE, stderr=subprocess.PIPE )

  for length in (l.split('=')[1] for l in proc.stdout.read().split("\n") if 'ID_LENGTH' in l):
    return length

def mencoder_cmd(c, file):
  global dry_run

  if ext_opts:
    c.append(ext_opts)

  if not dry_run:
    proc = subprocess.Popen( c, stdout=subprocess.PIPE, stderr=subprocess.STDOUT, close_fds=True )

    ''' unblock file descriptor ''' 
    fcntl.fcntl(
        proc.stdout.fileno(),
        fcntl.F_SETFL,
        fcntl.fcntl(proc.stdout.fileno(), fcntl.F_GETFL) | os.O_NONBLOCK,
    )

    while True:
      lines = None
      time.sleep(1)
      proc.poll()

      try:
        lines = proc.stdout.read().split("\n")
      except IOError as err:
        continue

      if proc.returncode != None:
        for l in (l for l in lines if proc.returncode > 0): 
          print l
         
	sys.stdout.write("\n")
        break

      for line in lines:
        match = re.search('([0-9]{1,3}%)', line)
	if not match:
	  continue
	sys.stdout.write("\rTask: \033[0;32m%s\033[0m File: \033[0;32m%s\033[0m Dest: \033[0;32m%s\033[0m Est. Size: \033[0;32m%sMB\033[0m Progress: \033[0;32m%s\033[0m" % 
        ( file['task'], file['name'], file['dst'], file['nsize'], match.group(1) ) )
	sys.stdout.flush()

  else:
    c.insert(0,'echo')
    proc = subprocess.Popen( c, shell=False )

def prepare_files(path, files):
  global default_tasks

  if not os.path.isfile('/usr/bin/mencoder'):
    print("mencoder is not installed!")
    quit()

  for file in files:

    name   = file['name']
    length = file['length'] = get_length( path, name )
    size   = file['size']   = os.path.getsize( os.path.join( path, name ) ) / 1024 / 1024

    file['tasks'] = default_tasks

    if int(size) <= int(estimated_size):
      del( file['tasks'][file['tasks'].index('split')] )

    if not os.path.splitext(file['name'])[1][1:] in [ 'mkv', 'mp4' ] or not convert:
      del( file['tasks'][file['tasks'].index('convert')] )
      ''' avi 1 1/5 size of a mkv '''

    if len(file['tasks']) == 0:
      print "No tasks for file %s" % file['name']

    for task in file['tasks']:
      file['task'] = task
      if task == 'split':
        split_files(file)
      if task == 'convert':
        mencoder_to_avi(file)
        

def split_files(file):

  name   = file['name']
  size   = file['size']
  length = file['length']

  if get_diskspace_avail(path) < file['size']:
    print "Not enough space available, at least %sMB required!" % (file['size'])
    quit()

  (src,sfx) = os.path.splitext( file['name'] )

  ''' create count of splitted files calculated from estimated_size '''
  file['parts'] = ( size // estimated_size + ( (size // estimated_size) // (size // estimated_size) ) )

  print file['parts']


  for part in range( file['parts'] ):
    duration = Decimal( length ) / Decimal( file['parts'] )
    file['task'] = "split"
    file['nsize']  = ( Decimal(duration) / Decimal(length)  ) * size
    mencoder_create_file( file, part, duration )

def mencoder_to_avi(file):

  (src, _)	 = os.path.splitext(file['name'])
  sfx		 = "avi"
  file['dst'] 	 = "%s.%s" % (src, sfx )
  file['nsize']  = file['size'] * 1.5

  # mencoder "Original.mkv" -ovc lavc -lavcopts vcodec=mpeg4:vhq:vbitrate=6000 -oac mp3lame -lameopts vbr=3 -o "New.avi"
  ''' @todo: check if -ovc copy -oac copy is possible '''
  c = [ 'mencoder', '-o', os.path.join(path, file['dst']), '-ovc', 'lavc', '-lavcopts', 'aspect=16/9,vcodec=mpeg4:vhq:vbitrate=6000', 
	'-oac', 'mp3lame', '-lameopts', 'vbr=3', '-mc', '0', '-noskip', os.path.join(path, file['name']) ]

  proc = mencoder_cmd(c, file)

def mencoder_create_file( file, part , duration ):

  (src,sfx)   = os.path.splitext( file['name'] )
  file['dst'] = "%s_%i%s" % (src, part + 1, sfx) 

  cmd = [ 'mencoder', '-mc', '0', '-ss', str(duration * part),  '-o', os.path.join(path, file['dst']), '-ovc', 'copy', '-oac', 'copy', os.path.join(path, file['name']) ]  

  max_duration = duration * part + duration
  if ( Decimal( max_duration ) < Decimal( file['length'] ) ):
    cmd.extend( ['-endpos', str(duration)] )

  proc = mencoder_cmd(cmd, file)

def usage():
  print "Usage:\n\t%s -i <path to infile> [-s <size in MB>] [-d|--dry-run]" % (sys.argv[0])
  quit()

def main():

  global estimated_size
  global path
  global dry_run
  global convert
  global ext_opts
  optlist = None

  try:
    optlist, args = getopt.getopt( sys.argv[1:], "hi:s:dce:", [ "help", "infile=", "size=", "dry-run", "extra-opts" ] )
  except getopt.GetoptError as err:
    print "Error: %s" % ( str(err) )
    usage()

  for opt, arg in optlist:
    if opt in ("-h", "--help"):
      usage()
    if opt in ("-i", "--infile"):
      if not os.path.exists(arg):
        usage()
      else:
        path = arg
    if opt in ("-s", "--size"):
      estimated_size = int(arg)
    if opt in ("-d", "--dry-run"):
      dry_run = True
    if opt in ("-c", "--convert"):
      convert = True
    if opt in ("-e", "--extra-opts"):
      ext_opts = arg

  if not path:
    usage()

  if not estimated_size:
    estimated_size = DEF_EST_SIZE

  files.append( { 'name': os.path.basename(path) } )
  path = os.path.dirname( os.path.abspath(path) )
  
  filter = re.sub('[0-9]{1,2}', '?', files[0]['name'])

  for file in os.listdir(path):
    if not fnmatch( file, filter ):
      continue
    if file in ( f['name'] for f in files):
      continue
    msg = "Further file found: %s. Add to list? [y|n]:" % (file)
    if raw_input(msg).lower() in  ['y','yes']:
      files.append( { 'name': file } )

  prepare_files(path, files)

if __name__ == '__main__':
  main()