-
-
Save majora2007/724354d081627cfd96c24b8eefef4ec3 to your computer and use it in GitHub Desktop.
#!/usr/bin/python | |
# Removes non-LANG audio tracks and subtitles from mkv files in a directory. | |
# Original script by greenbender at https://forum.videohelp.com/threads/343271-BULK-remove-non-English-tracks-from-MKV-container | |
# Modified by Joseph Milazzo for updated MkvMerge commands. | |
# 12/3/2021: Updated to Python 3.9 | |
import os | |
import re | |
import sys | |
import subprocess | |
import json | |
# change this for other languages (3 character code) | |
LANG = "eng" | |
# set this to the path for mkvmerge | |
MKVMERGE = "mkvmerge.exe" | |
AUDIO_RE = re.compile( | |
r"Track ID (\d+): audio \([A-Z0-9_/]+\) [number:\d+ uid:\d+ codec_id:[A-Z0-9_/]+ codec_private_length:\d+ language:([a-z]{3})") | |
SUBTITLE_RE = re.compile( | |
r"Track ID (\d+): subtitles \([A-Z0-9_/]+\) [number:\d+ uid:\d+ codec_id:[A-Z0-9_/]+ codec_private_length:\d+ language:([a-z]{3})(?: track_name:\w*)? default_track:[01]{1} forced_track:([01]{1})") | |
if len(sys.argv) < 2: | |
print("Please supply an input directory") | |
sys.exit() | |
in_dir = sys.argv[1] | |
for root, dirs, files in os.walk(in_dir): | |
for f in files: | |
# filter out non mkv files | |
if not f.endswith(".mkv"): | |
continue | |
# path to file | |
path = os.path.join(root, f) | |
# build command line | |
cmd = [MKVMERGE, "-J", path] | |
# get mkv info | |
mkvmerge = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE) | |
stdout, stderr = mkvmerge.communicate() | |
if mkvmerge.returncode != 0: | |
print >> sys.stderr, "mkvmerge failed to identify", path | |
continue | |
# find audio and subtitle tracks | |
audio = [] | |
subtitle = [] | |
info_json = json.loads(stdout) | |
tracks = info_json['tracks'] | |
for track in tracks: | |
track['properties']['id'] = track['id'] | |
if track['type'] == 'audio': | |
audio.append(track) | |
elif track['type'] == 'subtitles': | |
subtitle.append(track) | |
# filter out files that don't need processing | |
if len(audio) < 2 and len(subtitle) < 2: | |
print >> sys.stderr, "nothing to do for", path | |
continue | |
# filter out tracks that don't match the language | |
audio_lang = filter(lambda a: a['properties']['language'] == LANG, audio) | |
subtitle_lang = filter(lambda a: a['properties']['language'] == LANG, subtitle) | |
# filter out files that don't need processing | |
if len(audio_lang) == 0 and len(subtitle_lang) == 0: | |
print >> sys.stderr, "no tracks with that language in", path | |
continue | |
# build command line | |
cmd = [MKVMERGE, "-o", path + ".temp"] | |
if len(audio_lang): | |
cmd += ["--audio-tracks", ",".join([str(a['id']) for a in audio_lang])] | |
for i in range(len(audio_lang)): | |
cmd += ["--default-track", ":".join([str(audio_lang[i]['id']), "0" if i else "1"])] | |
if len(subtitle_lang): | |
cmd += ["--subtitle-tracks", ",".join([str(s['id']) for s in subtitle_lang])] | |
for i in range(len(subtitle_lang)): | |
cmd += ["--default-track", ":".join([str(subtitle_lang[i]['id']), "0"])] | |
cmd += [path] | |
# process file | |
print >> sys.stderr, "Processing", path, "...", | |
mkvmerge = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE) | |
stdout, stderr = mkvmerge.communicate() | |
if mkvmerge.returncode != 0: | |
print >> sys.stderr, "Failed" | |
print(stdout) | |
continue | |
print >> sys.stderr, "Succeeded" | |
# overwrite file | |
os.remove(path) # Don't overwrite | |
os.rename(path + ".temp", path) |
@lindelux Done
Nice! It would be useful to specify more than one language to keep. (I'd like to keep both Norwegian and English)
@vidarak You should be able to easily do that by changing
subtitle_lang = filter(lambda a: a['properties']['language'] == LANG, subtitle)
to
subtitle_lang = filter(lambda a: a['properties']['language'] in LANG, subtitle)
and have LANG = ["eng", "Norwegian code"]
Thanks! I found that it didn't work with Python 3.9 and the additional language needed some extra comparisons. I'm not a developer but think I managed to fix it....
#!/usr/bin/python3
# Removes non-LANG audio tracks and subtitles from mkv files in a directory.
# Original script by greenbender at https://forum.videohelp.com/threads/343271-BULK-remove-non-English-tracks-from-MKV-container
# Modified by Joseph Milazzo for updated MkvMerge commands.
# 12/3/2021: Updated to Python 3.9
import os
import re
import sys
import subprocess
import json
def print_to_stderr(*a):
print(*a, file = sys.stderr)
# change this for other languages (3 character code)
LANG = ["eng","nor"]
# set this to the path for mkvmerge
MKVMERGE = "/usr/bin/mkvmerge"
AUDIO_RE = re.compile(
r"Track ID (\d+): audio \([A-Z0-9_/]+\) [number:\d+ uid:\d+ codec_id:[A-Z0-9_/]+ codec_private_length:\d+ language:([a-z]{3})")
SUBTITLE_RE = re.compile(
r"Track ID (\d+): subtitles \([A-Z0-9_/]+\) [number:\d+ uid:\d+ codec_id:[A-Z0-9_/]+ codec_private_length:\d+ language:([a-z]{3})(?: track_name:\w*)? default_track:[01]{1} forced_track:([01]{1})")
if len(sys.argv) < 2:
print("Please supply an input directory")
sys.exit()
in_dir = sys.argv[1]
for root, dirs, files in os.walk(in_dir):
for f in files:
# filter out non mkv files
if not f.endswith(".mkv"):
continue
# path to file
path = os.path.join(root, f)
# build command line
cmd = [MKVMERGE, "-J", path]
# get mkv info
mkvmerge = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
stdout, stderr = mkvmerge.communicate()
if mkvmerge.returncode != 0:
print_to_stderr("mkvmerge failed to identify "+ path)
continue
# find audio and subtitle tracks
audio = []
subtitle = []
info_json = json.loads(stdout)
tracks = info_json['tracks']
for track in tracks:
track['properties']['id'] = track['id']
if track['type'] == 'audio':
audio.append(track)
elif track['type'] == 'subtitles':
subtitle.append(track)
# filter out files that don't need processing
if len(audio) < 2 and len(subtitle) < 2:
print_to_stderr("nothing to do for " + path)
continue
# filter out tracks that don't match the language
audio_lang = list(filter(lambda a: a['properties']['language'] in LANG, audio))
subtitle_lang = list(filter(lambda a: a['properties']['language'] in LANG, subtitle))
# filter out files that don't need processing
if audio_lang == audio and subtitle_lang == subtitle:
print_to_stderr("nothing to do for " + path)
continue
# filter out files that don't need processing
if len(audio_lang) == 0 and len(subtitle_lang) == 0:
print_to_stderr("no tracks with that language in " + path)
continue
# build command line
cmd = [MKVMERGE, "-o", path + ".temp"]
if len(audio_lang):
cmd += ["--audio-tracks", ",".join([str(a['id']) for a in audio_lang])]
for i in range(len(audio_lang)):
cmd += ["--default-track", ":".join([str(audio_lang[i]['id']), "0" if i else "1"])]
if len(subtitle_lang):
cmd += ["--subtitle-tracks", ",".join([str(s['id']) for s in subtitle_lang])]
for i in range(len(subtitle_lang)):
cmd += ["--default-track", ":".join([str(subtitle_lang[i]['id']), "0"])]
cmd += [path]
# process file
print_to_stderr("Processing " + path + "...")
mkvmerge = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
stdout, stderr = mkvmerge.communicate()
if mkvmerge.returncode != 0:
print_to_stderr("Failed")
print(stdout)
continue
print_to_stderr("Succeeded")
# overwrite file
os.remove(path) # Don't overwrite
os.rename(path + ".temp", path)
@majora2007 what are the licensing terms of this gist?
@vidarak what are the licensing terms for your edit of @majora2007's gist?
free to use for whatever. no terms.
Licensing is free to use as well. There was no license on the code i found floating around.
thank you for the fast responses. according to https://stackoverflow.com/a/71817350 you're supposed to ask the authors if none is provided
i get this error when i run it how can i fix it?
moe@moeserver:/hdd/TEST$ python3 removeNonEnglish.py /hdd/TEST
Traceback (most recent call last):
File "/hdd/TEST/removeNonEnglish.py", line 90, in
print >> sys.stderr, "Processing", path, "...",
TypeError: unsupported operand type(s) for >>: 'builtin_function_or_method' and '_io.TextIOWrapper'. Did you mean "print(, file=<output_stream>)"?
moe@moeserver:/hdd/TEST$
@i0moe that error means the code still has some python 2 and the print statement needs to be rewritten to use python 3 syntax print().
@i0moe It's been a while, but that's probably fixed in my modified script which worked for me on python 3.9?
@majora2007 would you please update this script to keep multiple audio/subtitles tracks in a specified language?
@majora2007 would you please update this script to keep multiple audio/subtitles tracks in a specified language?
Check this line to reconfigure the languages to keep: LANG = ["eng","nor"]
If you need to differentiate between subtitle and audio languages you can check my fork of the file at https://github.com/seantcanavan/apex-plex/blob/main/main.py
The lines in question are:
# change this for other languages (3 character code)
AUDIO_LANG = ["eng", "jpn"]
SUBTITLE_LANG = ["eng"]
Unfortunately it does not work on python 3.11
File "script.py", line 73, in <module>
if len(audio_lang) == 0 and len(subtitle_lang) == 0:
^^^^^^^^^^^^^^^
TypeError: object of type 'filter' has no len()
Hey,
It doesn't seem to work, os.walk(in_dir): seems to return nothing, all my mkv files are in 'Z:\Season 1' folder
But the program doesn't find any files
It seems it is a problem if the files are on a network drive
Hey, It doesn't seem to work, os.walk(in_dir): seems to return nothing, all my mkv files are in 'Z:\Season 1' folder But the program doesn't find any files
It seems it is a problem if the files are on a network drive
There are a lot of issues in general when working across operating systems and file systems over the network.
I spent hours one time trying to debug something and I was running Python on Windows but accessing files on a Linux shared drive. Windows is not case sensitive and linux is and it was a huge learning moment for me.
I would encourage you to try running the script on the same platform as your network drive or dig further into the exact file systems you're running on both devices and why you might be seeing an issue that you are.
Could you update this for Python 3.8 please?