path: root/subsync.py

#!/usr/bin/env python

import json
import os
import os.path
import re
import subprocess
import sys
import time
from multiprocessing import Pool

import chardet

os.chdir(os.path.dirname(os.path.realpath(__file__)))
os.chdir('..')


class mopen:
    def __init__(self, filename, mode='r'):
        self.real_encoding = chardet.detect(open(filename, 'rb').read())['encoding'].lower()
        self.file_o = open(filename, mode, encoding=self.real_encoding)

    def __getattr__(self, name):
        return getattr(self.file_o, name)

    def __enter__(self):
        return self.file_o

    def __exit__(self, type, value, traceback):
        self.file_o.close()


def remove_styles(lines, styles):
    res = []
    for line in lines:
        if line.startswith("Dialogue: "):
            style = line.split(',')[3]
            if style not in styles:
                continue
        res.append(line)
    return res


def count_style_occurences(lines, style):
    res = 0
    for line in lines:
        if line.startswith("Dialogue: "):
            line_style = line.split(',')[3]
            if style == line_style:
                res += 1
    return res


def get_example_lines(lines, style, sample=5):
    res = []
    for line in lines:
        if line.startswith("Dialogue: "):
            line_style = line.split(',')[3]
            if style == line_style:
                res.append(line)
                sample -= 1
                if sample == 0:
                    return res
    return res


def get_styles(lines):
    res = []
    for line in lines:
        if line.startswith('Style: '):
            res.append(line[7:].split(',')[0])
    return res


def select_keep_styles(all_styles, lines):
    styles_occ = sorted([(s, count_style_occurences(lines, s)) for s in all_styles], key=lambda x: x[1], reverse=True)
    for i, s in enumerate(styles_occ):
        print("[{}] {} ({} lines use this style)".format(i, s[0], s[1]))
        if i < 3:
            print(''.join(get_example_lines(lines, s[0])))
    keep_indices = input("Which sub styles to keep? Enter numbers separated by spaces or leave blank for all: ").split()
    if not keep_indices:
        return all_styles
    return [s[0] for i, s in enumerate(styles_occ) if str(i) in keep_indices]


def extract_eng_subs(mkv, index=None):
    print("\nExtracting english subtitles, this may take a while...\n")
    mkv_json = json.loads(subprocess.check_output([
        "ffmpeg",
        "-v",
        "quiet",
        "-print_format",
        "json",
        "-show_streams",
        "-select_streams",
        "s",
        mkv
    ]))

    # Use the subtitle track as provided. If there is not one,
    # ask instead.
    all_streams = {s["index"]: s for s in mkv_json["streams"]}
    if index not in all_streams:
        index = None
    if index is None:
        if not mkv_json.get("streams"):
            raise Exception("No subtitle streams to extract? Can't do any syncing. {}".format(mkv))
        elif len(mkv_json["streams"]) == 1:
            index = mkv_json["streams"][0]["index"]
        else:
            print("[id]: Tag Information")
            for s in mkv_json["streams"]:
                tags = 'Unknown'
                try:
                    tags = str(s['tags'])
                except:
                    pass
                print(f"[{s['index']}]:\n{tags}\n")
            index = int(input("Pick the stream to retime against: "))

    # Extract
    stream = all_streams[index]
    codec_name = stream['codec_name']
    if codec_name == "subrip": codec_name = "srt"
    if codec_name not in ("srt", "ssa", "ass", "idx"):
        print("\n\n")
        print("#" * 20)
        print("Embedded subtitle stream is an image based subtitle file.")
        print("These are not supported by Alass and cannot be used.")
        print("#" * 20)
        sys.exit(1)
    extracted = mkv.replace(".mkv", f".EXTRACTED.{codec_name}")
    subprocess.run(["mkvextract", "tracks", mkv, f"{index}:{extracted}"])
    return index


def fix_styling():
    extracted_subs = [f for f in os.listdir() if '.EXTRACTED.' in f]
    if extracted_subs[0].split('.')[-1] != 'ass':
        print("Extracted subs are not .ASS format. Skipping style removal.")
        return
    all_lines = sum([mopen(f).readlines() for f in extracted_subs], [])
    all_styles = list(set(get_styles(all_lines)))
    keep_styles = select_keep_styles(all_styles, all_lines)
    for sub in extracted_subs:
        with mopen(sub) as ass:
            lines = ass.readlines()
        lines = remove_styles(lines, keep_styles)
        with open(sub, 'w', encoding="utf-8") as ass:
            ass.write('\r\n'.join(lines))


def retime_based_on_audio(mkv, srt, conf):
    retimed = mkv.replace(".mkv", ".ja.RETIMED.{}".format(conf['ext']))
    subprocess.run(["alass", "--split-penalty", conf['split_pen'], mkv, srt, retimed])
    os.remove(srt)
    os.rename(retimed, mkv.replace(".mkv", ".ja.{}".format(conf['ext'])))


def retime(mkv, srt, conf):
    retimed = mkv.replace(".mkv", ".ja.RETIMED.{}".format(conf['ext']))
    extracted = [f for f in os.listdir() if mkv.replace(".mkv", ".EXTRACTED.") in f][0]
    print(["alass", "--split-penalty", conf['split_pen'], extracted, srt, retimed])
    subprocess.run(
        ["alass", "--split-penalty", conf['split_pen'], extracted, srt, retimed])
    os.remove(srt)
    os.rename(retimed, mkv.replace(".mkv", ".ja.{}".format(conf['ext'])))
    os.remove(extracted)


def fix_common_errors(srt):
    lines = mopen(srt).readlines()
    lines = [x for i, x in enumerate(lines) if
             not (x.strip() == "" and i + 1 < len(lines) and not lines[i + 1].strip().isdigit())]
    with open(srt, 'w', encoding="utf-8") as subfile:
        subfile.write(''.join(lines))


def tryint(s):
    try:
        return int(s)
    except:
        return s


def alphanum_key(s):
    return [tryint(c) for c in re.split('([0-9]+)', s)]


if __name__ == '__main__':
    # Guess the local mkv/srt pairs that need syncing
    files = os.listdir()
    CONF = {}
    print('[1] .srt files (Default)')
    print('[2] .ass files')
    choice = input("What file type are your Japanese subs? ")
    if choice.strip() == "":
        choice = 1
    CONF['ext'] = ["", "srt", "ass"][int(choice)]
    srts = sorted([f for f in files if f.endswith(f".{CONF['ext']}")], key=alphanum_key)
    mkvs = [f for f in files if f.endswith(".mkv")]
    while len(mkvs) > len(srts):
        print(
            f"Found more .mkv files than .{CONF['ext']} files. This can occur when OP/ED/extras are in the same "
            f"folder as episodes.")
        print("Please enter a pattern to filter out of mkv results. (e.g. OP)")
        pattern = input(">>> Pattern: ")
        removed = [m for m in mkvs if pattern in m]
        mkvs = [m for m in mkvs if pattern not in m]
        print("Ignoring filtered mkv files:\n{}".format("\n".join(removed)))
    mkvs = sorted([f for f in mkvs], key=alphanum_key)
    if len(mkvs) == 0 or len(mkvs) != len(srts):
        print(f"ERROR: Some .{CONF['ext']} files don't have a matching .mkv!")
        print(f"{len(srts)} {CONF['ext']} files found. {len(mkvs)} mkv files found")
        print(f"Ensure auto-sub-retimer folder is in same folder as your .mkv and .{CONF['ext']} files.")
        print(f"Ensure you have an equal number of .mkv and .{CONF['ext']} files!")
        input("press enter to exit...")
        exit(1)
    if len(sys.argv) > 1 and sys.argv[1] == "rename_only":
        for srt, mkv in zip(srts, mkvs):
            os.rename(srt, mkv.replace('.mkv', ".ja.{}".format(CONF['ext'])))
        input("Rename finished!")
        sys.exit(0)
    pool = Pool()
    print("[1] Retime using embedded subs. (Default)")
    print("[2] Retime using audio.")
    retime_choice = input("Enter desired option number: ")
    SPLIT_PEN = input("Split Penalty? Only change this if default resulted in mis-timed subs. (Default 7): ").strip()
    if not SPLIT_PEN:
        SPLIT_PEN = "7"
    CONF['split_pen'] = SPLIT_PEN
    for srt in srts:
        fix_common_errors(srt)
    if "2" in retime_choice:
        pool.starmap(retime_based_on_audio, zip(mkvs, srts, [CONF] * len(mkvs)))
    else:
        index = None
        for mkv in mkvs:
            index = extract_eng_subs(mkv, index)
        print("English Sub Extraction Complete.")
        fix_styling()
        pool.starmap(retime, zip(mkvs, srts, [CONF] * len(mkvs)))
    print("\nSuccess!\n")
    time.sleep(3)