sub_ref_sound_extractor.py

import csv
import subprocess
import sys
from datetime import timedelta
from pathlib import Path


ffmpeg_exe = Path("C:\\ffmpeg\\ffmpeg.exe")
field_headers = ["begin", "end", "text", "filename"]


mp3_a, csv_a = sys.argv[1], sys.argv[2]
fudge_a, pad_a = sys.argv[3], sys.argv[4]
mp3_p = Path(__file__).parent / mp3_a
csvi_p = Path(__file__).parent / csv_a

out_pd = Path(__file__).parent / f"{mp3_p.stem}"
if not out_pd.exists():
    out_pd.mkdir(parents=True, exist_ok=True)

csvo_p = out_pd / f"{csvi_p.stem}_out.csv"

with csvi_p.open(mode="r", encoding="utf-8") as csvi_f, \
        csvo_p.open(mode="w", encoding="utf-8", newline="") as csvo_f:
    reader = csv.DictReader(csvi_f)
    writer = csv.writer(csvo_f, quoting=csv.QUOTE_NONNUMERIC)
    writer.writerow(field_headers)
    for i, r in enumerate(reader, 1):
        begin_s, end_s, text = r["begin"], r["end"], r["text"]
        # Parse begin and end times into timedeltas
        hours, minutes, seconds = begin_s[0:8].split(":")
        hours, minutes, seconds = int(hours), int(minutes), int(seconds)
        begin = timedelta(hours=hours, minutes=minutes, seconds=seconds)
        hours, minutes, seconds = end_s[0:8].split(":")
        hours, minutes, seconds = int(hours), int(minutes), int(seconds)
        end = timedelta(hours=hours, minutes=minutes, seconds=seconds)
        # Fudge the times to re-align them with the mp3 properly
        fudge = int(fudge_a)
        begin = begin - timedelta(seconds=fudge)
        end = end - timedelta(seconds=fudge)
        # Apply the padding to the times
        pad = int(pad_a)
        begin = begin - timedelta(seconds=pad)
        end = end + timedelta(seconds=pad)
        # Create the output mp3 path
        out_pf = out_pd / f"{i}.mp3"
        # Write an updated entry to the CSV subtitle extraction list
        writer.writerow([begin, end, text, out_pf.name])
        # Form the ffmpeg command
        command = [str(ffmpeg_exe), "-y",
                   "-i", str(mp3_p),
                   "-ss", f"{begin}",
                   "-to", f"{end}",
                   str(out_pf)]
        # Run the command
        subprocess.run(command)