Audio_into_chunks.py

import os
from pydub import AudioSegment
import whisper
from deep_translator import GoogleTranslator


# @title Audio into chunks
def audio_into_chunks_transcribe_translate(audio_file,lang):
  chunk_length_seconds=11
  output_format="wav"
  # Check if file exists
  if not os.path.exists(audio_file):
    raise ValueError(f"FLAC file not found: {audio_file}")
  Transcribe_Text=[]
  # Load the FLAC audio
  audio_segment = AudioSegment.from_file(audio_file, format="flac")
  #load Model For Transcribe
  model = whisper.load_model("medium")
  # Get total audio duration in milliseconds
  total_duration_ms = audio_segment.duration_seconds * 1000

  # Calculate chunk duration in milliseconds
  chunk_duration_ms = chunk_length_seconds * 1000

  # Split audio into chunks
  start_time = 0
  chunk_num = 1
  while start_time < total_duration_ms:
    # Get the end time for the current chunk
    end_time = min(start_time + chunk_duration_ms, total_duration_ms)

    # Extract the current chunk
    chunk = audio_segment[start_time:end_time]

    # Generate output filename with sequential numbering
    output_filename = f"{os.path.splitext(os.path.basename(audio_file))[0]}_chunk_{chunk_num}.{output_format}"

    # Export the chunk as the specified format
    chunk.export(output_filename, format=output_format)

    # Update start time for the next chunk
    start_time += chunk_duration_ms
    chunk_num += 1
    #transcribe Chunks
    result = model.transcribe(output_filename)
    #translate the transcribe data
    translator=GoogleTranslator(source='auto',target=lang)
    data_trans=translator.translate(result['text'])
    Transcribe_Text.append(data_trans)
    print(data_trans)
    print(result['text'])

  return Transcribe_Text
  print(f"FLAC file '{flac_filepath}' successfully split into {chunk_num - 1} chunks.")