TranscribeYouTubePlaylist.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "!pip install faster-whisper\n",
        "!pip install pytubefix"
      ],
      "metadata": {
        "id": "ae2oMqO35dyR"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from pytubefix import Playlist\n",
        "from pytubefix import YouTube\n",
        "\n",
        "from faster_whisper import WhisperModel\n",
        "import torch\n",
        "\n",
        "from datetime import timedelta\n",
        "from tqdm import tqdm\n",
        "import os\n",
        "\n",
        "from google.colab import files\n",
        "\n",
        "def transcribe_audio(path, file_name):\n",
        "\t# Run on GPU with FP16\n",
        "\tmodel = WhisperModel(\"large-v2\", device=\"cuda\", compute_type=\"float16\")\n",
        "\n",
        "\t# or run on GPU with INT8\n",
        "\t# model = WhisperModel(\"large-v2\", device=\"cuda\", compute_type=\"int8_float16\")\n",
        "\t# or run on CPU with INT8\n",
        "\t# model = WhisperModel(\"large-v2\", device=\"cpu\", compute_type=\"int8\")\n",
        "\n",
        "\tprint(\"Whisper model loaded.\")\n",
        "\n",
        "\tsegments, info = model.transcribe(path, beam_size=5)\n",
        "\tprint(\"Detected language '%s' with probability %f\" % (info.language, info.language_probability))\n",
        "\n",
        "\tsegmentId = 1\n",
        "\ttext = \"\"\n",
        "\tfor segment in tqdm(segments):\n",
        "\t\tstartTime = str(0)+str(timedelta(seconds=int(segment.start)))+',000'\n",
        "\t\tendTime = str(0)+str(timedelta(seconds=int(segment.end)))+',000'\n",
        "\t\ttxt = segment.text\n",
        "\t\ttext = text + f\"{segmentId}\\n{startTime} --> {endTime}\\n{txt[1:] if txt[0] == ' ' else txt}\\n\\n\"\n",
        "\t\tsegmentId+=1\n",
        "\n",
        "\tsrtFilename = f'{file_name}.srt'\n",
        "\twith open(srtFilename, 'a', encoding='utf-8') as srtFile:\n",
        "\t\tsrtFile.write(text)\n",
        "\tfiles.download(srtFilename)\n",
        "\n",
        "\t#optional, to clean cache\n",
        "\ttorch.cuda.empty_cache()\n",
        "\n",
        "\treturn file_name\n",
        "\n",
        "def download_audio_as_wav(yt, video_title):\n",
        "\tstream = yt.streams.filter(only_audio=True).first()\n",
        "\tif stream:\n",
        "\t\tprint(f\"Downloading audio for '{video_title}'...\")\n",
        "\t\tstream.download(filename=f\"{video_title}.mp3\")\n",
        "\t\t# Here you can convert the downloaded mp3 file to WAV if you prefer, using a library like pydub.\n",
        "\t\t# The conversion process requires the pydub library and ffmpeg.\n",
        "\t\t# Example code for conversion:\n",
        "\t\t# from pydub import AudioSegment\n",
        "\t\t# audio = AudioSegment.from_mp3(f\"{video_title}.mp3\")\n",
        "\t\t# audio.export(f\"{video_title}.wav\", format=\"wav\")\n",
        "\t\tprint(f\"Audio downloaded for '{video_title}', starts converting..\")\n",
        "\t\ttranscribe_audio(f\"{video_title}.mp3\", video_title)\n",
        "\t\tos.remove(f\"{video_title}.mp3\")\n",
        "\telse:\n",
        "\t\tprint(f\"No audio available for '{video_title}'\")\n",
        "\n",
        "def main():\n",
        "\tplaylist_link = input(\"Enter the link to the YouTube playlist: \")\n",
        "\tplaylist = Playlist(playlist_link)\n",
        "\tplaylist._video_regex = r\"\\\"url\\\":\\\"(/watch\\?v=[\\w-]*)\"\n",
        "\tvideos = playlist.video_urls\n",
        "\tall = False\n",
        "\n",
        "\tfor video_url in videos:\n",
        "\t\tyt = YouTube(video_url)\n",
        "\t\tvideo_title = yt.title\n",
        "\t\tif not all:\n",
        "\t\t\tprint(f\"Video Title: {video_title}\")\n",
        "\t\t\tchoice = input(\"Do you want to (e)xit, (s)kip, (p)roceed or download (a)ll? (s/p/a): \").lower()\n",
        "\n",
        "\t\t\tif choice == \"a\":\n",
        "\t\t\t\tall = True\n",
        "\t\tif choice == \"e\":\n",
        "\t\t\treturn;\n",
        "\t\tif choice == \"s\":\n",
        "\t\t\tprint(f\"Skipping '{video_title}'\")\n",
        "\t\telif choice == \"p\" or all:\n",
        "\t\t\tdownload_audio_as_wav(yt, video_title)\n",
        "\t\telse:\n",
        "\t\t\tprint(\"Invalid choice. Skipping...\")\n",
        "\n",
        "if __name__ == \"__main__\":\n",
        "\tmain()"
      ],
      "metadata": {
        "id": "FChYyLJ9PDtJ"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}