Skip to content

Latest commit

 

History

History
143 lines (121 loc) · 3.66 KB

README.md

File metadata and controls

143 lines (121 loc) · 3.66 KB

Audio to SRT Conversion and Algolia Upload Guide

Audio to SRT Conversion

Convert a single audio file to SRT:

  1. Navigate to the project directory: 
    cd batch-folder-audio-to-srt
  2. Set up a virtual environment: 
    python3 -m venv venv
source venv/bin/activate  # For Linux/Mac
    For Windows: 
    venv\Scripts\activate
  3. Install dependencies: 
    pip install git+https://github.com/openai/whisper.git
  4. Convert an audio file (.m4a) to an SRT subtitle file: 
    python3 audio-to-srt.py "/path/to/audio/file.m4a"
  5. The generated .srt file will be saved in the ./SrtFiles directory where the script is executed.

Batch convert all audio files in a folder to SRT:

  1. Navigate to the project directory: 
    cd batch-folder-audio-to-srt
  2. Run the batch conversion script: 
    python3 run-for-folder.py
    Note: The input folder is hardcoded in the script. All SRT files will be saved to the ./SrtFiles directory.

Upload SRT to Algolia Database

Upload a single SRT file to Algolia:

  1. Use the .env-template to create and fill out a .env file, including your Algolia write access API key (find it here).
  2. Navigate to the upload-srt-to-algolia directory: 
    cd upload-srt-to-algolia
  3. Run a connection test: 
    python3 test.py
  4. Upload the SRT file: 
    python3 upload-srt-to-algolia.py "/path/to/subtitles.srt"

Upload an entire folder to Algolia:

  1. Run the following script to check the number of quotes that would be uploaded: 
    python3 find_quote_count.py "/path/to/srt/folder/"
  2. To upload all SRT files from a folder: 
    python3 upload_folder_to_algolia.py "/path/to/srt/folder/"

Benchmarks

Processing Speed Example:

  • File: "Livestream: TF2/CS:GO: April 14, 2012 (Pre Recorded) - Jerma985.m4a" (Length: 1:18:43)
  • Time Taken: 18 minutes 27 seconds
  • Processing Speed: 14.07 seconds per minute of audio

Estimate for processing a longer file:

  • File Length: 1:58:14 (118 minutes)
  • Estimated Time: 27.67 minutes
  • Example: 
    python3 convert.py "/path/to/long/audiofile.m4a"

Virtual Environment Setup Guide

Setting up and using a virtual environment:

  1. Activate the environment:
    • Windows: 
      venv_name\Scripts\activate
    • Mac/Linux: 
      source venv_name/bin/activate
  2. Deactivate the environment: 
    deactivate
  3. Manage dependencies:
    • Freeze installed dependencies to a file: 
      pip freeze > requirements.txt
    • Install dependencies from a file: 
      pip install -r requirements.txt

Convert YouTube to MP3

Download and convert YouTube videos to high-quality MP3:

yt-dlp -f bestaudio -x --audio-format mp3 --audio-quality 0 --add-metadata https://www.youtube.com/watch?v=AL2IkW4JWl4

Convert MP3 to Text Transcript

  1. Install Whisper: 
    pip install -U openai-whisper
sudo apt update && sudo apt install ffmpeg
  2. Convert MP3 to text: 
    whisper audio.mp3 --model medium

WhisperX Diarization (Speaker Segmentation)

  1. Install Conda.
  2. Initialize Conda: Getting Started.
  3. Follow the steps for installing WhisperX: WhisperX.
  4. Ensure ffmpeg is installed and accessible from the command line.