Whisper Transcriber

Whisper Transcriber is a command-line application designed to transcribe audio files using OpenAI's Whisper API. It includes features for language selection, logging, and cleanup of temporary files, with built-in checks for file validity and size before processing.

---

Features

• Transcribe audio files using OpenAI's Whisper API.
• Language selection for transcription (fr for French, en for English).
• Logging for debugging and tracking transcription activities.
• Automatic cleanup of temporary files such as .pyc, __pycache__, etc.

---

Installation

Prerequisites

1. Clone the repository:

   git clone https://github.com/franckferman/whisper-transcriber.git
   cd whisper-transcriber

2. Install dependencies with Poetry:

   poetry install

3. Alternatively, you can use pip:

   pip install -r requirements.txt

---

Usage

Transcription

To transcribe an audio file, use the following command:

poetry run whisper-transcriber transcribe -f <path_to_audio_file> -k <API_KEY> -l <language_code>

Example

poetry run whisper-transcriber transcribe -f "audio/sample.mp3" -k "your_openai_api_key" -l "en"

Options

• -f, --file: Path to the audio file to transcribe.
• -k, --key: OpenAI API key for authentication.
• -l, --lang: Language code for transcription (fr or en).
• -o, --output: File path to save the transcription output as JSON.
• --debug: Enable debug logging, which creates a log file transcription.log.

Cleanup

To remove temporary files and logs from the project directory:

poetry run whisper-transcriber clean --log

Options

• --log: Enable logging for the cleanup process.

---

Development

Running Tests

To test the transcription and cleanup functions, ensure all necessary dependencies are installed:

poetry install --with dev

Then, run tests using your preferred test runner.

Formatting

• The project uses black, flake8, and mypy for formatting, linting, and type-checking.
  • Format code with: black .
  • Lint code with: flake8 .
  • Type-check code with: mypy .

---

License

This project is licensed under the GNU AGPLv3. See the LICENSE file for details.