Text Book Reader

A suite of tools to split Project Gutenberg texts into scenes and read them aloud using MLX-based streaming TTS.

Features

• Book Splitter: Automatically detects chapter/act structure and splits long text files into Markdown scene files.
• Terminal Reader: CLI-based reader with session persistence and keyboard navigation.
• GUI Reader: Tkinter-based application with folder selection, start/stop controls, and custom bookmark saving.
• Streaming TTS: Low-latency audio synthesis using mlx-community/chatterbox-turbo-fp16.

Installation

1. Ensure you have Miniforge installed.
2. Create and activate the specialized TTS environment:

   conda create -n tts python=3.12
   conda activate tts

3. Install dependencies:

   pip install mlx-audio sounddevice numpy pytest

Usage

1. Splitting a Book

Take a Project Gutenberg text (like dracula.txt) and split it into scenes:

python sample_code/splitter.py sample_book/dracula.txt sample_book/scenes

2. GUI Reader (Recommended)

Launch the graphical interface to browse scenes and listen:

python sample_code/gui_reader.py

• Click Select Book Folder and point to sample_book/scenes.
• Click Start to begin reading.
• Use Save Bookmark to persist your exact position in a scene to a file.

3. CLI Reader

For terminal enthusiasts:

python sample_code/book_reader.py sample_book/scenes

• Enter: Next scene
• p: Previous scene
• s: Skip current audio
• q: Quit and save position

Testing

Run the test suite using pytest:

pytest tests/