AudioBookmarks

A personal Python tool for managing my audiobook bookmarks from various platforms. This is a personal project created for my own use and learning purposes.

Note on Accuracy: This tool uses heuristic methods to identify and transcribe audio snippets from bookmarks. While it works well in most cases (approximately 95% accuracy), it may occasionally fail to find the correct section or transcribe an incorrect portion of the audio. Review the resulting notes carefully for accuracy. Re-running the tool may yield different results.

Disclaimer

This tool is intended for personal use only. It was created as a learning exercise and to solve a personal need. Please be aware that automated interactions with audiobook platforms may be against their terms of service. Use at your own risk.

Overview

This tool automates the process of extracting and processing audiobook bookmarks from Libby and Hoopla:

1. Browser Automation: Uses Playwright to navigate to your bookmarks in each platform
2. Audio Extraction: Captures the audio around each bookmark timestamp
3. Transcription: Converts the audio to text using OpenAI's Whisper API
4. Quote Selection: Processes the transcribed text to extract the most relevant quote

Note: This tool uses OpenAI's APIs for transcription and text processing, which will incur costs based on their pricing.

Installation

1. Clone the repository:

git clone https://github.com/yourusername/audiobookmarks.git
cd audiobookmarks

2. Create and activate a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows, use: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Install Playwright browsers:

playwright install

5. Set up environment variables: Create a .env file in the root directory with the following variables:

BOOKS_DATA_DIRECTORY=/path/to/store/book/data
BROWSER_DATA_DIRECTORY=/path/to/store/browser/data # To bypass login after first run
NOTES_DIRECTORY=/path/to/store/notes
OPENAI_API_KEY=your_api_key_here  # Required for audio transcription and quote extraction
HOOPLA_USERNAME=your_hoopla_username
HOOPLA_PASSWORD=your_hoopla_password

Usage

First-Time Setup: The first time you use this tool with Libby, you must run it in debug mode (using the --debug flag) to manually sign in to your account. The browser session will be cached in your specified BROWSER_DATA_DIRECTORY, allowing subsequent runs to use these cached credentials automatically.

Run the script using the following command:

python main.py --platform <platform> --book <book_name> [--debug]

Arguments:

• --platform: The audiobook platform ("libby" or "hoopla")
• --book: The name of the audiobook
• --debug: (Optional) Run in debug mode with visible browser. Required for first-time setup.

Example:

# First time setup for Libby
python main.py --platform libby --book "Project Hail Mary" --debug

# Subsequent runs can be headless
python main.py --platform libby --book "Project Hail Mary"

Output Format

The tool creates a markdown file in your specified NOTES_DIRECTORY with the following structure:

• File Name: <Book Title>.md
• Metadata: Author information in YAML frontmatter
• Content Structure:
  • Organized by chapters
  • Each bookmark includes:
    • Bookmark number
    • Timestamp information
    • Any notes you added in the platform (for both Libby and Hoopla)
    • Transcribed quote from the audio
  • Additional details:
    • Libby shows time into chapter
    • Hoopla shows both chapter-specific and book-wide progress percentages

Example output:

---
Author: "[[Author Name]]"
---

## Chapter 1
#### Bookmark 1 *(5:30 in)*
**Your note from the platform (if any)**
Transcribed quote from the audiobook...

---

## Chapter 2
#### Bookmark 2 *(2:15 in)*
Transcribed quote from the audiobook...