hashhaze

A high-performance Rust implementation of the BlurHash algorithm. HashHaze is a CLI tool that generates compact, visually pleasing image placeholders using the BlurHash encoding scheme.

Features

• Fast, parallel processing of multiple images using Tokio
• Support for various image formats
• Configurable X and Y components for hash generation
• Efficient CPU utilization with automatic thread management
• Simple CLI interface

Installation

From Source

# Clone the repository
git clone https://github.com/yourusername/hashhaze.git
cd hashhaze

# Build the project
cargo build --release

# The binary will be available at target/release/hashhaze

Usage

# Basic usage with a single image
hashhaze input.jpg

# Process multiple images
hashhaze image1.jpg image2.png directory/

# Customize components (default: -x 4 -y 3)
hashhaze -x 5 -y 4 input.jpg

# Process an entire directory of images
hashhaze path/to/images/

Options

• -x, --components-x <NUMBER>: Number of X components for BlurHash (default: 4)
• -y, --components-y <NUMBER>: Number of Y components for BlurHash (default: 3)

How It Works

HashHaze implements the BlurHash algorithm, which creates a compact string representation of an image placeholder. The algorithm works by:

1. Breaking down the image into a grid (specified by X and Y components)
2. Calculating DCT (Discrete Cosine Transform) coefficients
3. Encoding the results into a base83 string

The resulting hash can be used to generate a blurred placeholder while the original image loads.

Performance

The tool is optimized for performance:

• Parallel processing using Tokio async runtime
• Automatic CPU core detection for optimal thread usage
• Memory-efficient image processing

Dependencies

• clap: Command-line argument parsing
• tokio: Async runtime for parallel processing
• imx: Image processing
• xio: File I/O operations
• futures: Async utilities
• num_cpus: CPU core detection
• anyhow: Error handling
• thiserror: Custom error types

License

This project is licensed under the MIT License - see the LICENSE.md file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Acknowledgments

• Original BlurHash algorithm by Wolt
• All contributors to the project