shrinkray

A high-performance Rust library for image processing powered by libvips.

Features

• Fast image processing - Built on libvips for exceptional performance
• Fluent API - Intuitive builder pattern for chaining operations
• Format conversion - Support for JPEG, PNG, WebP, and AVIF
• Resizing & cropping - Multiple fit modes (crop, clip, max)
• Filters & effects - Vintage, sepia, monochrome, duotone, and more
• Image adjustments - Sharpen, blur, tint, rotation, and trimming
• Type-safe - Strong typing for colours, formats, and options

Installation

cargo add shrinkray

System Dependencies

This library requires libvips 8.17.0 or higher to be installed on your system. For older libvips versions (8.16.x), use shrinkray 1.0.1.

macOS:

brew install vips

Ubuntu/Debian:

apt-get install libvips-dev

Fedora/RHEL:

dnf install vips-devel

Usage

Basic Example

use shrinkray::ImageProcessor;
use shrinkray::options::{Format, Fit};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    // Read input image
    let image_bytes = std::fs::read("input.jpg")?;

    // Process the image
    let result = ImageProcessor::new(&image_bytes)
        .resize(800, 600)
        .quality(85)
        .format(Format::Webp)
        .fit(Fit::Crop)
        .process()?;

    // Save the result
    std::fs::write("output.webp", result.bytes())?;

    Ok(())
}

Create Thumbnails

let thumbnail = ImageProcessor::new(&image_bytes)
    .thumbnail(200)  // 200x200 square thumbnail
    .process()?;

Apply Filters

let filtered = ImageProcessor::new(&image_bytes)
    .vintage(80)
    .sharpen(40)
    .process()?;

Duotone Effect

let duotone = ImageProcessor::new(&image_bytes)
    .duotone(
        0, 50, 100,      // Shadow colour (dark blue)
        255, 165, 0      // Highlight colour (orange)
    )
    .duotone_alpha(75)   // 75% opacity
    .process()?;

Rotation and Trimming

let trimmed = ImageProcessor::new(&image_bytes)
    .rotate(90)
    .trim()  // Auto-trim transparent/whitespace borders
    .process()?;

Custom Aspect Ratio

let cropped = ImageProcessor::new(&image_bytes)
    .aspect_ratio(16, 9)
    .width(1920)
    .fit(Fit::Crop)
    .process()?;

Device Pixel Ratio (Retina Support)

let retina = ImageProcessor::new(&image_bytes)
    .width(400)
    .height(300)
    .device_pixel_ratio(2)  // Output will be 800x600
    .process()?;

Available Options

Resizing

• width(i32) - Set output width
• height(i32) - Set output height
• resize(i32, i32) - Set both width and height
• thumbnail(i32) - Create square thumbnail
• fit(Fit) - Resize behavior (Crop, Clip, Max)
• device_pixel_ratio(i32) - DPR multiplier
• aspect_ratio(i32, i32) - Target aspect ratio

Format & Quality

• format(Format) - Output format (Jpeg, Png, Webp, Avif)
• quality(i32) - Quality 1-100 (default: 75)
• lossless(bool) - Enable lossless compression

Adjustments

• rotate(u16) - Rotate 90, 180, or 270 degrees
• sharpen(u8) - Sharpening intensity 1-100
• blur(u8) - Blur amount 1-100
• trim() - Auto-trim borders
• trim_colour(u8, u8, u8) - Trim specific colour
• background(u8, u8, u8) - Background colour for padding/flattening

Filters

• grayscale() - Convert to grayscale
• monochrome(u8) - Monochrome with intensity
• sepia(u8) - Sepia tone filter
• vintage(u8) - Vintage film effect
• polaroid(u8) - Polaroid effect
• kodachrome(u8) - Kodachrome film simulation
• technicolor(u8) - Technicolor effect
• tint(u8, u8, u8) - Colour tint overlay
• duotone(u8, u8, u8, u8, u8, u8) - Duotone with shadow and highlight colours
• duotone_alpha(u8) - Duotone opacity 1-100

Safety Limits

• max_megapixels(f64) - Maximum input image size
• max_output_resolution(u32) - Maximum output dimension

ProcessedImage API

The process() method returns a ProcessedImage with these methods:

• bytes(&self) -> &[u8] - Get image bytes as a slice
• into_bytes(self) -> Vec<u8> - Consume and get owned bytes
• format(&self) -> Format - Get the output format
• mime_type(&self) -> &'static str - Get MIME type string

Full Server Implementation

This library powers the shrinkray image server, which provides:

• HTTP/S3/file backend routing
• HMAC signature verification
• Prometheus metrics
• OpenTelemetry tracing
• Kubernetes deployment configs

See the main repository for the complete server implementation.

License

MIT License - see LICENSE for details.