A computational graphics framework for generative art in Zig.

Sequences provides a sketch-based programming model for creating mathematical visualizations and algorithmic art. Write simulations in continuous space, accumulate results into rasters, and export high-quality images.

# Build and run a sequence
zig build sq -Doptimize=ReleaseFast -- rw1

# Run all tests
zig build test

# Check compilation
zig build check

Each sequence consists of:

• .zig - Implementation using the Context(Config) framework
• .zon - Configuration file with runtime parameters

pub fn sequence(ctx: sq.Context(Config)) !void {
    // Setup simulation space
    const sim_rect = sq.Rect2.ofExtentCentered(ctx.config.frame);
    var raster = try sq.GrayscaleRaster.init(ctx.allocator, pixel_extent);
    
    // Run simulation
    for (0..iterations) |_| {
        const point = simulate();
        if (sim_rect.contains(point)) {
            const uv = sim_rect.getUv(point);
            raster.splat(uv, intensity, .bilinear);
        }
    }
    
    // Post-process and export
    processor.processInPlace(raster.buffer);
    try raster.writeToFile(ctx.allocator, "out.png", .{});
}

Sequence Framework

• Context(Config) - Provides services to your sequence:
  • allocator - Memory allocation
  • rng - Seeded random number generation
  • config - Parsed configuration from .zon file
  • node - Progress tracking for long operations
  • timer - Performance measurement

Coordinate Systems

• Extent2 / Rect2 - Floating-point geometry with physical dimensions
• Extent2i / Rect2i - Integer pixel-space geometry
• Uv2 - UV coordinates [0,1)×[0,1)
• Seamless conversion between simulation, UV, and pixel spaces

Rasterization

• GrayscaleRaster - Accumulation buffer with point/bilinear/box filtering
• StdGrayscalePostProcessor - Tone mapping, normalization, inversion pipeline

Utilities

• ReservoirSampler - Random sampling from streams

• Resolution independent - Work in physical units (inches), render at any DPU
• Memory explicit - All allocations use provided allocators
• Performance conscious - SIMD vectors, fused operations, progress reporting
• Test driven - Comprehensive tests for core components

See src/sequences/ for examples including:

• rw1 - Random walk/Brownian motion trails
• if1 - Iterated function systems
• if2 - IFS with nonlinear transformations

Each demonstrates different techniques for point generation, accumulation, and post-processing.