Crate archetype_asset 

archetype_asset—A practical, high-performance bridge for loading assets into GPU memory.

This crate isn’t a full engine; it’s the “plumbing” that handles parallel decoding, memory-mapped I/O, and spatial preloading so you can focus on rendering. We follow a “Dumb Pipe” philosophy: we parse and buffer the data, but we don’t tell you how to draw it.

Highlights

• Renderer Agonstic: Produces raw buffers compatible with Vulkan, DX12, or Metal.
• Runtime Flexible: Compiles against Tokio, async-std, or even no-std if you provide a spawner.
• Spatial Awareness: Automatically manages a preload queue based on player proximity.
• Zero-Copy: Leverages memmap2 for massive HDR cubemaps and baked assets.

Quick Start

use archetype_asset::{AssetCache, MockGpu, runtime::mock::MockSpawner};

let gpu = MockGpu::new();
let spawner = MockSpawner::new();
let cache = AssetCache::new(gpu, spawner, 100 * 1024 * 1024);
let model = cache.get_or_load_model("model.glb")?;

Feature Flags

• gpu-vulkan: Enable Vulkan GPU backend
• runtime-tokio: Enable Tokio async runtime
• lod: Enable mesh simplification with meshopt
• spatial-preload: Enable spatial preloading features

Re-exports

pub use cache::metrics::AssetMetrics;

pub use cache::metrics::AssetMetricsHandle;

pub use cache::pool::AssetMemoryPool;

pub use cache::pool::MeshData;

pub use cache::pool::TextureData;

pub use cache::pool::VertexData;

pub use cache::AssetCache;

pub use gpu::mock::MockGpu;

pub use gpu::BufferUsage;

pub use gpu::GpuDevice;

pub use gpu::GpuError;

pub use gpu::GpuResult;

pub use gpu::GpuTextureFormat;

pub use gpu::TextureDescriptor;

pub use runtime::mock::MockSpawner;

pub use runtime::AsyncSpawner;

pub use runtime::JoinHandle;

pub use model::AlphaMode;

pub use model::LoadedModel;

pub use model::Material;

pub use model::Mesh;

pub use model::ModelError;

pub use model::ModelHandle;

pub use model::ModelLoader;

pub use model::Node;

pub use model::PrimitiveType;

pub use model::Transform;

pub use texture::Texture;

pub use texture::TextureError;

pub use texture::TextureFormat;

pub use texture::TextureLoader;

pub use ibl::IblHeader;

pub use ibl::MappedIblAsset;

pub use lod::default_thresholds;

pub use lod::generate_lod_levels;

pub use lod::DefaultSimplifier;

pub use lod::LodModel;

pub use lod::MeshSimplifier;

pub use async_loading::AsyncAssetError;

pub use async_loading::AsyncModelHandle;

pub use async_loading::LoadState;

pub use spatial::DistancePredictor;

pub use spatial::SpatialPredictor;

pub use renderer::Vertex;

Modules

async_loading

Async model loading with streaming support

cache

Asset caching with LRU eviction policy

gpu

GPU abstraction layer for backend-agnostic asset loading

ibl

Image-Based Lighting (IBL) asset loading and processing

loader

Loader module for asset format handling

lod

Level of Detail (LOD) system for mesh optimization

model

Model and Mesh types

renderer

Vertex types for rendering

runtime

Async runtime abstraction for flexible async execution

spatial

Spatial preloading module

texture

Texture loading and processing

Enums

AssetError

Main error type for asset operations

Constants

VERSION

Current version of the crate

Type Aliases

Result

Result type alias for asset operations