npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@reearth/ezu

v0.10.0

Published

Painterly map renderer — WebAssembly bindings

Readme

@reearth/ezu

Painterly map renderer compiled to WebAssembly. The JS side owns all I/O (HTTP, PMTiles, asset fetching); this package exposes a stateful Renderer that holds a parsed style document, its built graph, an in-memory brush bank, and a per-tile binding buffer that mirrors the style's sources block.

Install

npm install @reearth/ezu

Usage

The package ships four builds and picks one automatically via the exports map:

| Runtime | Build used | | -------------------- | ----------- | | Vite / Webpack / Rollup | bundler | | Browser (native ESM) | web | | Node.js | nodejs | | Cloudflare Workers | workerd (via the workerd condition) |

Every build also exists in a wasm SIMD128 flavor under @reearth/ezu/simd — see SIMD.

Bundler / Cloudflare Workers / Node.js

import { Renderer } from "@reearth/ezu";

const renderer = new Renderer(styleJson);
renderer.bindSource("basemap", mvtBytes);
const png = renderer.renderTile(z, x, y);
renderer.clearSources();

Browser (native ESM, no bundler)

import init, { Renderer } from "@reearth/ezu/web";

await init();
const renderer = new Renderer(styleJson);

Explicit build selection

If the auto-resolved entry doesn't fit your toolchain, import a build directly: @reearth/ezu/web, @reearth/ezu/bundler, @reearth/ezu/nodejs, or @reearth/ezu/workerd.

workerd reuses the bundler glue but instantiates the wasm itself, because Cloudflare Workers resolve a .wasm import to an uninstantiated WebAssembly.Module rather than to live exports. Importing @reearth/ezu/bundler on Workers therefore fails with __wbindgen_start is not a function.

The raw artifacts are exported too, for hosts that need to wire up instantiation themselves: @reearth/ezu/bundler/ezu_bg.js, @reearth/ezu/bundler/ezu_bg.wasm, and @reearth/ezu/web/ezu_bg.wasm.

SIMD

The same four builds are shipped a second time compiled with -C target-feature=+simd128, under the simd subpath. It resolves by the same conditions as the default entry, so one import covers every runtime:

import { Renderer } from "@reearth/ezu/simd";

The default entry (.) stays non-SIMD on purpose. Safari gained wasm SIMD128 only in 16.4, and a module that uses SIMD instructions fails to compile on older engines rather than degrading — so making SIMD the package default would hard-break those browsers. Opt in where you know the engine supports it: Cloudflare Workers (workerd), Node 16+, and Chrome/Firefox/Safari 16.4+.

simdEnabled() reports which flavor is loaded — true only in the simd builds:

import { simdEnabled } from "@reearth/ezu/simd";
simdEnabled(); // true

Individual builds and raw artifacts mirror the default layout: @reearth/ezu/simd/web, /simd/bundler, /simd/nodejs, /simd/workerd, @reearth/ezu/simd/bundler/ezu_bg.js, @reearth/ezu/simd/bundler/ezu_bg.wasm, and @reearth/ezu/simd/web/ezu_bg.wasm.

A future version could ship a single entry that feature-detects SIMD at load time and instantiates the matching wasm; that would double the bytes a browser may download, so for now the choice is explicit.

API

function simdEnabled(): boolean;

class Renderer {
  constructor(styleJson: string);

  setStyle(styleJson: string): number;            // → new node count
  readonly tileSize: number;

  // Bind tile bytes under a `sources.<name>` entry from the style.
  // The renderer dispatches on the source's declared `type`:
  //   - brush  → parse `.myb` JSON, persistent
  //   - image  → decode PNG/WebP, persistent
  //   - mvt/pmtiles → decode at render time, cleared by `clearSources`
  //   - dem    → decode + stitch at render time. Pass each neighbour
  //              tile with `{ coord: [dx, dy] }` (dx, dy ∈ {-1, 0, 1}).
  bindSource(name: string, bytes: Uint8Array,
             opts?: { coord?: [number, number] }): void;
  clearSources(): void;
  boundSources(): string[];     // names with at least one pending tile-scoped binding

  // Single unified render. `opts.format` picks the encoder; `tileSize`
  // / `pad` override the canvas for hi-DPI / preview.
  renderTile(z: number, x: number, y: number, opts?: {
    format?: "png" | "webp" | "rgba";      // default "png"
    tileSize?: number;
    pad?: number;
    // Render-time overrides for the style's declared `params`, validated
    // against the declarations (type, min/max) — an unknown name or an
    // out-of-range value throws. Names left out keep their default. One
    // built graph serves every combination.
    params?: Record<string, number | boolean | string>;
    png?: { compression?: "fast" | "default" | "best" };
  }): Uint8Array;

  free(): void;
}

Output formats

  • format: "png" (default) returns PNG bytes — feed to <img> via URL.createObjectURL(new Blob([buf], { type: "image/png" })).
  • format: "webp" returns lossless WebP bytes (~15 % smaller than PNG on painterly tiles). Pure-Rust encoder, no native deps.
  • format: "rgba" returns straight (un-premultiplied) 8-bit RGBA bytes (tile_size * tile_size * 4) — feed directly to ctx.putImageData(new ImageData(new Uint8ClampedArray(buf.buffer), w, h), 0, 0) and skip the PNG decode round trip.

WebP is lossless via image-webp and has no quality knob — for lossy WebP without C deps, render as "rgba" then call OffscreenCanvas.convertToBlob({ type: "image/webp", quality }).

Missing tiles

Don't bind any MVT source for that tile — renderTile returns the style's paper background. features source nodes see no tile.<layer> binding and emit an empty layer, so downstream paint nodes short-circuit.

Errors

Every fallible method throws a JavaScript Error whose .name discriminates the kind:

| .name | When | |---|---| | InvalidStyle | new Renderer(...) / setStyle(...) rejected the JSON | | BrushParse | registerBrush(...) or bindSource on a brush source couldn't parse the .myb JSON | | MvtDecode | bindSource (mvt/pmtiles) or render couldn't decode the MVT bytes | | DemDecode | bindSource (dem) couldn't decode the raster-DEM tile | | UnknownSource| bindSource was given a name not in the style's sources block, or coord was malformed | | RenderFailed | A node eval failed (e.g. missing brush, downcast mismatch) | | PngEncode | PNG encoding failed | | WebpEncode | WebP encoding failed |

try {
  await loadAndRender();
} catch (e) {
  if (e.name === "InvalidStyle") showStyleError(e.message);
  else if (e.name === "MvtDecode") showFetchError(e.message);
  else throw e;
}

License

MIT OR Apache-2.0