@xsynaptic/satteri-wrap-cjk

A Sätteri hast plugin that wraps runs of CJK characters in an element (default <span class="cjk">) so they can be styled or language-tagged apart from surrounding Latin text.

How it works

The plugin visits every text node and wraps its CJK runs in place. It uses ctx.parent() to climb each node's ancestors, skipping text inside code-like tags (code, pre, kbd, samp, script, style) so their content stays verbatim, and skipping any element or component already carrying the target attribute and value so existing wrappers are never doubled.

Install

npm install @xsynaptic/satteri-wrap-cjk

This package is ESM-only.

Use

import { satteri } from '@astrojs/markdown-satteri';
import { wrapCjk } from '@xsynaptic/satteri-wrap-cjk';

export default defineConfig({
	markdown: {
		processor: satteri({ hastPlugins: [wrapCjk({ value: 'cjk' })] }),
	},
});

By default it emits <span class="cjk">...</span> as a styling hook. To emit semantic lang tags instead, pass attribute: 'lang' together with a value such as 'zh', 'ja', or 'ko'.

.cjk {
	font-style: normal !important;
	text-decoration: none !important;
	word-break: keep-all !important;
}

Options

| Option | Default | Purpose | | --- | --- | --- | | element | 'span' | Wrapper element. | | attribute | 'class' | Attribute on the wrapper. Use 'lang' for semantic language tags. | | value | 'cjk' | Attribute value, and the preset selector when set to 'zh', 'ja', 'ko', or 'cjk'. | | regex | from value | Custom pattern. The g flag is added if missing. |

Presets are built from Unicode Script Extensions (so they track the engine's Unicode version) plus the full-width ASCII block: zh (Han, Bopomofo), ja (Hiragana, Katakana, Han), ko (Hangul, Han), cjk (all of the above).

License

MIT