rapier-html 1.2.8 → 1.2.9

package/dist/return-address.mjs

@@ -1,9 +1,5 @@
 // SPDX-License-Identifier: MIT
-// Shared by the writer and the page. Every page Rapier writes (page.mjs's wrap, the editor's Share and HTML export) carries
-// PAGE_SEED once, after its charset (which must stay in the first 1024 bytes): an HTML comment no person sees that tells an
-// agent what the file is and how to make one, never part of the carried document.
 export const PAGE_SEED = '<!-- This page is a Rapier document: Markdown in one HTML file, its source inside, offline. Read it in any browser; edit it in Rapier, https://rapier.website. Agents: https://rapier.website/agents. `npx rapier-html document.md` makes a page like this. -->';
-// A file can only return to this worker's one-use route.
 export const RETURN_ORIGIN = 'https://mcp.rapier.website';
 export function returnAddress(value) {
 	if (typeof value !== 'string' || /[\s\u0000-\u001f\u007f-\u009f\\@?#]/u.test(value))

package/dist/wrap.mjs

@@ -5,23 +5,47 @@ import {PAGE_SEED, returnAddress, returnExpiresAt} from './return-address.mjs';
 
 const START = '<script type="text/markdown" id="rapier-document"';
 const DRAWING_START = '<script type="text/plain" id="rapier-drawing"';
-// The original text is a display reference: the page opens on its diff from the current document.
 const END = '</script>';
 // The page's search words (rapier.html's RAPIER_SEO regions: description, canonical address, previews, structured data and the plain guide
-// a crawler reads) are rapier.website's own. A page that carries someone's document never claims that address or that description.
 const SEARCH_WORDS = /<!-- RAPIER_SEO_BEGIN -->[\s\S]*?<!-- RAPIER_SEO_END -->\n?/g;
 const OFFLINE_POLICY = "default-src 'none'; script-src 'unsafe-inline' 'wasm-unsafe-eval' blob:; style-src 'unsafe-inline'; " +
 	"img-src data: blob:; media-src data: blob:; font-src data: blob:; connect-src 'none'; worker-src blob:; manifest-src 'none'; base-uri 'none'; form-action 'none'";
 
-// One encoding, exactly reversible: no `<` before `/` or `!`, no CR, no NUL. engine.js _rapierDecodeCarried is its
-// inverse.
 export {encodeCarried, decodeCarried};
 
+const PREVIEW = /<meta (?:name="(?:description|generator)"|property="og:(?:title|description)") content="[^"]*" data-rapier-preview>\n?/g;
+const plainWords = text => text.replace(/!\[[^\]]*\]\[[^\]]*\]|!\[[^\]]*\]\([^)]*\)/g, '').replace(/\[([^\]]*)\](?:\([^)]*\)|\[[^\]]*\])/g, '$1')
+	.replace(/<[^>]*>/g, '').replace(/[*_`~=]+|\[\^[^\]]*\]/g, '').replace(/\s+/g, ' ').trim();
+export function pagePreview(markdown) {
+	const lines = String(markdown).replace(/^\uFEFF/, '').split(/\r?\n/);
+	let title = '', paragraph = [], fence = null, at = 0;
+	if (/^---\s*$/.test(lines[0] || '')) { at = lines.indexOf('---', 1) + 1 || lines.length; }
+	for (; at < lines.length; at++) {
+		const line = lines[at], marker = /^\s*(`{3,}|~{3,})/.exec(line)?.[1];
+		if (fence) { if (marker && marker[0] === fence[0] && marker.length >= fence.length) fence = null; continue; }
+		if (marker) { fence = marker; if (paragraph.length) break; continue; }
+		const heading = /^ {0,3}#{1,6}\s+(.*?)\s*#*\s*$/.exec(line);
+		if (heading) { if (paragraph.length) break; if (!title) title = plainWords(heading[1]); continue; }
+		if (!/\S/.test(line)) { if (paragraph.length) break; continue; }
+		if (!paragraph.length && /^\s*(?:[>|<]|[-*+]\s|\d+[.)]\s|\[[^\]]+\]:|!\[|\$\$|[-*_]{3,}\s*$)/.test(line)) continue;
+		paragraph.push(line);
+		if (title && paragraph.length > 8) break;
+	}
+	const words = plainWords(paragraph.join(' '));
+	const sentence = /^.{20,300}?[.!?](?=\s|$)/.exec(words)?.[0] || (words.length <= 200 ? words : words.slice(0, 200).replace(/\s+\S*$/, '') + '…');
+	return {title, description: sentence};
+}
+function previewHead(markdown, fallbackTitle) {
+	const {title, description} = pagePreview(markdown);
+	const meta = (key, name, value) => '<meta ' + key + '="' + name + '" content="' + value.replace(/&/g, '&amp;').replace(/"/g, '&quot;').replace(/</g, '&lt;') + '" data-rapier-preview>\n';
+	return (description ? meta('name', 'description', description) : '') + meta('property', 'og:title', title || fallbackTitle) +
+		(description ? meta('property', 'og:description', description) : '') + meta('name', 'generator', 'Rapier (https://rapier.website)');
+}
+
 function safeName(name, fallback) {
 	return [...String(name || fallback).replace(/[\\/\0"<>&]/g, '-')].slice(0, 256).join('').replace(/[\uD800-\uDBFF]$/, '') || fallback;
 }
 
-// Only the trusted shell head changes. Carried source and inert application bytes stay exact.
 export function offlineTemplate(pageHtml, address = null) {
 	const end = pageHtml.indexOf('</head>');
 	if (end < 0) throw new Error('not a Rapier page: no </head>');
@@ -35,14 +59,11 @@ export function offlineTemplate(pageHtml, address = null) {
 	const anchor = /<meta charset[^>]*>\n?/i.exec(head) || /<head[^>]*>\n?/.exec(head);
 	if (!anchor) throw new Error('not a Rapier page: no <head>');
 	const at = anchor.index + anchor[0].length;
-	// A minted return is an explicit Send back destination; boot and rendering stay local.
 	const policy = address == null ? OFFLINE_POLICY : OFFLINE_POLICY.replace("connect-src 'none'", 'connect-src ' + new URL(returnAddress(address)).origin);
 	return head.slice(0, at) + '<meta http-equiv="Content-Security-Policy" content="' + policy + '">\n' +
 		head.slice(at) + pageHtml.slice(end);
 }
 
-// What a carried page may hold beyond the base editor: Draw (its part and its letter sets) and the plug-ins a document
-// displays with. Notes and the font subsetter never travel; without the subsetter a page keeps its fonts whole.
 export const CHOICES = Object.freeze(['drawing', 'math', 'mermaid']);
 const INVENTORY = '<script type="application/json" id="rapier-builtin-plugins">';
 const element = (html, id) => {
@@ -52,15 +73,12 @@ const element = (html, id) => {
 	return {start, end: html[end] === '\n' ? end + 1 : end};
 };
 
-// The full carried editor cut to the base editor and `include` (ids from CHOICES). Only whole inert elements and the
-// inventory rows naming them are removed; every kept byte is unchanged.
 export function carry(pageHtml, include = []) {
 	const kept = new Set(include);
 	for (const id of kept) if (!CHOICES.includes(id)) throw new Error('a carried page includes ' + CHOICES.join(', ') + '; not ' + id);
 	if (kept.has('drawing')) kept.add('letters');
 	let html = pageHtml;
 	const cut = id => { const at = element(html, id); if (at) html = html.slice(0, at.start) + html.slice(at.end); };
-	// Parts beyond the base editor (Notes, page layout) never travel, as in the person's HTML editor file; Draw when chosen.
 	for (const id of [...new Set([...html.matchAll(/<script type="application\/rapier-runtime" id="(rapier-part-[a-z-]+)">/g)].map(match => match[1]))])
 		if (id !== 'rapier-part-drawing' || !kept.has('drawing')) cut(id);
 	const start = html.indexOf(INVENTORY);
@@ -72,8 +90,6 @@ export function carry(pageHtml, include = []) {
 	return html;
 }
 
-// The plug-ins a Markdown text displays with, read without a Markdown parser, so it leans toward carrying: math for a `$`
-// or `$$` span outside code, Mermaid for a mermaid fence (`flowchart(fence)` true marks one Rapier draws natively).
 export function portableNeeds(text, {base, flowchart} = {}) {
 	const needs = new Set();
 	for (const source of [text, base]) {
@@ -101,7 +117,6 @@ export function wrap(pageHtml, text, name, options) {
 	const opts = {...saved, ...(options || {})};
 	const docName = safeName(name, 'document.md');
 	if (opts.kind != null && !['markdown', 'text', 'code'].includes(opts.kind)) throw new Error('invalid document kind');
-	// `view`: the view the page opens on with the document behind it, `draw` or `notes`.
 	if (opts.view != null && !['draw', 'notes'].includes(opts.view)) throw new Error('the view is draw or notes');
 	const address = opts.return == null ? null : returnAddress(opts.return);
 	const expiry = opts.return_expires_at == null ? null : returnExpiresAt(opts.return_expires_at);
@@ -120,8 +135,6 @@ export function wrap(pageHtml, text, name, options) {
 		if (typeof opts.base === 'object') baseline = readBase(opts.base);
 		else {
 			if (typeof opts.base !== 'string') throw new Error('the base must be text');
-			// A returned Share page uses its own source dialect; its original base is the same carrier.
-			// Read that base before optional current-document history, which belongs to the current text.
 			const carriedBase = /^\s*<!doctype html/i.test(opts.base) ? takeBase(opts.base).base : null;
 			const pageBase = !carriedBase && /^\s*<!doctype html/i.test(opts.base) ? unwrap(opts.base) : null;
 			const source = carriedBase ? {text: carriedBase.text} : pageBase?.text != null ? readDocument(pageBase.text) : readDocument(opts.base);
@@ -130,21 +143,18 @@ export function wrap(pageHtml, text, name, options) {
 				by: opts.by == null ? 'Agent' : String(opts.by).trim(), at: opts.at == null ? new Date().toISOString() : String(opts.at)});
 		}
 	}
-	// Rewrapping retains the original text, hash and revision; attribution names this comparison.
 	if (baseline) baseline = readBase({...baseline, ...(opts.by != null ? {by: String(opts.by).trim()} : {}), ...(opts.at != null ? {at: String(opts.at)} : {})});
 	carriedBlocks += '\n' + writeBase(text, baseline);
-	// The agent seed goes after the charset (or <head>), once: a page that already carries it, wrapped again, loses its old one.
-	const unseeded = offlineTemplate(previous.html.replace(SEARCH_WORDS, '').split(PAGE_SEED + '\n').join(''), address);
+	const unseeded = offlineTemplate(previous.html.replace(SEARCH_WORDS, '').replace(PREVIEW, '').split(PAGE_SEED + '\n').join(''), address);
 	const headTag = /<meta charset[^>]*>\n?/i.exec(unseeded) || /<head[^>]*>\n?/.exec(unseeded);
 	if (!headTag) throw new Error('not a Rapier page: no <head>');
 	const seededAt = headTag.index + headTag[0].length;
 	const stripped = unseeded.slice(0, seededAt) + PAGE_SEED + '\n' + unseeded.slice(seededAt);
-	// Before Rapier's scripts: parsed before the engine boots.
 	const bodyTag = /<body[^>]*>/.exec(stripped);
 	if (!bodyTag) throw new Error('not a Rapier page: no <body>');
 	const at = bodyTag.index + bodyTag[0].length;
-	// A function, not a string: `$'`, `$&` and `` $` `` are replacement patterns.
-	const titled = stripped.replace(/<title>[^<]*<\/title>/, () => '<title>' + docName.replace(/&/g, '&amp;').replace(/</g, '&lt;') + ' — Rapier</title>');
+	const titled = stripped.replace(/<title>[^<]*<\/title>\n?/, all => '<title>' + docName.replace(/&/g, '&amp;').replace(/</g, '&lt;') + ' — Rapier</title>' +
+		(all.endsWith('\n') ? '\n' : '') + previewHead(text, docName.replace(/\.[^.]*$/, '')));
 	const shift = titled.length - stripped.length;
 	return titled.slice(0, at + shift) + '\n' + block + carriedBlocks + titled.slice(at + shift);
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "rapier-html",
-  "version": "1.2.8",
+  "version": "1.2.9",
   "description": "One command turns a Markdown file into one offline HTML file: the whole Rapier editor with the document inside. Opens in any browser, phone or desktop, with no install or account. Write, draw and paint watercolor, then Send back the edits to your AI agent.",
   "keywords": [
     "markdown",