Skip to content

Turn any idea, plan, or codebase into a beautiful interactive diagram. An agent skill for Claude Code, Codex, and more.

fix(website): publish standalone case discovery metadata

tt-a1ifaf4baa4a4c4
diff --git a/website/scripts/indexnow.mjs b/website/scripts/indexnow.mjs
index ab4c57e7..0b1f40de 100644
--- a/website/scripts/indexnow.mjs
+++ b/website/scripts/indexnow.mjs
@@ -2,6 +2,7 @@ import fs from 'node:fs';
import { pathToFileURL } from 'node:url';
import { canonicalOrigin, canonicalUrl, sitePages } from '../src/data/site-urls.mjs';
import { CASES } from '../src/data/gallery-presentation.mjs';
+import { PUBLISHED_CASES } from '../src/data/published-cases.mjs';
export const key = '0e4ac4a253b420619eb3de08ba664736';
export const keyLocation = `${canonicalOrigin}/${key}.txt`;
@@ -10,6 +11,7 @@ export const endpoint = 'https://api.indexnow.org/indexnow';
const allowedUrls = new Set([
...sitePages.flatMap(page => ['en', 'zh'].map(lang => canonicalUrl(page, lang))),
...CASES.map(entry => `${canonicalOrigin}/gallery/artifacts/${entry.output.replace(/\.html$/, '')}`),
+ ...PUBLISHED_CASES.map(entry => `${canonicalOrigin}/${entry.artifact.replace(/\.html$/, '')}`),
]);
const requireValue = (ok, message) => { if (!ok) throw new Error(message); };
@@ -33,7 +35,7 @@ export function parseSitemap(xml) {
entries.push(value);
body = body.slice(item[0].length);
}
- requireValue(entries.length === allowedUrls.size, 'Sitemap must contain all main pages and selected gallery artifacts');
+ requireValue(entries.length === allowedUrls.size, 'Sitemap must contain all main pages, selected gallery artifacts and published cases');
return entries;
}
diff --git a/website/scripts/publish-seo.mjs b/website/scripts/publish-seo.mjs
index ffd57193..ee1c14fd 100644
--- a/website/scripts/publish-seo.mjs
+++ b/website/scripts/publish-seo.mjs
@@ -4,6 +4,7 @@ import path from 'node:path';
import { createHash } from 'node:crypto';
import { canonicalOrigin, canonicalUrl, canonicalPath, sitePages } from '../src/data/site-urls.mjs';
import { CASES } from '../src/data/gallery-presentation.mjs';
+import { PUBLISHED_CASES } from '../src/data/published-cases.mjs';
export const CANONICAL_ORIGIN = canonicalOrigin;
export const SEO_BLOCK_START = '<!-- archify:published-artifact-metadata:start -->';
@@ -17,15 +18,24 @@ export function stripPublishedArtifactMetadata(html) {
return html.replace(/<!-- archify:published-artifact-metadata:start -->\n[\s\S]*?<!-- archify:published-artifact-metadata:end -->\n/g, '');
}
-export function artifactMetadata(entry, presentation) {
+function publicationMetadata(entry, { title, description, language, twitter = false }) {
const url = `${CANONICAL_ORIGIN}/${entry.artifact.replace(/\.html$/, '')}`;
const source = `${CANONICAL_ORIGIN}/${entry.input}`;
const data = {
'@context': 'https://schema.org', '@type': 'CreativeWork',
- name: presentation.titleEn, description: presentation.descriptionEn,
- url, isBasedOn: source, inLanguage: 'en',
+ name: title, description,
+ url, isBasedOn: source, inLanguage: language,
};
- return `${SEO_BLOCK_START}\n<link rel="canonical" href="${escape(url)}">\n<meta name="description" content="${escape(presentation.descriptionEn)}">\n<meta property="og:type" content="article">\n<meta property="og:title" content="${escape(presentation.titleEn)}">\n<meta property="og:description" content="${escape(presentation.descriptionEn)}">\n<meta property="og:url" content="${escape(url)}">\n<script type="application/ld+json">${JSON.stringify(data).replaceAll('<', '\\u003c')}</script>\n${SEO_BLOCK_END}\n`;
+ const twitterMetadata = twitter ? `<meta name="twitter:card" content="summary">\n<meta name="twitter:title" content="${escape(title)}">\n<meta name="twitter:description" content="${escape(description)}">\n` : '';
+ return `${SEO_BLOCK_START}\n<link rel="canonical" href="${escape(url)}">\n<meta name="description" content="${escape(description)}">\n<meta property="og:type" content="article">\n<meta property="og:title" content="${escape(title)}">\n<meta property="og:description" content="${escape(description)}">\n<meta property="og:url" content="${escape(url)}">\n${twitterMetadata}<script type="application/ld+json">${JSON.stringify(data).replaceAll('<', '\\u003c')}</script>\n${SEO_BLOCK_END}\n`;
+}
+
+export function artifactMetadata(entry, presentation) {
+ return publicationMetadata(entry, { title: presentation.titleEn, description: presentation.descriptionEn, language: 'en' });
+}
+
+export function caseMetadata(entry) {
+ return publicationMetadata(entry, { ...entry, twitter: true });
}
// Only gallery entries with passing artifact checks and unchanged proof hashes
@@ -52,15 +62,29 @@ export function publishSeo(root, target) {
assert.ok(fs.existsSync(path.join(root, file)), `Missing indexable page: ${file}`);
}
const selected = selectedArtifacts(root);
- for (const { entry, presentation } of selected) {
+ // Check the explicit case inventory before writing any publication metadata.
+ // These are existing authored examples, not new generation or proof receipts.
+ for (const entry of PUBLISHED_CASES) {
+ for (const file of [entry.artifact, entry.input]) {
+ assert.ok(fs.existsSync(path.join(root, file)), `Missing published case: ${file}`);
+ }
+ }
+ const publications = [
+ ...selected.map(({ entry, presentation }) => ({ entry, metadata: artifactMetadata(entry, presentation) })),
+ ...PUBLISHED_CASES.map(entry => ({ entry, metadata: caseMetadata(entry) })),
+ ];
+ for (const { entry, metadata } of publications) {
const file = path.join(root, entry.artifact);
const original = stripPublishedArtifactMetadata(fs.readFileSync(file, 'utf8'));
- assert.match(original, /<\/head>/i, `Artifact missing head: ${entry.id}`);
- fs.writeFileSync(file, original.replace(/<\/head>/i, `${artifactMetadata(entry, presentation)}</head>`));
+ assert.match(original, /<\/head>/i, `Artifact missing head: ${entry.artifact}`);
+ // Callback insertion keeps authored $ replacement tokens literal (#826),
+ // and retains the original closing tag bytes, including its letter case.
+ fs.writeFileSync(file, original.replace(/<\/head>/i, closingTag => `${metadata}${closingTag}`));
}
const urls = [
...sitePages.flatMap(page => ['en', 'zh'].map(lang => canonicalUrl(page, lang))),
...selected.map(({ entry }) => `${CANONICAL_ORIGIN}/${entry.artifact.replace(/\.html$/, '')}`),
+ ...PUBLISHED_CASES.map(entry => `${CANONICAL_ORIGIN}/${entry.artifact.replace(/\.html$/, '')}`),
];
fs.writeFileSync(path.join(root, 'sitemap.xml'), `<?xml version="1.0" encoding="UTF-8"?>\n<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">\n${urls.map(url => ` <url><loc>${escape(url)}</loc></url>`).join('\n')}\n</urlset>\n`);
fs.writeFileSync(path.join(root, 'robots.txt'), `User-agent: *\nAllow: /\n\nSitemap: ${CANONICAL_ORIGIN}/sitemap.xml\n`);
diff --git a/website/src/data/published-cases.mjs b/website/src/data/published-cases.mjs
new file mode 100644
index 00000000..d1c989bf
--- /dev/null
+++ b/website/src/data/published-cases.mjs
@@ -0,0 +1,39 @@
+// Website discovery copy for existing standalone cases. This inventory does not
+// regenerate diagrams or assert renderer validation; source artifacts own that.
+export const PUBLISHED_CASES = [
+ {
+ artifact: 'cases/life/leave-plan.workflow.html',
+ input: 'cases/life/leave-plan.workflow.json',
+ title: '一年的年假怎么排',
+ description: '查看一份中文年假安排流程图:休假节点、剩余额度和规则提醒相互对应。此为示例计划,不代表真实日历或公司的休假政策。',
+ language: 'zh-CN',
+ },
+ {
+ artifact: 'cases/life/job-search.lifecycle.html',
+ input: 'cases/life/job-search.lifecycle.json',
+ title: 'Where Each Job Application Stands',
+ description: 'Explore an example job-application lifecycle, from saved roles and submitted applications through interviews, waiting states, and final outcomes.',
+ language: 'en',
+ },
+ {
+ artifact: 'cases/life/monthly-money.dataflow.html',
+ input: 'cases/life/monthly-money.dataflow.json',
+ title: 'Where a Paycheck Goes Each Month',
+ description: 'Follow an example monthly paycheck through checking, bills, emergency savings and retirement transfers in an interactive data-flow diagram. Amounts are illustrative.',
+ language: 'en',
+ },
+ {
+ artifact: 'cases/life/renting.sequence.html',
+ input: 'cases/life/renting.sequence.json',
+ title: '租一套房,一来一回都在这里',
+ description: '用中文时序图查看租客、中介和房东之间的看房、议价、签约、付款与交接步骤。金额和流程仅为示例。',
+ language: 'zh-CN',
+ },
+ {
+ artifact: 'cases/mco-runtime.architecture.html',
+ input: 'cases/mco-runtime.architecture.json',
+ title: 'MCO Runtime Architecture',
+ description: 'Explore the MCO runtime architecture at pinned revision 9f1a1cf: entry points, provider invocation, streamed answers and artifacts. This is an authored source-backed diagram, not an execution trace.',
+ language: 'en',
+ },
+];
diff --git a/website/test/agent-diagrams.test.mjs b/website/test/agent-diagrams.test.mjs
index a9486d33..2a9d4780 100644
--- a/website/test/agent-diagrams.test.mjs
+++ b/website/test/agent-diagrams.test.mjs
@@ -1,6 +1,7 @@
import { test } from 'node:test';
import assert from 'node:assert/strict';
import fs from 'node:fs';
+import { stripPublishedArtifactMetadata } from '../scripts/publish-seo.mjs';
import { parse } from 'parse5';
import { parseSitemap } from '../scripts/indexnow.mjs';
@@ -109,6 +110,8 @@ test('description workflow exposes pinned evidence, honest limits and localized
}
for (const extension of ['json', 'html']) {
const file = `cases/life/leave-plan.workflow.${extension}`;
- assert.deepEqual(fs.readFileSync(new URL(file, output)), fs.readFileSync(new URL(`../../docs/${file}`, import.meta.url)), 'existing artifact bytes remain unchanged');
+ const published = fs.readFileSync(new URL(file, output));
+ const original = extension === 'html' ? Buffer.from(stripPublishedArtifactMetadata(published.toString('utf8'))) : published;
+ assert.deepEqual(original, fs.readFileSync(new URL(`../../docs/${file}`, import.meta.url)), 'existing artifact bytes remain unchanged outside the publication metadata block');
}
});
diff --git a/website/test/deployment-target.test.mjs b/website/test/deployment-target.test.mjs
index 6274c0d3..bdd1b242 100644
--- a/website/test/deployment-target.test.mjs
+++ b/website/test/deployment-target.test.mjs
@@ -3,8 +3,9 @@ import { test } from 'node:test';
import assert from 'node:assert/strict';
import fs from 'node:fs';
import { parse, parseFragment } from 'parse5';
-import { SEO_BLOCK_START, SEO_BLOCK_END, stripPublishedArtifactMetadata, artifactMetadata } from '../scripts/publish-seo.mjs';
+import { SEO_BLOCK_START, SEO_BLOCK_END, stripPublishedArtifactMetadata, artifactMetadata, caseMetadata } from '../scripts/publish-seo.mjs';
import { CASES } from '../src/data/gallery-presentation.mjs';
+import { PUBLISHED_CASES } from '../src/data/published-cases.mjs';
const cloudflare = process.env.ARCHIFY_SITE_TARGET === 'cloudflare';
const output = new URL(cloudflare ? '../dist-cloudflare/' : '../dist/', import.meta.url);
@@ -12,6 +13,22 @@ const nodes = node => [node, ...(node.childNodes || []).flatMap(nodes)];
const attr = (node, name) => node?.attrs?.find(item => item.name === name)?.value;
const pages = sitePages;
+test('deployment publishes only the declared case metadata while retaining every original artifact and source byte', () => {
+ for (const entry of PUBLISHED_CASES) {
+ const html = fs.readFileSync(new URL(entry.artifact, output), 'utf8');
+ assert.equal(html.split(SEO_BLOCK_START).length, 2);
+ assert.equal(html.split(SEO_BLOCK_END).length, 2);
+ const start = html.indexOf(SEO_BLOCK_START), end = html.indexOf(SEO_BLOCK_END) + SEO_BLOCK_END.length;
+ assert.ok(start > html.indexOf('<head>') && end < html.indexOf('</head>'));
+ assert.equal(html.slice(start, end) + '\n', caseMetadata(entry), entry.artifact);
+ const all = nodes(parse(html));
+ assert.equal(all.filter(node => attr(node, 'rel') === 'canonical').length, 1);
+ assert.equal(attr(all.find(node => node.tagName === 'html'), 'lang'), entry.language);
+ assert.deepEqual(Buffer.from(stripPublishedArtifactMetadata(html)), fs.readFileSync(new URL(`../../docs/${entry.artifact}`, import.meta.url)));
+ assert.deepEqual(fs.readFileSync(new URL(entry.input, output)), fs.readFileSync(new URL(`../../docs/${entry.input}`, import.meta.url)));
+ }
+});
+
test('built deployment uses host-compatible generated assets and canonical metadata in both languages', () => {
for (const lang of ['en', 'zh']) for (const page of pages) {
const file = page === 'index' ? `${lang === 'zh' ? 'zh' : 'index'}.html` : `${lang === 'zh' ? 'zh/' : ''}${page}.html`;
diff --git a/website/test/helpers/community-fixture.mjs b/website/test/helpers/community-fixture.mjs
index 212c094f..937de4f6 100644
--- a/website/test/helpers/community-fixture.mjs
+++ b/website/test/helpers/community-fixture.mjs
@@ -3,6 +3,7 @@ import fs from 'node:fs';
import os from 'node:os';
import path from 'node:path';
import { fileURLToPath } from 'node:url';
+import { PUBLISHED_CASES } from '../../src/data/published-cases.mjs';
export const websiteRoot = fileURLToPath(new URL('../../', import.meta.url));
const repoRoot = path.resolve(websiteRoot, '..');
@@ -26,7 +27,7 @@ export function createCommunityFixture() {
'website/src/', 'website/public/', 'website/astro.config.mjs', 'website/scripts/', 'website/package.json',
'archify/package.json', 'archify/skill-release.json', 'archify/recipes/',
'scripts/site-copy.mjs', 'scripts/check-community-packages.mjs',
- 'docs/gallery/', 'docs/assets/', 'docs/cases/life/'], { cwd: repoRoot, encoding: 'utf8' })
+ 'docs/gallery/', 'docs/assets/', ...PUBLISHED_CASES.flatMap(entry => [`docs/${entry.artifact}`, `docs/${entry.input}`])], { cwd: repoRoot, encoding: 'utf8' })
.split('\0').filter(Boolean);
for (const relative of tracked) copy(relative);
fs.symlinkSync(path.join(websiteRoot, 'node_modules'), path.join(root, 'website/node_modules'), process.platform === 'win32' ? 'junction' : 'dir');
diff --git a/website/test/indexnow.test.mjs b/website/test/indexnow.test.mjs
index da9b8409..59764863 100644
--- a/website/test/indexnow.test.mjs
+++ b/website/test/indexnow.test.mjs
@@ -3,9 +3,10 @@ import assert from 'node:assert/strict';
import fs from 'node:fs';
import { canonicalUrl, sitePages, canonicalOrigin } from '../src/data/site-urls.mjs';
import { CASES } from '../src/data/gallery-presentation.mjs';
+import { PUBLISHED_CASES } from '../src/data/published-cases.mjs';
import { notifyIndexNow, parseSitemap, endpoint, key, keyLocation, sitemapLocation } from '../scripts/indexnow.mjs';
-const urls = [...sitePages.flatMap(page => ['en', 'zh'].map(lang => canonicalUrl(page, lang))), ...CASES.map(entry => `${canonicalOrigin}/gallery/artifacts/${entry.output.replace(/\.html$/, '')}`)];
+const urls = [...sitePages.flatMap(page => ['en', 'zh'].map(lang => canonicalUrl(page, lang))), ...CASES.map(entry => `${canonicalOrigin}/gallery/artifacts/${entry.output.replace(/\.html$/, '')}`), ...PUBLISHED_CASES.map(entry => `${canonicalOrigin}/${entry.artifact.replace(/\.html$/, '')}`)];
const xml = `<?xml version="1.0" encoding="UTF-8"?><urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">${urls.map(url => `<url><loc>${url}</loc></url>`).join('')}</urlset>`;
function mock({ sitemap = xml, hostedKey = key, getStatus = 200, postStatus = 200 } = {}) {
const calls = [];
@@ -55,7 +56,15 @@ test('public preflight requires HTTP 200 and matching deployed key before POST',
test('sitemap XML parser rejects foreign, query, source, duplicate, incomplete and malformed inventories', () => {
assert.deepEqual(parseSitemap(xml), urls);
+ assert.equal(urls.length, 31);
for (const sitemap of [xml.replace('https://archify.si/', 'http://archify.si/'), xml.replace('https://archify.si/', 'https://other.example/'), xml.replace(urls[0], `${urls[0]}?preview=1`), xml.replace(urls[0], `${canonicalOrigin}/gallery/manifest.json`), xml.replace(urls[0], urls[1]), xml.replace(/<url>[\s\S]*?<\/url>/, ''), xml.replace('<loc>', '<wrong>'), '<!DOCTYPE foo>' + xml]) {
assert.throws(() => parseSitemap(sitemap));
}
+ for (const entry of PUBLISHED_CASES) {
+ const url = `${canonicalOrigin}/${entry.artifact.replace(/\.html$/, '')}`;
+ for (const replacement of [url + '.html', url + '?lang=zh', `${canonicalOrigin}/${entry.input}`, `${canonicalOrigin}/zh/${entry.artifact.replace(/\.html$/, '')}`, `${canonicalOrigin}/cases/unlisted-example`]) {
+ assert.throws(() => parseSitemap(xml.replace(url, replacement)), /outside the published|Noncanonical/);
+ }
+ assert.throws(() => parseSitemap(xml.replace(`<url><loc>${url}</loc></url>`, '')), /must contain all/);
+ }
});
diff --git a/website/test/mco-source-case.test.mjs b/website/test/mco-source-case.test.mjs
index 6926d65f..26a940a6 100644
--- a/website/test/mco-source-case.test.mjs
+++ b/website/test/mco-source-case.test.mjs
@@ -3,6 +3,7 @@ import assert from 'node:assert/strict';
import fs from 'node:fs';
import { parse } from 'parse5';
import { parseSitemap } from '../scripts/indexnow.mjs';
+import { stripPublishedArtifactMetadata } from '../scripts/publish-seo.mjs';
const cloudflare = process.env.ARCHIFY_SITE_TARGET === 'cloudflare';
const output = new URL(cloudflare ? '../dist-cloudflare/' : '../dist/', import.meta.url);
const origin = cloudflare ? 'https://archify.si/' : 'https://tt-a1i.github.io/archify/';
@@ -57,8 +58,11 @@ test('MCO case exposes pinned source, honest limits, stable sections and nested
test('MCO case inventory includes both wrappers and preserves original artifact and typed source bytes', () => {
const urls = parseSitemap(fs.readFileSync(new URL('sitemap.xml', output), 'utf8'));
for (const prefix of ['', 'zh/']) assert.ok(urls.includes(`https://archify.si/${prefix}cases/mco-runtime`));
+ assert.ok(urls.includes('https://archify.si/cases/mco-runtime.architecture'), 'the standalone artifact is distinct from the explanatory article');
for (const extension of ['html', 'json']) {
const file = `cases/mco-runtime.architecture.${extension}`;
- assert.deepEqual(fs.readFileSync(new URL(file, output)), fs.readFileSync(new URL(`../../docs/${file}`, import.meta.url)), file);
+ const published = fs.readFileSync(new URL(file, output));
+ const original = extension === 'html' ? Buffer.from(stripPublishedArtifactMetadata(published.toString('utf8'))) : published;
+ assert.deepEqual(original, fs.readFileSync(new URL(`../../docs/${file}`, import.meta.url)), file);
}
});
diff --git a/website/test/migration.test.mjs b/website/test/migration.test.mjs
index c52a54d8..649dba1f 100644
--- a/website/test/migration.test.mjs
+++ b/website/test/migration.test.mjs
@@ -5,7 +5,8 @@ import path from 'node:path';
import { fileURLToPath } from 'node:url';
import { parse, parseFragment } from 'parse5';
import { CASES } from '../src/data/gallery-presentation.mjs';
-import { artifactMetadata, SEO_BLOCK_START, SEO_BLOCK_END, stripPublishedArtifactMetadata } from '../scripts/publish-seo.mjs';
+import { PUBLISHED_CASES } from '../src/data/published-cases.mjs';
+import { artifactMetadata, caseMetadata, SEO_BLOCK_START, SEO_BLOCK_END, stripPublishedArtifactMetadata } from '../scripts/publish-seo.mjs';
const root = fileURLToPath(new URL('../', import.meta.url));
const docs = path.resolve(root, '../docs');
@@ -122,8 +123,9 @@ test('all existing non-page public URLs retain exact bytes outside the declared
const selected = new Map(CASES.map(presentation => {
const entry = manifest.entries.find(item => item.id === presentation.id);
assert.ok(entry, presentation.id);
- return [entry.artifact, { entry, presentation }];
+ return [entry.artifact, { entry, metadata: artifactMetadata(entry, presentation), isCase: false }];
}));
+ for (const entry of PUBLISHED_CASES) selected.set(entry.artifact, { entry, metadata: caseMetadata(entry), isCase: true });
function publishedBytes(file) {
const bytes = fs.readFileSync(path.join(dist, file));
const proof = selected.get(file.replaceAll(path.sep, '/'));
@@ -134,9 +136,9 @@ test('all existing non-page public URLs retain exact bytes outside the declared
const start = html.indexOf(SEO_BLOCK_START), end = html.indexOf(SEO_BLOCK_END) + SEO_BLOCK_END.length;
assert.ok(start > html.indexOf('<head>') && end < html.indexOf('</head>'), `${file}: metadata belongs only in head`);
const block = html.slice(start, end) + '\n';
- assert.equal(block, artifactMetadata(proof.entry, proof.presentation), `${file}: exact publication metadata`);
+ assert.equal(block, proof.metadata, `${file}: exact publication metadata`);
const metadataNodes = (parseFragment(block).childNodes || []).filter(node => node.tagName);
- assert.deepEqual(metadataNodes.map(node => node.tagName), ['link', 'meta', 'meta', 'meta', 'meta', 'meta', 'script']);
+ assert.deepEqual(metadataNodes.map(node => node.tagName), ['link', 'meta', 'meta', 'meta', 'meta', 'meta', ...(proof.isCase ? ['meta', 'meta', 'meta'] : []), 'script']);
const structured = metadataNodes.at(-1);
assert.equal(attr(structured, 'type'), 'application/ld+json', `${file}: no executable script permitted`);
assert.equal(structured.attrs.length, 1);
diff --git a/website/test/publish-seo.test.mjs b/website/test/publish-seo.test.mjs
index 805a2dd9..ee479049 100644
--- a/website/test/publish-seo.test.mjs
+++ b/website/test/publish-seo.test.mjs
@@ -4,14 +4,17 @@ import fs from 'node:fs';
import os from 'node:os';
import path from 'node:path';
import { fileURLToPath } from 'node:url';
+import { parse, parseFragment } from 'parse5';
import { CASES } from '../src/data/gallery-presentation.mjs';
+import { PUBLISHED_CASES } from '../src/data/published-cases.mjs';
import { canonicalPath, canonicalUrl, sitePages } from '../src/data/site-urls.mjs';
-import { publishSeo, selectedArtifacts, stripPublishedArtifactMetadata } from '../scripts/publish-seo.mjs';
+import { publishSeo, selectedArtifacts, caseMetadata, SEO_BLOCK_START, SEO_BLOCK_END, stripPublishedArtifactMetadata } from '../scripts/publish-seo.mjs';
const docs = fileURLToPath(new URL('../../docs/', import.meta.url));
function fixture() {
const root = fs.mkdtempSync(path.join(os.tmpdir(), 'archify-search-'));
fs.cpSync(path.join(docs, 'gallery'), path.join(root, 'gallery'), { recursive: true });
+ fs.cpSync(path.join(docs, 'cases'), path.join(root, 'cases'), { recursive: true });
for (const page of sitePages) for (const lang of ['en', 'zh']) {
const route = canonicalPath(page, lang).slice(1);
const file = !route || route.endsWith('/') ? `${route}index.html` : `${route}.html`;
@@ -25,7 +28,8 @@ for (const name of ['cloudflare', 'github']) test(`${name}: published discovery
const root = fixture();
try {
const urls = publishSeo(root, { name });
- assert.equal(urls.length, sitePages.length * 2 + CASES.length);
+ assert.equal(urls.length, sitePages.length * 2 + CASES.length + PUBLISHED_CASES.length);
+ assert.equal(urls.length, 31, '14 localized pages, 12 gallery artifacts and 5 standalone cases');
assert.equal(new Set(urls).size, urls.length);
for (const page of sitePages) for (const lang of ['en', 'zh']) assert.ok(urls.includes(canonicalUrl(page, lang)));
const sitemap = fs.readFileSync(path.join(root, 'sitemap.xml'), 'utf8');
@@ -42,15 +46,105 @@ for (const name of ['cloudflare', 'github']) test(`${name}: published discovery
assert.equal(structured.isBasedOn, `https://archify.si/${entry.input}`);
assert.deepEqual(fs.readFileSync(path.join(root, entry.input)), fs.readFileSync(path.join(docs, entry.input)));
}
+ for (const entry of PUBLISHED_CASES) {
+ const original = fs.readFileSync(path.join(docs, entry.artifact));
+ const published = fs.readFileSync(path.join(root, entry.artifact), 'utf8');
+ assert.deepEqual(Buffer.from(stripPublishedArtifactMetadata(published)), original, `${entry.artifact}: exact original bytes outside the metadata block`);
+ assert.deepEqual(fs.readFileSync(path.join(root, entry.input)), fs.readFileSync(path.join(docs, entry.input)), `${entry.input}: unchanged source bytes`);
+ const source = JSON.parse(fs.readFileSync(path.join(docs, entry.input), 'utf8'));
+ assert.equal(entry.title, source.meta.title);
+ assert.equal(entry.language, source.meta.locale || 'en');
+ assertCaseMetadata(published, entry);
+ assert.ok(urls.includes(`https://archify.si/${entry.artifact.replace(/\.html$/, '')}`));
+ }
+ const casesOnce = PUBLISHED_CASES.map(entry => fs.readFileSync(path.join(root, entry.artifact)));
const once = fs.readFileSync(path.join(root, `gallery/artifacts/${CASES[0].output}`), 'utf8');
publishSeo(root, { name });
assert.equal(fs.readFileSync(path.join(root, `gallery/artifacts/${CASES[0].output}`), 'utf8'), once, 'repeat build is idempotent');
+ for (const [index, entry] of PUBLISHED_CASES.entries()) assert.deepEqual(fs.readFileSync(path.join(root, entry.artifact)), casesOnce[index], `${entry.artifact}: repeat publication is idempotent`);
assert.equal(fs.readFileSync(path.join(root, 'robots.txt'), 'utf8'), 'User-agent: *\nAllow: /\n\nSitemap: https://archify.si/sitemap.xml\n');
assert.equal(fs.existsSync(path.join(root, '_headers')), name === 'cloudflare');
if (name === 'cloudflare') assert.equal(fs.readFileSync(path.join(root, '_headers'), 'utf8'), '/*.json\n X-Robots-Tag: noindex\n');
} finally { fs.rmSync(root, { recursive: true, force: true }); }
});
+const nodes = node => [node, ...(node.childNodes || []).flatMap(nodes)];
+const attr = (node, name) => node?.attrs?.find(item => item.name === name)?.value;
+const text = node => (node.childNodes || []).map(child => child.value || '').join('');
+
+function assertCaseMetadata(html, entry) {
+ const all = nodes(parse(html));
+ const head = nodes(all.find(node => node.tagName === 'head'));
+ const url = `https://archify.si/${entry.artifact.replace(/\.html$/, '')}`;
+ const canonical = all.filter(node => node.tagName === 'link' && attr(node, 'rel') === 'canonical');
+ assert.equal(canonical.length, 1);
+ assert.ok(head.includes(canonical[0]));
+ assert.equal(attr(canonical[0], 'href'), url);
+ assert.equal(attr(all.find(node => node.tagName === 'html'), 'lang'), entry.language);
+ const original = nodes(parse(fs.readFileSync(path.join(docs, entry.artifact), 'utf8')));
+ assert.equal(text(head.find(node => node.tagName === 'title')), text(original.find(node => node.tagName === 'title')), 'original document title is preserved');
+ assert.equal(html.split(SEO_BLOCK_START).length, 2);
+ assert.equal(html.split(SEO_BLOCK_END).length, 2);
+ const start = html.indexOf(SEO_BLOCK_START), end = html.indexOf(SEO_BLOCK_END) + SEO_BLOCK_END.length;
+ assert.ok(start > html.indexOf('<head>') && end < html.indexOf('</head>'));
+ const block = html.slice(start, end) + '\n';
+ assert.equal(block, caseMetadata(entry));
+ const metadata = nodes(parseFragment(block)).filter(node => node.tagName);
+ assert.deepEqual(metadata.map(node => node.tagName), ['link', 'meta', 'meta', 'meta', 'meta', 'meta', 'meta', 'meta', 'meta', 'script']);
+ assert.deepEqual(metadata.at(-1).attrs, [{ name: 'type', value: 'application/ld+json' }], 'the only inserted script is non-executable structured data');
+ const meta = key => metadata.find(node => attr(node, 'name') === key || attr(node, 'property') === key);
+ for (const key of ['description', 'og:description', 'twitter:description']) assert.equal(attr(meta(key), 'content'), entry.description);
+ for (const key of ['og:title', 'twitter:title']) assert.equal(attr(meta(key), 'content'), entry.title);
+ assert.equal(attr(meta('og:type'), 'content'), 'article');
+ assert.equal(attr(meta('og:url'), 'content'), url);
+ assert.equal(attr(meta('twitter:card'), 'content'), 'summary');
+ assert.deepEqual(JSON.parse(text(metadata.at(-1))), {
+ '@context': 'https://schema.org', '@type': 'CreativeWork', name: entry.title, description: entry.description,
+ url, isBasedOn: `https://archify.si/${entry.input}`, inLanguage: entry.language,
+ });
+ assert.doesNotMatch(block, /hreflang|og:image|twitter:image|archify-social-preview/, 'no invented translation or stale preview image');
+}
+
+test('case discovery copy keeps the source-backed and illustrative boundaries explicit', () => {
+ const byPath = file => PUBLISHED_CASES.find(entry => entry.artifact === file);
+ assert.match(byPath('cases/life/leave-plan.workflow.html').description, /示例计划.*不代表真实日历或公司的休假政策/);
+ assert.match(byPath('cases/life/renting.sequence.html').description, /金额和流程仅为示例/);
+ assert.match(byPath('cases/life/monthly-money.dataflow.html').description, /Amounts are illustrative/);
+ const mco = byPath('cases/mco-runtime.architecture.html');
+ const source = JSON.parse(fs.readFileSync(path.join(docs, mco.input), 'utf8'));
+ assert.ok(mco.description.includes(source.meta.repository.revision.slice(0, 7)));
+ assert.match(mco.description, /authored source-backed diagram, not an execution trace/);
+});
+
+test('case metadata escapes HTML and JSON while retaining literal replacement tokens', () => {
+ const root = fixture();
+ const entry = PUBLISHED_CASES[0];
+ const original = { title: entry.title, description: entry.description };
+ try {
+ entry.title = 'Price $& with $` and $\' and $$ <offer> "quoted"';
+ entry.description = 'Text </script><script>alert("not executable")</script> & <meta name="extra">';
+ publishSeo(root, { name: 'github' });
+ assertCaseMetadata(fs.readFileSync(path.join(root, entry.artifact), 'utf8'), entry);
+ } finally {
+ Object.assign(entry, original);
+ fs.rmSync(root, { recursive: true, force: true });
+ }
+});
+
+test('publication rejects every missing selected case HTML or JSON before writing metadata', () => {
+ const root = fixture();
+ try {
+ const gallery = path.join(root, `gallery/artifacts/${CASES[0].output}`);
+ const original = fs.readFileSync(gallery);
+ for (const entry of PUBLISHED_CASES) for (const file of [entry.artifact, entry.input]) {
+ fs.rmSync(path.join(root, file));
+ assert.throws(() => publishSeo(root, { name: 'cloudflare' }), /Missing published case/);
+ assert.deepEqual(fs.readFileSync(gallery), original, 'missing inventory must not partially publish gallery metadata');
+ fs.copyFileSync(path.join(docs, file), path.join(root, file));
+ }
+ } finally { fs.rmSync(root, { recursive: true, force: true }); }
+});
+
test('publication fails closed for missing localized pages or changed proof artifacts/sources', () => {
const root = fixture();
try {
diff --git a/website/test/seo-dollar-copy.test.mjs b/website/test/seo-dollar-copy.test.mjs
new file mode 100644
index 00000000..bae69e28
--- /dev/null
+++ b/website/test/seo-dollar-copy.test.mjs
@@ -0,0 +1,46 @@
+import assert from 'node:assert/strict';
+import fs from 'node:fs';
+import os from 'node:os';
+import path from 'node:path';
+import test from 'node:test';
+import { createHash } from 'node:crypto';
+import { publishSeo, stripPublishedArtifactMetadata } from '../scripts/publish-seo.mjs';
+import { CASES } from '../src/data/gallery-presentation.mjs';
+import { canonicalPath, sitePages } from '../src/data/site-urls.mjs';
+
+const digest = value => createHash('sha256').update(value).digest('hex');
+
+// Regression ported from #826, with the selected case fixture added and exact
+// original closing-tag bytes checked alongside literal replacement tokens.
+test('SEO insertion preserves replacement tokens in authored gallery copy', t => {
+ const root = fs.mkdtempSync(path.join(os.tmpdir(), 'archify-seo-dollar-'));
+ t.after(() => fs.rmSync(root, { recursive: true, force: true }));
+ fs.cpSync(new URL('../../docs/cases/', import.meta.url), path.join(root, 'cases'), { recursive: true });
+ for (const page of sitePages) for (const lang of ['en', 'zh']) {
+ const route = canonicalPath(page, lang).slice(1);
+ const file = path.join(root, route === '' ? 'index.html' : `${route}.html`);
+ fs.mkdirSync(path.dirname(file), { recursive: true });
+ fs.writeFileSync(file, '<html></html>');
+ }
+ const html = '<html><head></HEAD><body>Diagram</body></html>';
+ const entries = CASES.map(item => {
+ const artifact = `gallery/artifacts/${item.output}`, input = `gallery/sources/${item.input}`;
+ const source = '{}';
+ for (const [name, content] of [[artifact, html], [input, source]]) {
+ const file = path.join(root, name);
+ fs.mkdirSync(path.dirname(file), { recursive: true });
+ fs.writeFileSync(file, content);
+ }
+ return { id: item.id, artifact, input, checks: [{ ok: true }], artifactSha256: digest(html), sourceSha256: digest(source) };
+ });
+ fs.writeFileSync(path.join(root, 'gallery/manifest.json'), JSON.stringify({ entries }));
+ const original = CASES[0].titleEn;
+ t.after(() => { CASES[0].titleEn = original; });
+ CASES[0].titleEn = 'Price $& with $` and $\' and $$';
+ publishSeo(root, { name: 'github' });
+ const published = fs.readFileSync(path.join(root, entries[0].artifact), 'utf8');
+ const json = /<script type="application\/ld\+json">([\s\S]*?)<\/script>/.exec(published)[1];
+ assert.equal(JSON.parse(json).name, CASES[0].titleEn);
+ assert.equal((published.match(/<\/head>/gi) || []).length, 1);
+ assert.equal(stripPublishedArtifactMetadata(published), html);
+});