Janady13/thatstupidcomputer / packages/site/search.ts

1import { getPage, ROUTES, type SitePage, type RoutePath } from './content.ts'; 2import type { PublicSiteConfig } from './truth.ts'; 3import { ROBOTS_POLICIES, PUBLISHER_META } from '../discovery/registry.ts'; 4 5export interface PageMetadata { 6 title: string; 7 description: string; 8 canonical?: string; 9 robots: string; 10 openGraph: { type: 'website'; siteName: string; title: string; description: string; url?: string }; 11} 12export const NOT_FOUND = { 13 title: 'Page not found', 14 description: 'This page could not be found. Explore the services or return to the workshop.', 15}; 16export const absoluteUrl = (path: string, config: PublicSiteConfig): string => `${config.canonicalOrigin}${path}`; 17export function metadataForPage(page: SitePage | undefined, config: PublicSiteConfig): PageMetadata { 18 const title = `${page?.title ?? NOT_FOUND.title} | ${config.siteName}`; 19 const description = page?.description ?? NOT_FOUND.description; 20 const canonical = page ? absoluteUrl(page.path, config) : undefined; 21 const robots = config.deployment === 'production' && page 22 ? 'index, follow, max-image-preview:large' : 'noindex, nofollow'; 23 return { title, description, ...(canonical ? { canonical } : {}), robots, 24 openGraph: { type: 'website', siteName: config.siteName, title, description, ...(canonical ? { url: canonical } : {}) } }; 25} 26export function breadcrumbsForPage(page: SitePage): Array<{ label: string; path: RoutePath }> { 27 const output: Array<{ label: string; path: RoutePath }> = []; 28 const visited = new Set<RoutePath>(); 29 let current: SitePage | undefined = page; 30 while (current) { 31 if (visited.has(current.path)) throw new Error('Cycle in page hierarchy'); 32 visited.add(current.path); 33 output.unshift({ label: current.navLabel, path: current.path }); 34 current = current.parent ? getPage(current.parent) : undefined; 35 } 36 return output; 37} 38type GraphNode = Record<string, unknown> & { '@id': string; '@type': string }; 39export function buildGraph(page: SitePage | undefined, config: PublicSiteConfig): { '@context': string; '@graph': GraphNode[] } | undefined { 40 if (!page) return undefined; 41 const url = absoluteUrl(page.path, config); 42 const ref = (id: string) => ({ '@id': id }); 43 const website: GraphNode = { 44 '@type': 'WebSite', '@id': config.websiteId, url: `${config.canonicalOrigin}/`, name: config.siteName, inLanguage: 'en', 45 ...(config.publisher ? { publisher: ref(config.publisher.id) } : {}), 46 ...(config.brand ? { about: ref(config.brand.id) } : {}), 47 }; 48 const webpage: GraphNode = { 49 '@type': page.path === '/about/' ? 'AboutPage' : page.path === '/contact/' ? 'ContactPage' : page.path === '/services/' ? 'CollectionPage' : 'WebPage', 50 '@id': `${url}#webpage`, url, name: page.title, description: page.description, inLanguage: 'en', isPartOf: ref(config.websiteId), 51 ...(page.path !== '/' ? { breadcrumb: ref(`${url}#breadcrumbs`) } : {}), 52 }; 53 const graph: GraphNode[] = [website, webpage]; 54 const sections = page.sections.filter(section => section.id); 55 const ids = new Set<string>(); 56 for (const section of sections) { 57 const id = section.id!; 58 if (!/^[a-z][a-z0-9]*(?:-[a-z0-9]+)*$/.test(id) || ids.has(id) || ['website', 'webpage', 'breadcrumbs', 'service', 'main-content'].includes(id)) throw new Error('Invalid or duplicate public answer anchor'); 59 ids.add(id); 60 graph.push({ '@type': 'WebPageElement', '@id': `${url}#${id}`, url: `${url}#${id}`, 61 name: section.heading, text: [...(section.paragraphs ?? []), ...(section.items ?? [])].join('\n\n'), 62 inLanguage: 'en', cssSelector: `#${id}`, isPartOf: ref(webpage['@id']), 63 ...(section.sources?.length ? { citation: section.sources.map(source => source.href) } : {}), 64 }); 65 } 66 if (sections.length) webpage.hasPart = sections.map(section => ref(`${url}#${section.id}`)); 67 if (page.path !== '/') graph.push({ 68 '@type': 'BreadcrumbList', '@id': `${url}#breadcrumbs`, 69 itemListElement: breadcrumbsForPage(page).map((crumb, index) => ({ '@type': 'ListItem', position: index + 1, name: crumb.label, item: absoluteUrl(crumb.path, config) })), 70 }); 71 if (page.service) { 72 const serviceId = `${url}#service`; 73 webpage.mainEntity = ref(serviceId); 74 graph.push({ '@type': 'Service', '@id': serviceId, url, name: page.service.name, 75 serviceType: page.service.serviceType, description: page.intro, mainEntityOfPage: ref(webpage['@id']), 76 ...(config.provider ? { provider: ref(config.provider.id) } : {}), 77 ...(config.brand ? { brand: ref(config.brand.id) } : {}), 78 }); 79 } 80 const externalIds = [config.publisher?.id, config.provider?.id, config.brand?.id].filter(Boolean); 81 if (graph.some((node) => externalIds.includes(node['@id']))) throw new Error('Canonical identity collides with a site-owned document or service'); 82 return { '@context': 'https://schema.org', '@graph': graph }; 83} 84export function escapeHtml(value: string): string { 85 return value.replace(/&/g, '&amp;').replace(/</g, '&lt;').replace(/>/g, '&gt;').replace(/"/g, '&quot;').replace(/'/g, '&#39;'); 86} 87/** Safe inside a script element; JSON remains parseable after HTML-dangerous characters are escaped. */ 88export function safeJson(value: unknown): string { 89 return JSON.stringify(value).replace(/</g, '\\u003c').replace(/>/g, '\\u003e').replace(/&/g, '\\u0026').replace(/\u2028/g, '\\u2028').replace(/\u2029/g, '\\u2029'); 90} 91export function renderHead(page: SitePage | undefined, config: PublicSiteConfig): string { 92 const metadata = metadataForPage(page, config); 93 const attr = (kind: 'name' | 'property', key: string, value: string) => `<meta ${kind}="${key}" content="${escapeHtml(value)}">`; 94 const head = [ 95 `<title>${escapeHtml(metadata.title)}</title>`, 96 attr('name', 'description', metadata.description), attr('name', 'robots', metadata.robots), 97 ...PUBLISHER_META.map(({ name, content }) => attr('name', name, config.deployment === 'production' && page ? content : 'noindex, nofollow')), 98 ...(metadata.canonical ? [`<link rel="canonical" href="${escapeHtml(metadata.canonical)}">`] : []), 99 attr('property', 'og:type', metadata.openGraph.type), attr('property', 'og:site_name', config.siteName), 100 attr('property', 'og:title', metadata.title), attr('property', 'og:description', metadata.description), 101 ...(metadata.canonical ? [attr('property', 'og:url', metadata.canonical)] : []), 102 attr('name', 'twitter:card', config.socialImage ? 'summary_large_image' : 'summary'), 103 attr('name', 'twitter:title', metadata.title), attr('name', 'twitter:description', metadata.description), 104 ]; 105 if (config.socialImage) head.push( 106 attr('property', 'og:image', config.socialImage.url), attr('property', 'og:image:alt', config.socialImage.alt), 107 attr('property', 'og:image:width', String(config.socialImage.width)), attr('property', 'og:image:height', String(config.socialImage.height)), 108 attr('name', 'twitter:image', config.socialImage.url), attr('name', 'twitter:image:alt', config.socialImage.alt), 109 ); 110 if (config.verification?.google) head.push(attr('name', 'google-site-verification', config.verification.google)); 111 if (config.verification?.bing) head.push(attr('name', 'msvalidate.01', config.verification.bing)); 112 const graph = buildGraph(page, config); 113 if (graph) head.push(`<script id="site-structured-data" type="application/ld+json">${safeJson(graph)}</script>`); 114 return head.join('\n'); 115} 116export function sitemapXml(config: PublicSiteConfig): string { 117 const urls = config.deployment === 'production' ? ROUTES.map((page) => ` <url><loc>${escapeHtml(absoluteUrl(page.path, config))}</loc></url>`).join('\n') : ''; 118 return `<?xml version="1.0" encoding="UTF-8"?>\n<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">\n${urls}\n</urlset>\n`; 119} 120/** Public search access with explicit, independent deployment choices for broader model-use permissions. */ 121export function robotsTxt(config: PublicSiteConfig): string { 122 if (config.deployment === 'production' && config.crawlerPolicy.source !== 'deployment-config') throw new Error('Production robots requires an explicitly configured crawler policy'); 123 const utility = ['Disallow: /api/', 'Disallow: /private/', 'Disallow: /native/', 'Disallow: /__']; 124 const policyGroup = (agent: string, policy: 'allow' | 'deny') => [`User-agent: ${agent}`, ...(policy === 'allow' ? ['Allow: /', ...utility] : ['Disallow: /']), '']; 125 const groups = [ 126 '# Public search discovery; private resources also require server-side access controls.', 127 ...(config.deployment === 'preview' ? ['# Preview HTML and response headers carry noindex. Crawling stays open so the directive can be read.'] : []), 128 'User-agent: *', 'Allow: /', ...utility, '', 129 '# Provider modules own discovery, training and user-retrieval controls.', 130 '# Google-Extended governs specified Gemini training and grounding, not Google Search.', 131 '# User-triggered robots behavior varies; private resources require authorization.', 132 ...ROBOTS_POLICIES.flatMap(({agent,policy})=>policyGroup(agent, 133 policy==='openaiTraining' ? config.crawlerPolicy.openaiTraining ?? config.crawlerPolicy.training : 134 policy==='anthropicTraining' ? config.crawlerPolicy.anthropicTraining ?? config.crawlerPolicy.training : 135 policy==='mistralTraining' ? config.crawlerPolicy.mistralTraining ?? config.crawlerPolicy.training : 136 policy==='googleExtended' ? config.crawlerPolicy.googleExtended : policy)), 137 ]; 138 if (config.deployment === 'production') groups.push(`Sitemap: ${config.canonicalOrigin}/sitemap.xml`, ''); 139 return groups.join('\n'); 140} 141