{profile.name} — {t(\"role\")}

[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8,"_9":10},"actionData","errors","root",{"_308":309,"_310":311},"routes/_app.x.$handle.$name",{"_101":100,"_32":45,"_109":91,"_27":110,"_13":14,"_111":-5,"_11":112,"_113":114,"_115":116,"_117":118,"_119":120,"_20":-5,"_121":-5,"_122":-5,"_123":25,"_124":125,"_126":60,"_127":-5,"_128":-5,"_129":130},"routes/_app.x.$handle.$name.branch.$branch.version.$version.diff",{"_11":12,"_13":14,"_15":16,"_17":-5,"_18":-5},"project",{"_43":44,"_32":45,"_38":46,"_47":-5,"_48":25,"_49":50,"_51":50,"_52":53,"_54":55,"_18":56,"_57":58,"_59":60,"_61":-5,"_62":-5,"_63":64,"_65":66,"_67":68,"_69":70,"_71":72,"_73":-5,"_74":75,"_76":75,"_77":78,"_79":78,"_80":78,"_81":75,"_82":75,"_83":84,"_85":86,"_87":88,"_89":25},"version",30,"changes",[19],"aiChatInfo","userId",{"_20":21,"_22":23,"_24":25,"_26":25,"_27":28,"_29":28},"file",{"_30":31,"_32":33,"_34":20,"_35":37,"_38":39,"_40":14,"_41":42,"_24":25},"previous",{"_30":31,"_32":33,"_34":20,"_35":36},"deleted",false,"deletedHere","path","posts/football-stats-pipeline.md","latestPath","fileId","60514658-1f07-11f1-8a80-42dde27851f2","name","football-stats-pipeline.md","fileType","content","---\nslug: football-stats-pipeline\ndate: Dec 5, 2025\nreadTime_en: 10 min read\nreadTime_pt: 10 min leitura\ntitle_en: Building a real-time football stats pipeline with AI enrichment\ntitle_pt: Pipeline de estatísticas de futebol em tempo real com IA\nexcerpt_en: How I wired fbref data → scheduled workers → GPT analysis → live API in a weekend, entirely on serverless infrastructure.\nexcerpt_pt: Como liguei dados do fbref → workers agendados → análise GPT → API em direto num fim de semana, tudo em infraestrutura serverless.\ntags: AI, TypeScript, Data Engineering\n---\n\n## The goal\n\nI wanted a personal API that could answer questions like: *\"Which midfielders in the Premier League have the best progressive pass ratio in the last 5 games?\"* — with fresh data, not a CSV from 2022.\n\nThe plan: scrape fbref → process and store → enrich with AI → serve via REST. All on Val.town, zero infrastructure to manage.\n\n## Architecture\n\n\\`\\`\\`\n[fbref scraper cron]\n ↓\n[raw events store (Val blob)]\n ↓\n[processing worker] ← runs every 15 min\n ↓\n[AI enrichment] ← GPT-4o summaries + tags\n ↓\n[HTTP API val] ← public REST endpoints\n\\`\\`\\`\n\n## The scraper\n\nfbref doesn't have an official API, so this involved parsing their HTML tables. The tricky part is rate limiting — scrape too fast and you get blocked. I settled on one request per 3 seconds with randomised jitter:\n\n\\`\\`\\`ts\nasync function fetchWithDelay(url: string, minMs = 2000, jitter = 1000) {\n const delay = minMs + Math.random() * jitter;\n await new Promise(r => setTimeout(r, delay));\n return fetch(url, { headers: { \"User-Agent\": \"personal-stats-bot/1.0\" } });\n}\n\\`\\`\\`\n\n## AI enrichment\n\nAfter storing raw stats, a separate worker calls the OpenAI API to generate:\n- A plain-English summary of a player's recent form\n- Automatic tags (e.g. \"in-form\", \"injury-return\", \"high-press-specialist\")\n- A form score (0–100)\n\nThe prompt is simple but the output is surprisingly useful for filtering.\n\n## Result\n\nThe API now serves ~50 endpoints with sub-100ms response times (mostly from Val blob cache). Total cost: ~$2/month in OpenAI credits. Total infrastructure managed: zero.\n\nThe code lives on Val.town at [val.town/u/nmsilva](https://www.val.town/u/nmsilva) — some vals are public if you want to poke around.\n","---\nslug: football-stats-pipeline\ndate: Dec 5, 2025\nreadTime_en: 10 min read\nreadTime_pt: 10 min leitura\ntitle_en: Building a real-time football stats pipeline with AI enrichment\ntitle_pt: Pipeline de estatísticas de futebol em tempo real com IA\nexcerpt_en: How I wired fbref data → scheduled workers → GPT analysis → live API in a weekend, entirely on serverless infrastructure.\nexcerpt_pt: Como liguei dados do fbref → workers agendados → análise GPT → API em direto num fim de semana, tudo em infraestrutura serverless.\ntags: AI, TypeScript, Data Engineering\n---\n\n## The goal\n\nI wanted a personal API that could answer questions like: *\"Which midfielders in the Premier League have the best progressive pass ratio in the last 5 games?\"* — with fresh data, not a CSV from 2022.\n\nThe plan: scrape fbref → process and store → enrich with AI → serve via REST. All on Val.town, zero infrastructure to manage.\n\n## Architecture\n\n```\n[fbref scraper cron]\n ↓\n[raw events store (Val blob)]\n ↓\n[processing worker] ← runs every 15 min\n ↓\n[AI enrichment] ← GPT-4o summaries + tags\n ↓\n[HTTP API val] ← public REST endpoints\n```\n\n## The scraper\n\nfbref doesn't have an official API, so this involved parsing their HTML tables. The tricky part is rate limiting — scrape too fast and you get blocked. I settled on one request per 3 seconds with randomised jitter:\n\n```ts\nasync function fetchWithDelay(url: string, minMs = 2000, jitter = 1000) {\n const delay = minMs + Math.random() * jitter;\n await new Promise(r => setTimeout(r, delay));\n return fetch(url, { headers: { \"User-Agent\": \"personal-stats-bot/1.0\" } });\n}\n```\n\n## AI enrichment\n\nAfter storing raw stats, a separate worker calls the OpenAI API to generate:\n- A plain-English summary of a player's recent form\n- Automatic tags (e.g. \"in-form\", \"injury-return\", \"high-press-specialist\")\n- A form score (0–100)\n\nThe prompt is simple but the output is surprisingly useful for filtering.\n\n## Result\n\nThe API now serves ~50 endpoints with sub-100ms response times (mostly from Val blob cache). Total cost: ~$2/month in OpenAI credits. Total infrastructure managed: zero.\n\nThe code lives on Val.town at [val.town/u/nmsilva](https://www.val.town/u/nmsilva) — some vals are public if you want to poke around.\n\n---pt---\n\n## O objectivo\n\nQueria uma API pessoal que respondesse a perguntas como: *\"Quais médios da Premier League têm o melhor rácio de passes progressivos nos últimos 5 jogos?\"* — com dados frescos, não um CSV de 2022.\n\nO plano: fazer scraping do fbref → processar e armazenar → enriquecer com IA → servir via REST. Tudo no Val.town, zero infraestrutura a gerir.\n\n## Arquitectura\n\n```\n[cron scraper fbref]\n ↓\n[armazenamento de eventos raw (Val blob)]\n ↓\n[worker de processamento] ← corre a cada 15 min\n ↓\n[enriquecimento com IA] ← resumos GPT-4o + tags\n ↓\n[HTTP API val] ← endpoints REST públicos\n```\n\n## O scraper\n\nO fbref não tem uma API oficial, por isso isto envolvia fazer parse das tabelas HTML. A parte complicada é o rate limiting — scraping demasiado rápido e ficas bloqueado. Cheguei a um pedido por cada 3 segundos com jitter aleatório:\n\n```ts\nasync function fetchWithDelay(url: string, minMs = 2000, jitter = 1000) {\n const delay = minMs + Math.random() * jitter;\n await new Promise(r => setTimeout(r, delay));\n return fetch(url, { headers: { \"User-Agent\": \"personal-stats-bot/1.0\" } });\n}\n```\n\n## Enriquecimento com IA\n\nApós armazenar estatísticas raw, um worker separado chama a API da OpenAI para gerar:\n- Um resumo em linguagem simples da forma recente de um jogador\n- Tags automáticas (ex: \"em forma\", \"regresso de lesão\", \"especialista em pressing\")\n- Uma pontuação de forma (0–100)\n\nO prompt é simples mas o output é surpreendentemente útil para filtrar.\n\n## Resultado\n\nA API serve agora ~50 endpoints com tempos de resposta abaixo de 100ms (principalmente cache do Val blob). Custo total: ~$2/mês em créditos OpenAI. Infraestrutura gerida: zero.\n\nO código está no Val.town em [val.town/u/nmsilva](https://www.val.town/u/nmsilva) — alguns vals são públicos se quiseres explorar.\n","createdAt",["D",1773425984502],"branchVersion","branchId","5680e34e-1d41-11f1-9a2e-42dde27851f2","id","5679b0d8-1d41-11f1-9a2e-42dde27851f2","my-portfolio",["D",1773230188571],"deletedAt","legacyEnvVars","privacy","public","httpPrivacy","description","Starter template for a markdown blog","imageUrl","https://imagedelivery.net/iHX6Ovru0O7AjmyT5yZRoA/44a31e5c-3e8a-48a8-d067-5785c0b08200/square512?exp=1789534799&sig=e3425e6cfc6d61e4bb656977644d074e6da3176e4285f9c272871f43226a9d62","1e4d2157-cdd6-42c8-a0ed-7d707498307a","pinnedFileId","019cdcc1-687b-73e8-a526-d410d2c5adc7","pinPreview",true,"pinMigrated","screenshotVersion","likeCount",0,"orgId","eec31d76-5f30-4859-9ad3-ead49204cdd9","tursoName","5679b0d81d4111f19a2e42dde27851f2","tursoHostname","5679b0d81d4111f19a2e42dde27851f2-val-town.aws-us-east-2.turso.io","tursoMeasuredAt",["D",1789521041830],"pinned","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm","user",{"_101":100,"_108":104},"org",{"_101":100,"_102":-5,"_103":104,"_18":56,"_105":106,"_107":25},"branch",{"_13":90,"_32":91,"_43":42,"_92":44,"_93":-5,"_94":-5,"_95":-5,"_96":-5,"_97":-5,"_98":45,"_99":100},"isOwned",47,"main","projectId","forkedBranchId","forkedBranchVersion","forkedProjectPrivacy","forkedProjectId","forkedProjectOrgId","projectName","userHandle","nmsilva","handle","displayName","avatar","https://img.clerk.com/eyJ0eXBlIjoicHJveHkiLCJzcmMiOiJodHRwczovL2ltYWdlcy5jbGVyay5kZXYvb2F1dGhfZ2l0aHViL2ltZ18yZlNYVHdvT3FqSExKd0Y5Qmx5WllLWVR6UjMifQ","tier","free","canRestrictVals","profileImageUrl","branchName","","adoptOrgId",{"_43":44,"_32":45,"_38":303,"_47":-5,"_48":25,"_49":50,"_51":50,"_52":53,"_54":55,"_18":56,"_57":58,"_59":60,"_61":-5,"_62":-5,"_63":64,"_65":66,"_67":68,"_69":70,"_71":304,"_73":-5,"_74":75,"_76":75,"_77":78,"_79":78,"_80":78,"_81":75,"_82":75,"_83":305,"_85":306,"_87":307,"_89":25},"tree",[182,189,196,197,198],"files",[182,183,184,185,186,187,188,189,190,191,192,193,194,195,196,197,198],"counts",{"_172":64,"_173":150,"_174":64,"_175":176,"_177":178},"contents",[131,132,133,134,135],"upstreamProject","upstreamBranch","emptyEnvVars","flags",[],"hasHttpFiles","grants","orgs","customDomains",{},{"_43":167,"_32":168,"_30":169,"_138":-5,"_41":42,"_40":142,"_140":60,"_38":170,"_13":64,"_34":165,"_35":-5,"_24":25,"_145":-5,"_27":171},{"_43":161,"_32":162,"_30":163,"_138":-5,"_41":42,"_40":150,"_140":60,"_38":164,"_13":64,"_34":165,"_35":-5,"_24":25,"_145":-5,"_27":166},{"_43":154,"_32":155,"_30":156,"_138":-5,"_41":42,"_40":150,"_140":60,"_38":157,"_13":64,"_34":158,"_35":159,"_24":25,"_145":-5,"_27":160},{"_43":147,"_32":148,"_30":149,"_138":-5,"_41":42,"_40":150,"_140":60,"_38":151,"_13":64,"_34":20,"_35":152,"_24":25,"_145":-5,"_27":153},{"_43":136,"_32":137,"_30":58,"_138":-5,"_41":42,"_40":139,"_140":25,"_38":141,"_13":142,"_34":143,"_35":144,"_24":25,"_145":-5,"_27":146},"74bf29e8-1f07-11f1-97c1-42dde27851f2","index.tsx","parentId",26,"latest",["D",1773425230902],12,"http","/** @jsxImportSource https://esm.sh/react@18.2.0 */\nimport { renderToString } from \"https://esm.sh/react-dom@18.2.0/server\";\nimport rosetta from \"https://esm.sh/rosetta@1.1.0\";\n\n// Load post markdown files at startup (cached per cold start)\nconst BASE_URL = \"https://esm.town/v/nmsilva/my-portfolio/posts\";\nconst [ormVsRawSqlMd, valtownScratchpadMd, reactServerComponentsMd, footballStatsPipelineMd] =\n await Promise.all([\n fetch(`${BASE_URL}/orm-vs-raw-sql.md`).then((r) => r.text()),\n fetch(`${BASE_URL}/valtown-scratchpad.md`).then((r) => r.text()),\n fetch(`${BASE_URL}/react-server-components-one-year-in.md`).then((r) => r.text()),\n fetch(`${BASE_URL}/football-stats-pipeline.md`).then((r) => r.text()),\n ]);\n\n// ─── I18N ────────────────────────────────────────────────────────────────────\n\nconst i18n = rosetta({\n en: {\n // Nav\n about: \"About\",\n projects: \"Projects\",\n writing: \"Writing\",\n skills: \"Skills\",\n contact: \"Contact\",\n\n // Sidebar\n role: \"Full-Stack Developer\",\n bio: \"I build robust backends and polished UIs. Focused on Node.js and React — I like systems that are fast, clean, and maintainable.\",\n\n // Section labels\n label_about: \"About\",\n label_work: \"Work\",\n label_blog: \"Blog\",\n label_toolbox: \"Toolbox\",\n label_contact: \"Get in touch\",\n\n // Intro paragraphs\n intro_1: \"I'm a full-stack developer based in Alcofra, Portugal, currently at Dragonboat where I build product roadmap tooling for engineering and PM teams worldwide.\",\n intro_2: \"My stack is Node.js + TypeScript on the backend and React on the frontend. Outside of work I tinker with data pipelines, sports analytics, and automation — mostly on Val.town.\",\n\n // Contact section\n contact_title: \"Contact\",\n contact_email_label: \"Email\",\n contact_github_label: \"GitHub\",\n contact_valtown_label: \"Val.town\",\n\n // Post page\n back: \"← Back to home\",\n read_article: \"Read article →\",\n\n // Footer\n built_with: \"Built with Val.town\",\n },\n pt: {\n // Nav\n about: \"Sobre\",\n projects: \"Projetos\",\n writing: \"Artigos\",\n skills: \"Competências\",\n contact: \"Contacto\",\n\n // Sidebar\n role: \"Programador Full-Stack\",\n bio: \"Construo backends robustos e interfaces cuidadas. Focado em Node.js e React — gosto de sistemas que sejam rápidos, limpos e fáceis de manter.\",\n\n // Section labels\n label_about: \"Sobre\",\n label_work: \"Trabalho\",\n label_blog: \"Blog\",\n label_toolbox: \"Ferramentas\",\n label_contact: \"Fala comigo\",\n\n // Intro paragraphs\n intro_1: \"Sou um programador full-stack de Alcofra, Portugal, atualmente na Dragonboat onde construo ferramentas de roadmap de produto para equipas de engenharia e produto em todo o mundo.\",\n intro_2: \"O meu stack é Node.js + TypeScript no backend e React no frontend. Fora do trabalho, experimento pipelines de dados, análise desportiva e automação — principalmente no Val.town.\",\n\n // Contact section\n contact_title: \"Contacto\",\n contact_email_label: \"Email\",\n contact_github_label: \"GitHub\",\n contact_valtown_label: \"Val.town\",\n\n // Post page\n back: \"← Voltar ao início\",\n read_article: \"Ler artigo →\",\n\n // Footer\n built_with: \"Feito com Val.town\",\n },\n});\n\n// ─── TRANSLATED DATA ─────────────────────────────────────────────────────────\n\nconst projects = {\n en: [\n {\n name: \"Dragonboat — Product Roadmap Platform\",\n desc: \"Building scalable backend services and React UIs for a leading product management platform used by thousands of product teams worldwide.\",\n tags: [\"Node.js\", \"React\", \"TypeScript\", \"PostgreSQL\"],\n link: \"https://dragonboat.io\",\n },\n {\n name: \"fbref Football Data Pipeline\",\n desc: \"Automated data ingestion and AI-enriched processing pipeline for football statistics — events, scheduled workers, and live scoring.\",\n tags: [\"TypeScript\", \"Val.town\", \"Cron\", \"AI\"],\n link: \"#\",\n },\n {\n name: \"Crypto & Real Estate Price Trackers\",\n desc: \"Lightweight scheduled monitors for crypto prices and Portuguese real estate listings, with change alerts.\",\n tags: [\"Node.js\", \"Val.town\", \"REST APIs\"],\n link: \"#\",\n },\n ],\n pt: [\n {\n name: \"Dragonboat — Plataforma de Roadmap\",\n desc: \"Construção de serviços backend escaláveis e UIs React para uma plataforma líder de gestão de produto utilizada por milhares de equipas em todo o mundo.\",\n tags: [\"Node.js\", \"React\", \"TypeScript\", \"PostgreSQL\"],\n link: \"https://dragonboat.io\",\n },\n {\n name: \"Pipeline de Dados de Futebol (fbref)\",\n desc: \"Pipeline automatizado de ingestão de dados e processamento enriquecido com IA para estatísticas de futebol — eventos, workers e resultados em direto.\",\n tags: [\"TypeScript\", \"Val.town\", \"Cron\", \"AI\"],\n link: \"#\",\n },\n {\n name: \"Monitores de Preços de Cripto e Imobiliário\",\n desc: \"Monitores agendados para preços de criptomoedas e listagens de imóveis em Portugal, com alertas de variação.\",\n tags: [\"Node.js\", \"Val.town\", \"REST APIs\"],\n link: \"#\",\n },\n ],\n};\n\n// Posts keep their content in English only (technical writing is usually in EN)\n// but titles/excerpts are translated\n\n// ─── POST LOADER ─────────────────────────────────────────────────────────────\n\nfunction parseFrontmatter(raw: string): { data: Record; content: string } {\n const match = raw.match(/^---\\n([\\s\\S]*?)\\n---\\n([\\s\\S]*)$/);\n if (!match) return { data: {}, content: raw };\n const data: Record = {};\n for (const line of match[1].split(\"\\n\")) {\n const i = line.indexOf(\":\");\n if (i === -1) continue;\n data[line.slice(0, i).trim()] = line.slice(i + 1).trim();\n }\n return { data, content: match[2].trim() };\n}\n\nfunction parsePost(raw: string) {\n const { data, content } = parseFrontmatter(raw);\n return {\n slug: data.slug,\n date: data.date,\n readTime: { en: data.readTime_en, pt: data.readTime_pt },\n title: { en: data.title_en, pt: data.title_pt },\n excerpt: { en: data.excerpt_en, pt: data.excerpt_pt },\n tags: data.tags.split(\",\").map((t: string) => t.trim()),\n content,\n };\n}\n\nconst posts = [\n ormVsRawSqlMd,\n valtownScratchpadMd,\n reactServerComponentsMd,\n footballStatsPipelineMd,\n].map(parsePost);\n\n\nconst skills = [\n { icon: \"🟩\", name: \"Node.js\" },\n { icon: \"⚛️\", name: \"React\" },\n { icon: \"🟦\", name: \"TypeScript\" },\n { icon: \"🐘\", name: \"PostgreSQL\" },\n { icon: \"🍃\", name: \"MongoDB\" },\n { icon: \"🐳\", name: \"Docker\" },\n { icon: \"☁️\", name: \"AWS\" },\n { icon: \"🔀\", name: \"Git\" },\n { icon: \"🚀\", name: \"REST APIs\" },\n { icon: \"🔗\", name: \"GraphQL\" },\n { icon: \"⚡\", name: \"Redis\" },\n { icon: \"🧪\", name: \"Jest\" },\n];\n\nconst profile = {\n name: \"Nuno Silva\",\n initials: \"NS\",\n location: { en: \"Alcofra, Portugal\", pt: \"Alcofra, Portugal\" },\n email: \"nuno4slb@gmail.com\",\n github: \"https://github.com/nmsilva\",\n valtown: \"https://www.val.town/u/nmsilva\",\n};\n\n// ─── MARKDOWN-LITE RENDERER ───────────────────────────────────────────────────\n\nfunction renderMarkdown(md: string): string {\n return md\n .trim()\n .replace(/```(\\w*)\\n([\\s\\S]*?)```/g, (_, _lang, code) =>\n `

${code.replace(//g, \">\")}

`\n )\n .replace(/^## (.+)$/gm, \"

$1

\")\n .replace(/\\*\\*(.+?)\\*\\*/g, \"$1\")\n .replace(/\\[(.+?)\\]\$(.+?)\$/g, '$1')\n .replace(/`([^`]+)`/g, \"$1\")\n .split(\"\\n\\n\")\n .map((block) =>\n block.startsWith(\"<\") ? block : block.trim() ? `

${block}

` : \"\"\n )\n .join(\"\\n\");\n}\n\n// ─── STYLES ──────────────────────────────────────────────────────────────────\n\nconst css = `\n *, *::before, *::after { box-sizing: border-box; margin: 0; padding: 0; }\n html { font-family: 'Inter', system-ui, -apple-system, BlinkMacSystemFont, sans-serif; font-size: 16px; }\n body { background-color: #fafafa; color: #18181b; -webkit-font-smoothing: antialiased; line-height: 1.6; }\n\n .glow {\n position: fixed; top: -20rem; left: 50%; transform: translateX(-50%);\n width: 80rem; height: 40rem; border-radius: 9999px; pointer-events: none; z-index: -1;\n background: radial-gradient(ellipse at center, rgba(99,102,241,0.07) 0%, rgba(139,92,246,0.04) 40%, transparent 70%);\n }\n\n .page { max-width: 72rem; margin: 0 auto; padding: 5rem 1.5rem; display: flex; gap: 0; align-items: flex-start; }\n\n /* ── Sidebar ── */\n .sidebar { width: 22rem; flex-shrink: 0; position: sticky; top: 4rem; padding-right: 4rem; }\n .avatar {\n width: 4.5rem; height: 4.5rem; border-radius: 9999px;\n background: linear-gradient(135deg, #e4e4e7 0%, #a1a1aa 100%);\n display: flex; align-items: center; justify-content: center;\n font-size: 1.5rem; font-weight: 700; color: #52525b;\n margin-bottom: 1.25rem; border: 2px solid #e4e4e7;\n }\n .sidebar-name { font-size: 1.5rem; font-weight: 700; color: #18181b; letter-spacing: -0.03em; line-height: 1.2; margin-bottom: 0.25rem; }\n .sidebar-role { font-size: 0.875rem; font-weight: 500; color: #71717a; margin-bottom: 1rem; }\n .sidebar-bio { font-size: 0.875rem; color: #52525b; line-height: 1.65; margin-bottom: 1.75rem; }\n\n .sidebar-nav { list-style: none; display: flex; flex-direction: column; gap: 0.125rem; margin-bottom: 1.75rem; }\n .sidebar-nav a {\n display: flex; align-items: center; gap: 0.625rem; padding: 0.375rem 0.75rem;\n border-radius: 0.5rem; text-decoration: none; font-size: 0.875rem; font-weight: 500;\n color: #52525b; transition: background 0.15s, color 0.15s;\n }\n .sidebar-nav a:hover { background: #f4f4f5; color: #18181b; }\n .sidebar-nav a .nav-dot { width: 0.375rem; height: 0.375rem; border-radius: 9999px; background: #d4d4d8; flex-shrink: 0; transition: background 0.15s; }\n .sidebar-nav a:hover .nav-dot { background: #6366f1; }\n\n .sidebar-socials { display: flex; gap: 0.875rem; align-items: center; margin-bottom: 1.25rem; }\n .sidebar-socials a { color: #a1a1aa; text-decoration: none; font-size: 0.8rem; font-weight: 500; transition: color 0.15s; }\n .sidebar-socials a:hover { color: #18181b; }\n\n /* ── Lang toggle ── */\n .lang-toggle { display: flex; align-items: center; gap: 0.25rem; }\n .lang-btn {\n padding: 0.2rem 0.5rem; border-radius: 0.375rem; font-size: 0.75rem; font-weight: 600;\n text-decoration: none; color: #a1a1aa; transition: background 0.15s, color 0.15s; letter-spacing: 0.03em;\n }\n .lang-btn:hover { color: #52525b; }\n .lang-btn.active { background: #f4f4f5; color: #18181b; }\n .lang-sep { color: #d4d4d8; font-size: 0.75rem; }\n\n /* ── Main ── */\n .main { flex: 1; min-width: 0; }\n .mobile-header { display: none; }\n\n /* ── Sections ── */\n .section { margin-bottom: 3.5rem; }\n .section-label { font-size: 0.6875rem; font-weight: 600; letter-spacing: 0.1em; text-transform: uppercase; color: #a1a1aa; margin-bottom: 1.25rem; }\n .section-title { font-size: 1.25rem; font-weight: 700; color: #18181b; letter-spacing: -0.025em; margin-bottom: 1.25rem; }\n hr { border: none; border-top: 1px solid #f4f4f5; margin: 2.5rem 0; }\n\n .intro-text { font-size: 1.0625rem; color: #3f3f46; line-height: 1.75; margin-bottom: 1.25rem; }\n .intro-text strong { color: #18181b; font-weight: 600; }\n .intro-text a { color: #18181b; font-weight: 600; text-decoration: underline; text-decoration-color: #d4d4d8; }\n .intro-text a:hover { text-decoration-color: #6366f1; }\n .intro-text:last-child { margin-bottom: 0; }\n\n /* ── Projects ── */\n .project-card { padding: 1.25rem 0; border-bottom: 1px solid #f4f4f5; display: flex; gap: 1.25rem; align-items: flex-start; }\n .project-card:last-child { border-bottom: none; }\n .project-icon { width: 2.5rem; height: 2.5rem; border-radius: 0.5rem; background: linear-gradient(135deg, #f4f4f5, #e4e4e7); border: 1px solid #e4e4e7; display: flex; align-items: center; justify-content: center; font-size: 1rem; flex-shrink: 0; }\n .project-info { flex: 1; min-width: 0; }\n .project-title { font-size: 0.9375rem; font-weight: 600; color: #18181b; text-decoration: none; line-height: 1.4; display: flex; align-items: center; gap: 0.375rem; }\n .project-title:hover { color: #6366f1; }\n .project-title .arrow { font-size: 0.75rem; color: #a1a1aa; }\n .project-desc { font-size: 0.875rem; color: #71717a; line-height: 1.55; margin-top: 0.25rem; margin-bottom: 0.75rem; }\n\n /* ── Tags ── */\n .tags { display: flex; flex-wrap: wrap; gap: 0.375rem; }\n .tag { display: inline-flex; align-items: center; padding: 0.125rem 0.5rem; border-radius: 9999px; background: #f4f4f5; color: #71717a; font-size: 0.6875rem; font-weight: 500; border: 1px solid #e4e4e7; letter-spacing: 0.01em; }\n\n /* ── Post cards ── */\n .post-card { padding: 1.5rem; border: 1px solid #f4f4f5; border-radius: 0.75rem; background: white; margin-bottom: 1rem; transition: border-color 0.15s, box-shadow 0.15s; }\n .post-card:hover { border-color: #e4e4e7; box-shadow: 0 4px 12px -2px rgba(0,0,0,0.05); }\n .post-meta { display: flex; align-items: center; gap: 0.625rem; font-size: 0.75rem; color: #a1a1aa; margin-bottom: 0.5rem; }\n .post-meta .sep { color: #d4d4d8; }\n .post-title { font-size: 0.9375rem; font-weight: 600; color: #18181b; text-decoration: none; line-height: 1.45; display: block; margin-bottom: 0.375rem; }\n .post-title:hover { color: #6366f1; }\n .post-excerpt { font-size: 0.875rem; color: #71717a; line-height: 1.6; margin-bottom: 0.75rem; }\n .read-link { font-size: 0.8125rem; font-weight: 500; color: #52525b; text-decoration: none; display: inline-flex; align-items: center; gap: 0.25rem; margin-top: 0.625rem; }\n .read-link:hover { color: #6366f1; }\n\n /* ── Skills ── */\n .skills-wrap { display: flex; flex-wrap: wrap; gap: 0.5rem; }\n .skill-chip { display: inline-flex; align-items: center; gap: 0.375rem; padding: 0.375rem 0.75rem; border-radius: 9999px; background: white; border: 1px solid #e4e4e7; font-size: 0.8125rem; font-weight: 500; color: #3f3f46; transition: border-color 0.15s, color 0.15s; }\n .skill-chip:hover { border-color: #a1a1aa; color: #18181b; }\n .skill-icon { font-size: 0.875rem; }\n\n /* ── Contact ── */\n .contact-row { display: flex; flex-direction: column; gap: 0.75rem; }\n .contact-item { display: flex; align-items: center; gap: 0.75rem; font-size: 0.875rem; color: #52525b; text-decoration: none; padding: 0.75rem 1rem; border: 1px solid #f4f4f5; border-radius: 0.625rem; background: white; transition: border-color 0.15s, color 0.15s; }\n .contact-item:hover { border-color: #d4d4d8; color: #18181b; }\n .contact-item .icon { font-size: 1rem; flex-shrink: 0; }\n .contact-detail { font-weight: 600; color: #18181b; font-size: 0.875rem; }\n .contact-sub { color: #a1a1aa; font-size: 0.8125rem; }\n\n /* ── Footer ── */\n footer { border-top: 1px solid #f4f4f5; padding-top: 1.5rem; display: flex; justify-content: space-between; align-items: center; flex-wrap: wrap; gap: 0.75rem; margin-top: 1rem; }\n footer p { font-size: 0.75rem; color: #a1a1aa; }\n .foot-links { display: flex; gap: 1.25rem; }\n .foot-links a { font-size: 0.75rem; color: #a1a1aa; text-decoration: none; }\n .foot-links a:hover { color: #71717a; }\n\n /* ── Post page ── */\n .post-page { max-width: 44rem; margin: 0 auto; padding: 3.5rem 1.5rem 5rem; }\n .back-link { display: inline-flex; align-items: center; gap: 0.375rem; font-size: 0.875rem; font-weight: 500; color: #71717a; text-decoration: none; margin-bottom: 2.5rem; transition: color 0.15s; }\n .back-link:hover { color: #18181b; }\n .post-header { margin-bottom: 2.5rem; }\n .post-header h1 { font-size: clamp(1.5rem, 3.5vw, 2.125rem); font-weight: 700; color: #18181b; letter-spacing: -0.03em; line-height: 1.2; margin-bottom: 0.75rem; }\n .post-body { color: #3f3f46; line-height: 1.75; }\n .post-body h2 { font-size: 1.1875rem; font-weight: 700; color: #18181b; margin: 2.25rem 0 0.875rem; letter-spacing: -0.025em; }\n .post-body p { margin-bottom: 1.25rem; }\n .post-body a { color: #6366f1; }\n .post-body strong { color: #18181b; font-weight: 600; }\n .post-body code { font-family: 'ui-monospace', 'SFMono-Regular', monospace; font-size: 0.875em; background: #f4f4f5; padding: 0.125rem 0.375rem; border-radius: 0.25rem; color: #52525b; }\n .post-body pre { background: #18181b; border-radius: 0.625rem; padding: 1.25rem; margin: 1.5rem 0; overflow-x: auto; }\n .post-body pre code { background: none; padding: 0; color: #d4d4d8; font-size: 0.8125rem; }\n .post-body ul { padding-left: 1.375rem; margin-bottom: 1.25rem; }\n .post-body ul li { margin-bottom: 0.375rem; }\n\n /* ── Responsive ── */\n @media (max-width: 1024px) {\n .sidebar { display: none; }\n .mobile-header { display: block; margin-bottom: 2.5rem; }\n .page { padding: 2.5rem 1.25rem; }\n }\n @media (max-width: 640px) {\n .page { padding: 1.75rem 1rem; }\n .post-page { padding: 2rem 1rem 3.5rem; }\n }\n`;\n\n// ─── COMPONENTS ──────────────────────────────────────────────────────────────\n\nfunction LangToggle({ lang, basePath }: { lang: string; basePath: string }) {\n return (\n

\n \n EN\n \n /\n \n PT\n \n

\n );\n}\n\nfunction Sidebar({ lang, t, basePath }: { lang: string; t: (k: string) => string; basePath: string }) {\n return (\n \n );\n}\n\nfunction MobileHeader({ lang, t, basePath }: { lang: string; t: (k: string) => string; basePath: string }) {\n return (\n

\n {profile.initials}\n

{profile.name}

{t(\"role\")}

\n \n

{t(\"bio\")}

\n {[\n { href: \"#intro\", key: \"about\" },\n { href: \"#projects\", key: \"projects\" },\n { href: \"#writing\", key: \"writing\" },\n { href: \"#skills\", key: \"skills\" },\n { href: \"#contact\", key: \"contact\" },\n ].map(({ href, key }) => (\n {t(key)}\n ))}\n

\n );\n}\n\n// ─── HOME PAGE ────────────────────────────────────────────────────────────────\n\nfunction HomePage({ lang, basePath }: { lang: string; basePath: string }) {\n i18n.locale(lang);\n const t = (key: string) => i18n.t(key) as string;\n const proj = projects[lang as \"en\" | \"pt\"];\n\n return (\n \n \n \n \n \n {profile.name} — {t(\"role\")}\n \n \n \n

Diff · v30