Arryt — kuryer logistikasi v2

• Bu nima: yetkazib berish/kuryerlik boshqaruv tizimi (jumladan, Restaurant chain (UZ) ekotizimida ishlatiladi — qarang: crm_chopar dagi check_arryt). Menejerlar uchun admin panel + REST/type-safe API + analitika + Flutter haydovchi ilovasi + navbatlar/cron. Buyurtmalarni tayinlash, haydovchilarni kuzatish, narxlash dvigateli, terminallar bo'yicha balanslar.
• Turi / holati / roli: web-app (ko'p ilovali + mobil) / active / lead (case study bo'yicha "Lead Full-Stack Developer", 3 kishilik jamoa; git da foydalanuvchi bosh muallif — 173/329 ≈ 53%, undan keyin Ilhombek 90, Shahzod 34).
• Faollik davri: 2023-11-04 → 2026-03-22 (~2.5 yil); new_arryt — bu arryt (#8) ning zamonaviy qayta yozilgan varianti (Elysia/Bun/Next15).

• API (`api/`): ElysiaJS (Bun) + Drizzle ORM (+typebox/zod) + PostgreSQL + Redis; BullMQ/Bull (joblar); DuckDB (@evan/duckdb — analitika); OpenTelemetry (@elysiajs/opentelemetry, auto-instrumentations, OTLP exporter — taqsimlangan tracing); server-timing, bearer-auth, Eden.
• Admin (`admin/`, `new_admin/`): Next.js 15 + React 19, Tanstack Query, shadcn/ui, real-time (WebSocket).
• Mobil (`mob/`): Flutter (Dart) — haydovchi ilovasi: BLoC, GraphQL, ObjectBox (offline), Firebase (push), fon rejimda GPS kuzatuvi.
• Boshqa: mcp_server/ (maxsus MCP server!), protos/ + proto.ts (gRPC/protobuf), duck_api/ (DuckDB analitikasi), queue/, crons/, transfer_db/.
• DevOps: Docker, PM2, GitHub Actions, Nginx.

• Elysia/Bun/Next15 (v2) da type-safe API (Eden) bilan to'liq qayta yozish.
• PostgreSQL ustida DuckDB analitikasi (1M+ yozuv bo'yicha hisobotlar).
• API uchun OpenTelemetry tracing.
• Flutter haydovchi ilovasi (offline rejim, GPS, push).
• Maxsus MCP server (mcp_server/) — ehtimol AI ga logistika ma'lumotlariga kirish uchun.
• gRPC/protobuf qatlami (protos/, proto.ts).
• Arryt-Case-Study.md/html (portfolio case study) va arryt_monetization.md (monetizatsiya) tayyorlangan.
• Hajm: ~2.5 yilda 329 kommit, shundan 173 tasi foydalanuvchidan (bosh muallif).

(Kod/manifestlar + case study asosida; foydalanuvchi lead / bosh muallif.)

• PostgreSQL ustida DuckDB analitikasi (@evan/duckdb, duck_api/, order_data.duckdb): 1M+ buyurtma bo'yicha og'ir analitik so'rovlar embedded-OLAP DuckDB ga o'tkaziladi, bu tranzaksion Postgres ni yengillashtiradi. → Toza OLTP/OLAP ajratmasi — kam uchraydigan va kuchli tanlov.
• Taqsimlangan tracing (OpenTelemetry) Elysia API da (auto-instrumentations + OTLP): production kuzatuvchanligi va kechikishni optimallashtirish (case study bo'yicha <50ms). → Yetuk observability yondashuvi.
• Dinamik narxlash dvigateli: masofa/vaqtga asoslangan qoidalar, haydovchi bonuslari, mijozga xos tariflar, kunlik kafolat; qoidalar Redis da keshlanadi + o'zgarishlar auditi.
• 500+ haydovchini real vaqtda kuzatish: batareyani tejaydigan fon GPS, holatni avtomatik o'zgartirish uchun geofencing, xaritada markerlarni klasterlash.
• gRPC/protobuf + maxsus MCP server: type-safe servislararo qatlam + AI ga logistika ma'lumotlariga kirish.
• Yuqori yuklamali ishlov berish: kursorli paginatsiya, 1000+ buyurtma bo'yicha bulk operatsiyalar, indekslarni optimallashtirish.

• Topilgan sessiyalar: bu loyiha uchun lokal Claude Code sessiyalar katalogida 39 ta .jsonl fayl. Shuningdek .cursor / .cursorrules (26 KB) — Cursor + Claude Code.
• AI bilan nima qilingan: v2 qayta yozish/funksiyalar, ehtimol case study va MCP server generatsiyasi. Chuqur tahlil qilinmagan (katta hajm).
• Pattern: ko'p asbobli (Claude Code + Cursor), mahsulotning bir qismi sifatida maxsus MCP server (AI ga ma'lumotlarga kirish).