Library Management System

The API is intentionally small and typed end to end. Elysia route schemas validate query strings, params, and request bodies before a handler runs, while Mongoose models keep the library domain explicit. The server ships as a single Bun process with an OpenAPI reference at /docs.

Bun keeps the API, seed script, and Docker image small, and Elysia's schema-first routes gave me validation and the OpenAPI reference from one definition. MongoDB and Mongoose fit the document-shaped catalog and order data without needing a migration for every content change.

The catalog endpoint resolves author matches first, then searches books by title, ISBN, or author reference. Results are paginated with total pages, next/previous flags, and author documents populated. A single book can be fetched by Mongo ID or ISBN, which keeps the storefront and admin views on the same route.

• Search by title, ISBN, or author name
• Pagination metadata for storefront and admin tables
• Book details with populated authors, stock, and ratings
• Author search with prefix matching

Placing an order validates every line against live stock before creating the customer and order. To avoid accidental duplicates, the same phone number cannot open a second pending order with an identical set of books. When an order moves through pending, sold, delivered, or canceled, inventory and reporting stay in sync.

1export const orderRoutes = new Elysia({ prefix: "/orders" })2  .post(3    "/",4    async ({ body, status }) => {5      for (const item of body.items) {6        const book = await Book.findById(item.bookId);7 8        if (!book) {9          return status(404, `Book ${item.bookId} not found`);10        }11 12        if (book.quantity < item.quantity) {13          return status(14            409,15            `Not enough stock for "${book.title}", only ${book.quantity} left`,16          );17        }18      }19 20      return createOrder(body);21    },22    {23      body: t.Object({24        phone: t.String(),25        items: t.Array(26          t.Object({27            bookId: t.String(),28            quantity: t.Integer({ minimum: 1 }),29          }),30        ),31      }),32    },33  );

Customers can suggest a title that is missing from the catalog. The request records title, author, and notes, and warns when a matching title already has stock. Admins filter requests by status and mark them acquired or rejected, keeping the suggestion queue short and actionable.

The users plugin issues JWT cookies at register and login, keeps the token httpOnly, and exposes a /me endpoint for the current session, with passwords hashed using bcrypt. Cover images use a signed Cloudinary upload flow, and uploaded files are served through Elysia's static plugin. The project includes a seed script with authors, books, customers, orders, requests, and comments for local development.

• JWT cookie auth with /register, /login, and /me
• Signed uploads and static file serving
• OpenAPI reference at /docs
• Docker image based on oven/bun:alpine on port 8000