Skip to content

getting-started.md

Latest commit

 

History

History
133 lines (96 loc) · 3.66 KB

getting-started.md

File metadata and controls

133 lines (96 loc) · 3.66 KB

Getting Started with ObjectQL

ObjectQL is a universal, metadata-driven ORM designed for building dynamic business applications. Unlike traditional ORMs where you define schema in class files (like TypeORM entities), ObjectQL uses a Metadata-First approach.

Why ObjectQL?

  1. Metadata-Driven: Define your data models in JSON or YAML. Perfect for low-code platforms where schema changes at runtime.
  2. Universal Protocol: The query language is a JSON AST, making it easy for frontends or AI agents to generate queries.
  3. Action & Hook System: Built-in support for "Button Actions" (RPC) and "Triggers" (Hooks), allowing you to model Behavior alongside Data.

Installation

Install the core package and a driver (e.g., PostgreSQL or MongoDB).

npm install @objectql/core @objectql/driver-knex knex pg
# or
npm install @objectql/core @objectql/driver-mongo mongodb

Quick Start: The "Hello World"

Let's build a simple To-Do List backend.

1. Define Your Object

In ObjectQL, everything is an "Object" (like a Table or Collection).

ObjectQL uses filename-based identification - the object name is automatically inferred from the filename, making your metadata files cleaner.

# File: todo.object.yml
# Object name "todo" is automatically inferred from filename!

label: To-Do Item
fields:
  title:
    type: text
    label: Task Name
  completed:
    type: boolean
    default: false

Note: You no longer need to specify name: todo - it's inferred from the filename todo.object.yml!

2. Configure the Engine

Updated in v0.2: You can now use a simple connection string.

import { ObjectQL } from '@objectql/core';
import * as path from 'path';

async function main() {
    const db = new ObjectQL({
        // 1. Connection String (Protocol://Path)
        // Detects 'sqlite', 'postgres', 'mongodb' automatically
        connection: 'sqlite://data.db', 
        
        // 2. Schema Source
        // Where your *.object.yml files are located
        source: ['src/objects'],

        // 3. Load Presets (Optional)
        // Load standard objects from npm packages
        presets: ['@objectql/preset-auth']
    });

    await db.init();
    
    // ...
    
    // 3. Create Data (CRUD)
    // Create a context (representing a user request)
    const ctx = db.createContext({}); 
    const todoRepo = ctx.object('todo');

    const newTask = await todoRepo.create({
        title: 'Learn ObjectQL',
        completed: false
    });
    console.log('Created:', newTask);

    // 4. Query Data
    const tasks = await todoRepo.find({
        filters: [['completed', '=', false]]
    });
    console.log('Pending Tasks:', tasks);
}

main();

Adding Business Logic

ObjectQL shines when you need to add logic.

Adding a Hook

Triggers logic automatically when data changes.

app.on('beforeCreate', 'todo', async (ctx) => {
    if (ctx.doc.title === 'Sleep') {
        throw new Error("Cannot sleep yet!");
    }
    // Auto-tagging
    ctx.doc.title = `[Task] ${ctx.doc.title}`;
});

Adding an Action

Defines a custom operation (RPC) that frontends can call.

// Define protocol
app.registerAction('todo', 'mark_done', async (ctx) => {
    const { id } = ctx;
    await ctx.object('todo').update(id, { completed: true });
    return { message: 'Good job!' };
});

// Invocation
await ctx.object('todo').execute('mark_done', 'id_123', {});

Next Steps