@udraft/core v1.8.10
uDraft - Write Once, Use Everywhere
uDraft is a language and stack agnostic code-generation tool that simplifies full-stack development by converting a single YAML file into code for rapid development. In the YAML file, you define models, enums, API endpoints, and validation rules, and uDraft automatically generates classes, database schemas, DTOs, API infrastructure, and even client APIs. This cuts down on repetitive coding and keeps everything consistent.
Using modular renderers, uDraft produces different outputs from the same source. For example, it can create:
TypeScript classes with class-validator decorators for Nest.js Mongoose schemas for MongoDB Plain models or client APIs for front-end applications Its flexible pipeline lets teams add custom renderers (for GraphQL, OpenAPI docs, ORMs, etc.) or use built-in ones. Unlike rigid scaffolding tools, uDraft separates the architecture design from the implementation details, allowing you to maintain one source of truth for your domain logic while easily updating services, databases, and clients.
📜 Core Concepts
Symbol Cheat Sheet
Symbol | Purpose | Example |
---|---|---|
/ | Node attribute | /schema: accounts |
+ | Model declaration | +user |
~ | Enum declaration | ~user-roles |
& | Reference by ID | owner[&user] |
> | Field renaming | passwordHash > pass |
$ | Special operations | $pick , $remove |
🏗️ Basic Structure
Draft Blueprint
draft:
/name: Project Name # Root attributes
/version: 1.0
module-name: # Component container
~enum-name: # Enum definition
KEY: value
+model-name: # Data model
field[type]:
- validation
feature-name: # API endpoint
/http: { method: post }
📦 Modules
Logical grouping of related components
authentication:
/description: User auth flows
~auth-methods: # Enum
EMAIL: email
GOOGLE: oauth2-google
🎚️ Enums
Fixed value collections
~user-status:
ACTIVE: active # Display : Storage
PENDING: pending
BANNED: banned
🧬 Models
Base Model
+timestamps:
createdAt[date]:
- required
updatedAt[date]:
Inherited Model
+user(timestamps): # Parenthesis inheritance
email[string]:
- unique
- format(email)
Relationships
posts[&post]: # ID reference (foreign key)
- array
profile[user-profile]: # Embedded document
🔄 Model Transformations
+user-dto(user): # Full inclusion: user-dto inherits all the user model fields
$remove: # Field exclusion: remove fields declared in this model by name
- passwordHash
- __v
$pick(user): # Selective inclusion: inherits only specific fields from user model
- _id > id # Field renaming
- email
🌐 Features (API Endpoints)
/create-user:
/http: { method: post, url: /users }
input:
+create-user-dto: # Declare new model for request adding some existing user fields
$pick(user):
- email
- password
output: user-dto # Use aleady declared user-dto model as is
✅ Field Validations
age[int]:
- min(13)
- max(120)
password[string]:
- minLength(8)
- regex(/[A-Z]/)
🚀 Complete Example
draft:
/name: Blog Platform
content:
~post-status:
DRAFT: draft
PUBLISHED: published
+post:
title[string]:
- required
content[string]:
- minLength(100)
status[post-status]:
- required(false)
create-post:
/http: { method: post, url: /posts }
input:
+create-post-request(post):
$remove:
- status
output: +post-response(post)
🛠️ Code Generation
Pipeline Flow
- Parse YAML - Load draft configuration
- Apply Renderers - Transform components to code
- Generate Output - Create files in target directories
Built-in Renderers
Renderer | Output |
---|---|
TSClassRenderer | TypeScript classes |
TSClassValidatorRenderer | TypeScript Validation decorators |
TSMongooseSchemaRenderer | TypeScript MongoDB schemas |
TSApiClientRenderer | TypeScript API Client With Axios |
DartClassRenderer | Dart classes |
DartApiClientRenderer | Dart API Client With Axios |
Using a Renderer
UDraft.load("project.yaml")
.begin("projects/backend/")
.pipeline([
new TSClassRenderer(),
new TSClassValidatorRenderer(),
new TSMongooseSchemaRenderer(),
])
.exec();
📁 Sample Output
// Generated MongoDB Schema: project/schemas/social/post-schema.ts
import { PostStatus } from "../../types/social/post-status";
import { Post } from "../../entities/social/post";
export const PostSchema = new Schema<Post>({
title: { type: String, required: true },
content: { type: String, minLength: 100 },
status: {
type: String,
enum: PostType,
},
});
export const PostModel = mongoose.model<Post>("accounts", PostSchema);
// Generated TypeScript Class with validators: project/entities/social/post.ts
import { PostStatus } from "../../types/post-status";
export class Post {
@IsString()
@IsRequired()
@IsNotEmpty()
title: string;
@IsString()
@IsOptional()
@MaxLength(100)
content?: string;
@IsOptional()
@IsEnum(PostStatus)
status?: PostStatus;
}
// Generated TypeScript Enum: project/types/social/post-status.ts
export enum PostStatus {
Draft = "dratf",
Publish = "publish",
}
💡 Pro Tips
- Reuse Models: Inherit common fields with
(parent-model)
- API Safety: Use
$remove
to exclude sensitive fields from DTOs - Validation First: Define rules in YAML for auto-generated checks
- Cross-References: Use
&model
for database relations - Renderer Stack: Combine multiple renderers for full-stack generation
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago
4 months ago