🛑 Pipes (Validation)

In NestJS, a Pipe is a class annotated with @Injectable() that implements the PipeTransform interface, typically used to validate or transform input data before it reaches the Controller.

In Ferrox, Pipes are implemented automatically through the Validation Extractor pattern, leveraging the validator crate.

High-Level Example

To validate incoming JSON data, you simply need to decorate your Data Transfer Object (DTO) with validation rules, and use the Json extractor in your controller.

use axum::{Json, routing::post, Router};
use validator::Validate;
use serde::{Deserialize, Serialize};

#[derive(Debug, Serialize, Deserialize, Validate)]
pub struct CreateUserDto {
    #[validate(email)]
    pub email: String,
    
    #[validate(length(min = 8, message = "Password too short"))]
    pub password: String,
}

// The framework automatically runs the Pipe (validation) before executing this function
async fn register_user(Json(payload): Json<CreateUserDto>) -> Json<&'static str> {
    payload.validate().unwrap(); // Trigger the Pipe
    
    Json("User Created!")
}

Low-Level Internal Details

While you can manually call .validate() inside the controller, Ferrox allows you to build custom Extractors that act as automatic Pipes.

By implementing FromRequest on a wrapper struct (e.g., ValidJson<T>), Ferrox will intercept the raw TCP bytes, deserialize them into JSON, and run the validator rules. If the rules fail, Ferrox intercepts the error and immediately returns an HTTP 400 Bad Request with a structured AppError::ValidationError, completely bypassing the Controller logic.