Generic Controller

Generic controller is a common Plumier controller with generic type signature, it take advantage of inheritance to possibly serve CRUD API based on ORM entity.

Enable Functionality

Generic controller supported TypeORM and Mongoose (with Plumier mongoose helper) entities. Enable the generic controller by installing the TypeORMFacility or MongooseFacility on the Plumier application.

import Plumier, { WebApiFacility } from "plumier"

import { TypeORMFacility } from "@plumier/typeorm"

new Plumier()

.set(new WebApiFacility())

.set(new TypeORMFacility())

Or if you are using mongoose helper like below

import Plumier, { WebApiFacility } from "plumier"

import { MongooseFacility } from "@plumier/mongoose"

new Plumier()

.set(new WebApiFacility())

.set(new MongooseFacility())

Above facilities is a common facility used if you are using TypeORM or Mongoose with Plumier. In context of generic controller above facilities will normalize entities to make it ready used by generic controller helpers.

Mark Entity Handled by Generic Controller

After installing facility above you need to mark specific entity that will be generated into CRUD API with @genericController() like below:

import { Entity, PrimaryGeneratedColumn } from "typeorm"

import { route } from "plumier"

@genericController()

@Entity()

class User {

@PrimaryGeneratedColumn()

id: number

@Column()

email: string

@Column()

name: string

}

Or if you using Mongoose helper

import { collection } from "@plumier/mongoose"

import { route } from "plumier"

@genericController()

@collection()

class User {

constructor(

public id: string,

public email: string,

public name: string

){}

}

Above code will generate six routes handled by generic controller implementation.

Method	Route	Description
POST	/users	Add new user
GET	/users/:id?select	Get user by ID
PUT	/users/:id	Replace user by ID (required validation used)
PATCH	/users/:id	Modify user by ID (required validation ignored)
DELETE	/users/:id	Delete user by ID
GET	/users	Get list of users with pagination, order, filter and projection

Its possible to create a generic controller using generic controller factory instead of using decorator, below code will result the same.

import { GenericController } from "@plumier/typeorm"

// or if using mongoose

// import { GenericController } from "@plumier/mongoose"

const UserController = GenericController(User)

// Returned controller is a valid controller class,

// so you can override its method like below

export class UserController extends GenericController(User) {

get(id:number) {

return super.get(id)

}

}

The factory receive controller builder factory configuration like below

export class UserController extends GenericController(User, c => {

c.mutators().authorize("Admin")

})

Getting and Saving Simple Relation

Relational data with single value (one to one or many to one) by default will be populated on each request. For example if we have entity below:

@genericController()

@Entity()

class Address {

/** other columns **/

@Column()

city:string

@Column()

address:string

}

@genericController()

@Entity()

class User {

/** other columns **/

@Column()

email:string

@OneToOne(x => Address)

address:Address

}

Above code generates 6 routes for each /address and /users. User entity contains relation property to Address entity which is a one to one relation. Issuing GET /users/:id will automatically populate the address, thus the response will be like below

{

"email": "john.doe@gmail.com",

// full address object populated

"address": {

"city": "Badung",

"address": "Jl Surapati No. 19 Kuta"

}

}

While GET /users/:id returns the full address object, saving address (POST, PUT, PATCH) only require the ID of the address like below

POST /users/123

{

"email": "john.doe@gmail.com",

"address": 456 //<-- address ID

}

Getting and Saving Array Relation

For array relation (one to many relation), Plumier provide a nested route to easily perform CRUD operation on child relation.

@Entity()

class User {

/** other columns **/

@Column()

name:string

@genericController()

@OneToMany(x => Email, x => x.user)

emails:Email[]

}

@Entity()

class Email {

/** other columns **/

@Column()

email:string

@Column()

description:string

@ManyToOne(x => User, x => x.emails)

user:User

}

Above code showing that we apply @genericController() on the User.emails relation. Using this setup will make Plumier generate a nested routes like below

Method	Route	Description
POST	/users/:pid/emails	Add new user's email
GET	/users/:pid/emails/:id?select	Get user's email by ID
PUT	/users/:pid/emails/:id	Replace user's email by ID
PATCH	/users/:pid/emails/:id	Modify user's email by ID
DELETE	/users/:pid/emails/:id	Delete user's email by ID
GET	/users/:pid/emails	Get list of user's emails with paging, filter, order and projection

Its possible to create a nested generic controller using generic controller factory, below code will result the same.

import { GenericController } from "@plumier/typeorm"

// or if using mongoose

// import { GenericController } from "@plumier/mongoose"

export const UserEmailController = GenericController([User, "emails"])

Its also possible to use many to one relation to generate nested routes, from previous example we can decorate the Email.user property. This feature useful to create nested routes without explicit relation between parent-children, for example to reduce number of relation properties on parent entity.

@Entity()

class User {

/** other columns **/

@Column()

name:string

}

@Entity()

class Email {

/** other columns **/

@Column()

email:string

@Column()

description:string

@genericController()

@ManyToOne(x => User)

user:User

}

Instead of using one to many relation, we can decorate the Email.user property and ignore User.animals property.

The generic controller factory for above feature also work the same as before, we specify tuple relation Animal and user as the parameter of the generic controller factory

import { GenericController } from "@plumier/typeorm"

// or if using mongoose

// import { GenericController } from "@plumier/mongoose"

export const UserEmailController = GenericController([Animal, "user"])

Inverse Property Population

If you are defined your one to many relation ORM configuration with inverse property, Generic controller will taking care of populating its value automatically. For example if you have ORM configuration like below.

@Entity()

class User {

/** other columns **/

@genericController()

@OneToMany(x => Email, x => x.user)

emails:Email[]

}

@Entity()

class Email {

/** other columns **/

@ManyToOne(x => User, x => x.emails)

user:User

}

Or if you are using mongoose helper like below

@collection()

class User {

/** other columns **/

@genericController()

@collection.ref(x => [Email], "user")

emails:Email[]

}

@Entity()

class Email {

/** other columns **/

@collection.ref(x => User)

user:User

}

You're defined inverse property user on the one to many relation, it help Plumier to understand which property will automatically populated. So adding new email using POST /users/{pid}/emails will make user automatically populated with pid value, without having to provided on request body.

Apply Multiple Decorators

Its possible to apply multiple @genericController() decorator on entity or entity relation, but the generated route must be unique or the route generator static check will shows errors.

@genericController(c => c.setPath("user-emails/:id"))

@genericController()

@Entity()

class Email {

/** other columns **/

@Column()

email:string

@Column()

description:string

@ManyToOne(x => User, x => x.emails)

user:User

}

Above will generates 12 routes like below

Method	Route	Description
POST	/emails	Add new email
GET	/emails/:id?select	Get email by ID
PUT	/emails/:id	Replace email by ID (required validation used)
PATCH	/emails/:id	Modify email by ID (required validation ignored)
DELETE	/emails/:id	Delete email by ID
GET	/emails	Get list of emails with pagination, order, filter and projection
POST	/user-emails	Add new email
GET	/user-emails/:id?select	Get email by ID
PUT	/user-emails/:id	Replace email by ID (required validation used)
PATCH	/user-emails/:id	Modify email by ID (required validation ignored)
DELETE	/user-emails/:id	Delete email by ID
GET	/user-emails	Get list of emails with pagination, order, filter and projection

Filters

Generic controller provided filter query string to narrow API response. The query respects the @authorize.read() and @authorize.writeonly(). Its mean you will not able to query

import { Entity, PrimaryGeneratedColumn } from "typeorm"

import { route } from "plumier"

@genericController()

@Entity()

class User {

@PrimaryGeneratedColumn()

id: number

// name will be searchable by anyone (respect route authorization)

@Column()

name: string

// authorize email field to be search able by admin

@authorize.read("Admin")

@Column()

email: string

}

Using above code enabled us to query the response result using syntax string like below

# filter by email

GET /users?filter=email=john.doe@gmail.com

# or more readable using group

GET /users?filter=(email=john.doe@gmail.com)

# filter by name, will return all users name start with john

GET /users?filter=(name='john'*)

# combine both filter, will return with AND operator

GET /users?filter=(email=john.doe@gmail.com and name=john)

Several filter supported based on property data type, for more information see the Query Language Specification

Delete Column

By default when you perform DELETE /users/{id} it will delete the user record permanently from the database, You can specify the delete flag by providing @entity.deleteColumn() decorator above the flag property with boolean datatype like below.

import { entity } from "plumier"

@Entity()

export class User {

@PrimaryGeneratedColumn()

id: number

/** other properties **/

@entity.deleteColumn()

@Column({ default: false })

deleted: boolean

}

Using above code when you request DELETE /users/{id} Plumier will set the deleted property into true automatically.

Query Strings

Both get by id and get list route has some extra query string to manipulate the response match your need.

Query String	Example	Default	Description
select	GET /users?select=name,email,dob	All properties	Select properties to include in JSON response
limit	GET /users?limit=20	50	Limit number of records returned in response
offset	GET /users?offset=3	0	Offset of the record set returned in response
filter	GET /users?filter=(name='john' and active=true)	-	Find records by property using exact comparison
order	GET /users?order=-createdAt,name	-	Order by properties, - for descending order

Above query string supported by generic controller and nested generic controller.

Get By ID

Get by ID for generic controller and nested generic controller supported select query string like below

GET /users/:id?select=name,email,dob

GET /users/:pid/pets/:id?select=name,active,dob

Get List

Get list supported all the query string, it can be combined in single request

GET /users?select=name,email,dob&filter=(email='john.doe@gmail.com')&order=-createdAt,name&offset=5

GET /users/:pid/pets/:id?select=name,dob&filter=(name=bingo)&order=-createdAt,name&offset=5

Custom Path Name

Plumier provide a default route path name based on entity name (pluralized), you can specify a new path name by provide it on the @genericController() parameter.

import { Entity, PrimaryGeneratedColumn } from "typeorm"

import { route } from "plumier"

@genericController("user-data/:uid")

@Entity()

class User {

@PrimaryGeneratedColumn()

id: number

@Column()

email: string

@Column()

name: string

}

Above code showing that we specify custom path name user-data/:uid, it will generate routes like below

Method	Route	Description
POST	/user-data	Add new user
GET	/user-data/:uid	Get user by ID
PUT	/user-data/:uid	Replace user by ID (required validation used)
PATCH	/user-data/:uid	Modify user by ID (required validation ignored)
DELETE	/user-data/:uid	Delete user by ID
GET	/user-data	Get list of users with paging, filter, order and projection

For nested generic controller you need to specify ID for parent and the child

@Entity()

class User {

/** other columns **/

@genericController("user-data/:uid/email-data/:eid")

@OneToMany(x => Email, x => x.user)

emails:Email[]

}

Above code will generate routes below

Method	Route	Description
POST	/user-data/:uid/email-data	Add new user's email
GET	/user-data/:uid/email-data/:eid	Get user's email by ID
PUT	/user-data/:uid/email-data/:eid	Replace user's email by ID
PATCH	/user-data/:uid/email-data/:eid	Modify user's email by ID
DELETE	/user-data/:uid/email-data/:eid	Delete user's email by ID
GET	/user-data/:uid/email-data	Get list of user's email with paging, filter, order and projection

Control Access To The Entity Properties

Plumier provide functionality to protect your data easily, you can use @authorize decorator to authorize user to write or read your entity property.

info

Refer to Security on how to setup user authorization on your Plumier application

import { Entity, PrimaryGeneratedColumn } from "typeorm"

import { route, authorize } from "plumier"

@genericController()

@Entity()

class User {

@PrimaryGeneratedColumn()

id: number

@Column()

email: string

@authorize.writeonly()

@Column()

password: string

@Column()

name: string

@authorize.write("SuperAdmin", "Admin")

@Column()

role: "SuperAdmin" | "Admin" | "User"

}

Above code showing that we apply @authorize decorator on password and role property which contains sensitive data. Using above configuration password will not be visible on any response, and role only can be set by SuperAdmin and Admin. Below list of authorization you can use to protect property of the entity

Decorator	Description
@authorize.route("SuperAdmin")	Protect route can be accessed by specific role (SuperAdmin)
@authorize.write("SuperAdmin")	Protect property only can be write by specific role (SuperAdmin)
@authorize.read("SuperAdmin")	Protect property only can be read by specific role (SuperAdmin)
@authorize.readonly()	Protect property only can be read and no other role can write it
@authorize.writeonly()	Protect property only can be write and no other role can read it

Control Access To The Generated Routes

You can specify authorization into specific generated route by providing more configuration on the @genericController() decorator.

import { Entity, PrimaryGeneratedColumn } from "typeorm"

import { route, authorize } from "plumier"

@genericController(c => c.mutators().authorize("SuperAdmin", "Admin"))

@Entity()

class User {

/** properties / columns */

}

Above code showing that we apply authorization to the mutators() http methods. mutators() is a syntax sugar to specify generic controller actions handled the POST, PUT, PATCH and DELETE. Using above configuration the result of the generated routes are like below

Action	Method	Route	Access
save	POST	/users	SuperAdmin, Admin
get	GET	/users/:id	Any user
replace	PUT	/users/:id	SuperAdmin, Admin
modify	PATCH	/users/:id	SuperAdmin, Admin
delete	DELETE	/users/:id	SuperAdmin, Admin
list	GET	/users	Any user

info

If no authorization specified, the default authorization for generated route is Authenticated means all authenticated user can access the route.

For nested routes (one to many relation) you can define the same configuration like above.

@Entity()

class User {

/** other columns **/

@genericController()

@authorize.route(c => c.mutators().authorize("SuperAdmin", "Admin"))

@OneToMany(x => Email, x => x.user)

emails:Email[]

}

Using above configuration the route access now is like below

Action	Method	Route	Access
save	POST	/users/:pid/emails	SuperAdmin, Admin
get	GET	/users/:pid/emails/:id	Any user
replace	PUT	/users/:pid/emails/:id	SuperAdmin, Admin
modify	PATCH	/users/:pid/emails/:id	SuperAdmin, Admin
delete	DELETE	/users/:pid/emails/:id	SuperAdmin, Admin
list	GET	/users/:pid/emails	Any user

Ignore Some Routes

In some case you may want to hide specific route generated. You can use the ignore() configuration to ignore specific methods

import { Entity, PrimaryGeneratedColumn } from "typeorm"

import { route } from "plumier"

@genericController(c => {

c.put().ignore()

c.patch().ignore()

c.delete().ignore()

})

@Entity()

class User {

/** properties / columns */

}

Above code showing that we specify the method one by one instead of using mutators() like the previous example. The result of the

Action	Method	Route
save	POST	/users
get	GET	/users/:id
list	GET	/users

It also can be applied on the entity relation (one to many) to ignore some nested routes like below

@Entity()

class User {

/** other columns **/

@genericController(c => {

c.put().ignore()

c.patch().ignore()

c.delete().ignore()

})

@OneToMany(x => Email, x => x.user)

emails:Email[]

}

Above code showing that we applied the ignore decorator on the entity relation, it will produce

Action	Method	Route
save	POST	/users/:pid/emails
get	GET	/users/:pid/emails/:id
list	GET	/users/:pid/emails

Transform Response

Its possible to transform the response result of the GET method using transformer() configuration. For example we have entities below

@Entity()

export class Comment {

@PrimaryGeneratedColumn()

id: number

@Column()

message: string

@ManyToOne(x => User)

user: User

@ManyToOne(x => Todo)

todo:Todo

}

You may want the response of GET /users only contains properties id, message, userName and userPicture, which userName and userPicture can be retrieved from user property.

First you need to create the model of the transformed entity.

export class CommentWithUser {

@noop()

id: number

@noop()

message: string

@noop()

userName: string

@noop()

userPicture: string

}

Then specify the transformation function using transformer configuration

@genericController(c => {

c.accessors().transformer(CommentWithUser, x => {

return {

id: x.id,

message: x.message,

userName: x.user.name,

userPicture: x.user.picture

}

})

})

@Entity()

export class Comment {

@PrimaryGeneratedColumn()

id: number

@Column()

message: string

@ManyToOne(x => User)

user: User

@ManyToOne(x => Todo)

todo:Todo

}

Above code showing that we specify the transformer configuration, the first parameter is the target model then the second parameter is the transformation function.

caution

When using response transform, the select query may still applied but the response will be based on what you returned in transform function.

Custom Query

Unlike transform response, with custom query you provide a custom database query for getOne and/or getMany method to get different response result than the default generic controller query provided.

import { route } from "plumier"

import { Entity, Column, PrimaryGeneratedColumn } from "typeorm"

import { noop } from "@plumier/reflect"

// for mongoose use import { transformFilter } from "@plumier/mongoose"

import { transformFilter } from "@plumier/typeorm"

@genericController(c => {

// custom query for GET /users/:id

c.getOne().custom(UserDto, async ({ id }) => {

const repo = getManager().getRepository(User)

return repo.findOne(id, { select: ["email"] })

})

// custom query for GET /users

c.getMany().custom([UserDto], async ({ limit, offset, filter }) => {

const repo = getManager().getRepository(User)

const where = transformFilter(filter)

return repo.find({ take: limit, skip: offset, where, select: ["email"] })

})

})

@Entity()

class User {

@PrimaryGeneratedColumn()

id: number

@Column()

email: string

@Column()

name: string

}

// the response will be like this

class UserDto {

@noop()

email: string

}

Using above code you provide a custom query that will be used by the GET /users/:id. To override query of GET /users you can use the getMany() method instead of getOne().

Signature of the custom query method is like below

custom(responseType, queryCallback)

• responseType is the model used to generate schema of the response. This model used by Open API generator to generate response schema, and also used by response authorization. Important to note that the @authorize.read() on the entity will not take effect if you use different model than the entity, you need to decorate the appropriate property accordingly.
• queryCallback is a function returned the database query, signature of the query callback is mostly the same for getOne() and getMany(), except the first parameter object, see below.

Query callback signature for getOne() and getMany() is like below

(param, ctx) => any

• param contains the action parameter such as id, limit, offset etc. The parameter is differ between getOne() and getMany().
• ctx is the request context

Entity Authorization Policy

Entity Authorization Policy (Entity Policy) is a custom auth policy designed to secure entity data by ID.

Entity policy requires unique combination between policy name and the entity type, the definition is like below.

entityPolicy(ENTITY_CLASS)

.register(POLICY_NAME, (auth:AuthorizationContext, id:any) => boolean)

• ENTITY_CLASS is the entity which authorization policy registered to.
• POLICY_NAME is the policy name, policy name may the same with other entity but must be unique when combined with entity class.
• Authorization callback is where you put the authorization logic, return true or Promise<true> to allow otherwise false to restrict.

Since the internal process of entity policy is quite complex, its a lot easier to understand entity policy from its implementation. For example we created an entity policy named ResourceOwner which means the current data is owned by the login user. Then its easy to understand that for entity User, the logic of ResourceOwner is when the provided ID by authorization callback is the same as the current login user ID like below.

// define policy for User entity

entityPolicy(User)

// ResourceOwner is when current login user id is the same as current User id

.register("ResourceOwner", (auth, id) => auth.user?.userId === id)

This policy then can be applied on the generic controller configuration builder, to secure specific route like below.

@genericController(c => {

c.delete().authorize("ResourceOwner")

})

@Entity()

export class User {

@PrimaryGeneratedColumn()

id:number

@Column()

email: string

@Column()

name: string

}

By using above configuration only the ResourceOwner of the User entity allowed to access DELETE /users/{id} route. By giving DELETE /users/12345 only user with user ID 12345 allowed to access it.

The policy also can be used to secure property using @authorize.read() or @authorize.write() like below.

@genericController()

@Entity()

export class User {

@PrimaryGeneratedColumn()

id:number

@authorize.read("ResourceOwner")

@Column()

email: string

@Column()

name: string

}

Above configuration will make email filed only visible (in response body) by the owner of user, this behavior applied when the User entity used as child property.

Entity Policy Restriction

Since entity policy designed to secure resources by ID, its important to note that route authorization (defined with @authorize.route() or ActionBuilder.authorize()) and parameter authorization (@authorize.write()) only works on route with entity ID specified.

Route	Description
POST /users	Not allowed, since no ID specified
DELETE /users/{id}	Allowed, Callback called with ID of User entity
POST /users/{id}/logs	Allowed, Callback called with ID of User entity
DELETE /users/{id}/logs/{logId}	Allowed, Callback called with ID of Log entity

Important to note that response authorization (@authorize.read()) has no restriction, its mean it will always work on all routes.

Request Hook

Request hook enables entity to have a method contains piece of code that will be executed during request. Request hook has some simple rule

• It executed before or after the entity saved into the database, use decorator @preSave() and @postSave()
• It will be executed only on request with http method POST, PUT, PATCH. By default it will execute on those three http methods except specified on the parameter.
• It can be specified multiple request hooks on single entity
• It can have parameter with parameter binding
• It possible to bind ActionResult (execution result of the controller) on request hook @postSave()

import { Entity, PrimaryGeneratedColumn } from "typeorm"

import { route, preSave } from "plumier"

import bcrypt from "bcrypt"

@genericController()

@Entity()

class User {

@PrimaryGeneratedColumn()

id: number

@Column()

email: string

@Column()

name: string

@Column()

password: string

@preSave()

async hashPassword(){

// executed before user saved

this.password = await bcrypt.hash(data.password, 10)

}

@postSave()

async sendConfirmationEmail(){

// executed after user saved

await mailer.sendTemplate("confirmation-email")

}

}

Above code will hash password before the entity saved into the database. Request hook has parameter binding feature, you can @bind any request part into the hook parameter, it works exactly like common Parameter Binding which also support name binding, model binding and decorator binding.

info

The ID of the current entity only accessible on @postSave using this.id, since on @postSave() the entity is not saved yet to database.

caution

Keep in mind that entity used for @preSave() and @postSave() is different, means if you using state variable to share between @preSave() and @postSave() its may not working like expected.

import { Entity, PrimaryGeneratedColumn } from "typeorm"

import { route, preSave } from "plumier"

import bcrypt from "bcrypt"

@genericController()

@Entity()

class User {

/** other properties **/

@preSave()

async hook(@bind.ctx() ctx:Context){

// ctx will contains context

}

@postSave()

async postHook(@bind.ctx() ctx:Context){

// ctx will contains context

}

}

Its possible to specify in which http method should the hook executed by specify http method on the request hook parameter

import { Entity, PrimaryGeneratedColumn } from "typeorm"

import { route, preSave } from "plumier"

import bcrypt from "bcrypt"

@genericController()

@Entity()

class User {

/** other properties **/

@preSave("put", "patch")

async hook(){

// this only executed on PUT and PATCH http method before entity saved

}

@postSave("put", "patch")

async postHook(){

// this only executed on PUT and PATCH http method after entity saved

}

}

Use Custom Generic Controller

When the default generic controller doesn't match your need, you can provide your own custom generic controllers. For example the default generic controller for GET method doesn't contains count of the match records usually used for table pagination on UI side. You can override this function by provide a new generic controller inherited from your ORM/ODM helper:

Generic Controller	Package	Description
TypeORMControllerGeneric<T, TID>	@plumier/typeorm	TypeORM generic controller implementation
TypeORMNestedControllerGeneric<P, T, PID, TID>	@plumier/typeorm	TypeORM One To Many generic controller implementation
MongooseControllerGeneric<T, TID>	@plumier/mongoose	Mongoose generic controller implementation
MongooseNestedControllerGeneric<P, T, PID, TID>	@plumier/mongoose	Mongoose One To Many generic controller implementation

First define the model represent the response schema returned by each controller using generic type like below

import {generic, type, noop} from "@plumier/reflect"

@generic.parameter("T")

export class Response<T> {

@type(["T"])

data: T[]

@noop()

count: number

}

Then create a new generic controller for both based on any of above generic controller like below

import {TypeORMControllerGeneric, TypeORMNestedControllerGeneric} from "@plumier/typeorm"

export class CustomControllerGeneric<T, TID> extends TypeORMControllerGeneric<T, TID> {

@type(Response, "T")

async list(offset: number, limit: number, filter: any, select: SelectQuery, order: any, ctx: Context) {

const data = await super.list(offset, limit, filter, select, order, ctx)

const count = await this.repo.count(filter)

return { data, count }

}

}

export class CustomNestedControllerGeneric<P, T, PID, TID> extends TypeORMNestedControllerGeneric<P, T, PID, TID>{

@type(Response, "T")

async list(pid: PID, offset: number, limit: number, filter: any, select: SelectQuery, order: any, ctx: Context) {

const data = await super.list(pid, offset, limit, filter, select, order, ctx)

const count = await this.repo.count(pid, filter)

return { data, count }

}

}

Using above code, we simply call the super.list method and use the repository count method to create the response match the Response data structure.

Note that we define the return type of the method as @type(Response, "T"), this means we specify that the Response.data property is of type of T of the generic controller.

Next, we need to register above custom generic controller on the Plumier application like below

new Plumier()

.set(new WebApiFacility())

.set(new TypeORMFacility())

// Make sure to register the controller under the `TypeORMFacility` or `MongooseFacility`.

.set({

genericController: [CustomControllerGeneric, CustomNestedControllerGeneric]

})

Keep in mind that the controller created using generic controller factory GenericController(Entity) or GenericController([Entity, "children"]) will keep using the default generic controllers, to solve the issue you can use the generic controller factory factory like below.

import { createGenericControllerTypeORM } from "@plumier/typeorm"

// for mongoose use below import

// import { createGenericControllerMongoose } from "@plumier/mongoose"

export const CustomGenericController = createGenericControllerTypeORM([CustomControllerGeneric, CustomNestedControllerGeneric])

With above code we created a new generic controller factory which based on our custom generic controllers. Export the factory and use it anywhere on your code like usual generic controller factory.

const UsersController = CustomGenericController(User)

// or

class UsersController extends CustomGenericController(User) { }