# Developing `MicroModules`

This guide assumes you have gone through Using MicroModules and already know the basic concepts about MicroModules.

Technically, a MicroModule is only a javascript class that extends from the MicroModule superclass. This superclass specifies certain properties that define how the module behaves, mainly what are its dependencies and how the modules loads itself.

# Properties of `MicroModules`

All MicroModules can define these properties:

id: a unique identifier, which means that there should be only one instance of that module in the application.
bindings: specify all the bindings that this module has registered to the MicroOrchestrator container during initialization.
dependencies: identifier of all the modules that this module depends on. They need to be present when the module is loaded.
submodules: all the modules that compose this bigger module. They will be loaded just before the present module is.

# `MicroModules` loading lifecycle

Here is the flow of how the MicroOrchestrator loads MicroModules, when executing the loadModules() function:

Register all modules ids to the list of modules being loaded at current time.

Now, for each module being loaded:

Load dependencies first:

Check the dependencies list an make sure that the modules with specified ids there are already loaded or being loaded at the same time.
If there are dependencies that are being loaded at the same time of the present module, load the dependencies first.
After that, if some dependency is missing, fail to load the module.

Load submodules: execute the load lifecycle for each of the submodules.
Execute the onLoad function of the module being loaded.

# Defining your `MicroModule`

In this guide we are going to be using typescript to define our modules, but you can define them in javascript just as easily, without specifying the types of each variable.

# 1. Create a class extending from `MicroModule`

The first step is to define a standard class extending MicroModule. The only required property to set is the onLoad function:

import { interfaces } from 'inversify';
import { MicroModule } from '@uprtcl/micro-orchestrator';

export class MyModule extends MicroModule {
  // This function will be executed when the module is being loaded
  public async onLoad(container: interfaces.Container): Promise<void> {
    // Here we should execute any loading functions that our module requires to work
  }
}

The inversify container can be modified through the Container API. But for now, we won't code our onLoad function just yet: first we're going to define dependencies and submodules.

# 2. Declare dependencies to other modules (optional)

Declaring dependencies is as easy as including the module's id in the dependency list property:

import { interfaces } from 'inversify';
import { CortexModule } from '@uprtcl/cortex';
import { MicroModule } from '@uprtcl/micro-orchestrator';

export class MyModule extends MicroModule {
  dependencies = [CortexModule.id];

  // This function will be executed when the module is being loaded
  public async onLoad(container: interfaces.Container): Promise<void> {
    // Here we should execute any loading functions that our module requires to work
  }
}

This is expressing that MyModule will only work when CortexModule is also loaded in the MicroOrchestrator.

# 3. Compose different submodules (optional)

Adding submodules is easy as well, adding the submodules property:

import { interfaces } from 'inversify';
import { CortexModule, PatternsModule } from '@uprtcl/cortex';
import { MicroModule } from '@uprtcl/micro-orchestrator';
import { CustomPattern } from './custom-pattern';

export class MyModule extends MicroModule {
  dependencies = [CortexModule.id];

  get submodules() {
    return [new PatternsModule([new CustomPattern()])];
  }

  // This function will be executed when the module is being loaded
  public async onLoad(container: interfaces.Container): Promise<void> {
    // Here we should execute any loading functions that our module requires to work
  }
}

This will load all submodules before executing the onLoad function of our module. If any of the submodules specifies an id it will only get loaded once: multiple loads of a module with the same id will only execute the onLoad function once.

# 4. Define module bindings and register services (optional)

Every module can register services, functions or any type of object inside the Container. This objects will be made available to any other module, as well as to all custom elements defined inside the customElements registry.

When the onLoad method is called, you can assume:

All dependencies will be available and loaded (you can get other module's bindings with container.get())
All submodules will be loaded

After this method is called, other modules will assume:

All bindings declared by this module are available in the container (bind them using container.bind())

import { interfaces } from 'inversify';

import { CortexModule } from '@uprtcl/cortex';
import { MicroModule, ElementsModule } from '@uprtcl/micro-orchestrator';

import { CustomPattern } from './custom-pattern';
import { HttpService } from './http-service';

export class MyModule extends MicroModule {
  dependencies = [CortexModule.id];

  get submodules() {
    return [new PatternsModule([new CustomPattern()])];
  }

  static bindings = {
    BackendService: 'http-service'
  };

  // This function will be executed when the module is being loaded
  public async onLoad(container: interfaces.Container): Promise<void> {
    // Here we should execute any loading functions that our module requires to work
    // And bind all services specified in the `bindings` property
    container.bind(MyModule.bindings.BackendService).to(HttpService);
  }
}

# 5. Define the module's id (optional)

If your module only needs to be loaded once and should serve as a base layer for other modules to depend on, you can specify the module's identifier with the id static property. This is akin to making the module 'singleton'.

This has some consequences:

The module will only be loaded once: if MicroOrchestrator finds a module with the same id that one module already present, it will not load the module a second time.
Other modules can specify a dependency to this module by including the module's id inside their dependencies property.

import { interfaces } from 'inversify';

import { CortexModule } from '@uprtcl/cortex';
import { MicroModule } from '@uprtcl/micro-orchestrator';

import { CustomPattern } from './custom-pattern';
import { HttpService } from './http-service';

export class MyModule extends MicroModule {
  static id = 'my-module';

  dependencies = [CortexModule.id];

  get submodules() {
    return [new PatternsModule([new CustomPattern()])];
  }

  static bindings = {
    BackendService: 'http-service'
  };

  // This function will be executed when the module is being loaded
  public async onLoad(container: interfaces.Container): Promise<void> {
    // Here we should execute any loading functions that our module requires to work
    // And bind all services specified in the `bindings` property
    container.bind(MyModule.bindings.BackendService).to(HttpService);
  }
}

← Using the Evees module What is Cortex? →

# Developing MicroModules

# Properties of MicroModules

# MicroModules loading lifecycle

# Defining your MicroModule

# 1. Create a class extending from MicroModule