serverless-models-plugin

serverless-models-plugin

This plugin adds the missing model support to Serverless 0.5.x.

Overview

The plugin lets you define your models within your project. Models are important as soon as you play around with the mobile SDK's generated by AWS Gateway. They define the typed results or inputs of your API definition and are mapped to classes in the SDK's. One advantage is, that if a model is used in more than one endpoints of your API, or a model references other models for its sub-objects, the generated SDK uses exactly the same class definition instance throughout the SDK.

As soon as you reference them within your endpoint definitions and deploy the endpiont, the needed models are uploaded to API Gateway. If the model already exists, the definition is updated.

Installation

1. Install the plugin module npm install serverless-models-plugin will install the latest version of the plugin.

   If you want to debug, you also can reference the source repository at a specific version or branch with npm install https://github.com/HyperBrain/serverless-models-plugin#<tag or branch name>

2. Activate the plugin in your Serverless project Add serverless-models-plugin to the plugins array in your s-project.json.

   {
     "name": "testp1",
     "custom": {},
     "plugins": [
       "serverless-models-plugin"
     ]
   }
   

Usage

Model definition

Within anywhere of your Serverless project space you can create a s-models.json or s-models.yaml (both formats are supported). Within this files define your models ( you can $ref models between s-models files).

functions/
└── function1
    ├── event.json
    ├── handler.js
    ├── s-function.json
    └── s-models.json

Example (YAML)

myModelOne:
  '$schema': 'http://json-schema.org/draft-04/schema#'
  type: object
  properties:
    myProp:
      type: string
    myProp2:
      type: number
myModelTwo:
  '$schema': 'http://json-schema.org/draft-04/schema#'
  type: array
  items:
    type: string

Example (JSON)

{
  "myModelOne": {
    "$schema": "http://json-schema.org/draft-04/schema#",
    "type": "object",
    "properties": {
      "myProp": {
        "type": "string"
      },
      "myProp2": {
        "type": "number"
      }
    }
  },
  "myModelTwo": {
    "$schema": "http://json-schema.org/draft-04/schema#",
    "type": "array",
    "items": {
      "type": "string"
    }
  }
}

Model references

Models can reference other models. You can do this easily by adding $ref properties that refer to another defined model. The plugin will take care of including and deploying referenced models properly.

Example (YAML)

myModelOne:
  '$schema': 'http://json-schema.org/draft-04/schema#'
  type: object
  properties:
    myProp:
      type: string
    myProp2:
      type: number
myModelTwo:
  '$schema': 'http://json-schema.org/draft-04/schema#'
  type: array
  items:
    $ref: myModelOne

Example (JSON)

{
  "myModelOne": {
    "$schema": "http://json-schema.org/draft-04/schema#",
    "type": "object",
    "properties": {
      "myProp": {
        "type": "string"
      },
      "myProp2": {
        "type": "number"
      }
    }
  },
  "myModelTwo": {
    "$schema": "http://json-schema.org/draft-04/schema#",
    "type": "array",
    "items": {
      "$ref": "myModelOne"
    }
  }
}

Using models in endpoints

If you want to declare a response (output) or request (input) model for an endpoint, you just have to add it to your s-function.json accordingly (see requestModels and responseModels properties):

{
  "name": "testfct2",
  "runtime": "nodejs",
  "description": "Serverless Lambda function for project: testp1",
  "customName": false,
  "customRole": false,
  "handler": "handler.handler",
  "timeout": 6,
  "memorySize": 1024,
  "authorizer": {},
  "custom": {
    "excludePatterns": []
  },
  "endpoints": [
    {
      "path": "testfct2",
      "method": "GET",
      "type": "AWS",
      "authorizationType": "none",
      "authorizerFunction": false,
      "apiKeyRequired": false,
      "requestParameters": {},
      "requestModels": {
        "application/json": "myModelOne"
      },
      "requestTemplates": {
        "application/json": ""
      },
      "responses": {
        "400": {
          "statusCode": "400"
        },
        "default": {
          "statusCode": "200",
          "responseParameters": {},
          "responseModels": {
            "application/json": "myModelTwo"
          },
          "responseTemplates": {
            "application/json": ""
          }
        }
      }
    }
  ],
  "events": [],
  "environment": {
    "SERVERLESS_PROJECT": "${project}",
    "SERVERLESS_STAGE": "${stage}",
    "SERVERLESS_REGION": "${region}"
  },
  "vpc": {
    "securityGroupIds": [],
    "subnetIds": []
  }
}

Helper commands

The plugin also adds some new commands to Serverless: sls models XXXXXX

list

Lists the defined model names

show

Shows specified model definitions as JSON or YAML. Usage: sls models show <model names> [--format json|yaml]

more to come

Releases

1.1.0

• s-models.[yaml|json] files are now used throughout the project hierarchy and merged on deploy.

1.0.0

• Initial release

1.0.1

• Documentation enhancements: issues #1, #2

1.1.0

• Documentation enhancements: issues #9
• Added support for placing s-models within anywhere of the project's space