Skip to content

Latest commit

聽

History

History
313 lines (232 loc) 路 9.3 KB

project-js.md

File metadata and controls

313 lines (232 loc) 路 9.3 KB
Raw

project.js - Project Configuration

The project.js is an essential file that stores a lot of important project information that is used for deployment.

Introduction

The project.js is used by many features of Jovo, for example to build platform specific project files with the Jovo CLI, or to add different stages for more convenient deployment.

Here is a list of all elements that can be added to the project.js:

Category Name Description
Platforms alexaSkill Alexa Skill project config
googleAction Google Action project config
nlu Includes natural language understanding tools like dialogflow
Deployment endpoint Endpoint for the voice platforms to reach your app. Default: Jovo Webhook.
host Deployment information (e.g. if hosted on Lambda, but endpoint is API Gateway)
stages Different deployment stages
defaultStage Default stage to use if not defined differently in the environment variables

This is what the default project.js looks like:

// ------------------------------------------------------------------
// JOVO PROJECT CONFIGURATION
// ------------------------------------------------------------------

{
	alexaSkill: {
		nlu: 'alexa',
	},
	googleAction: {
		nlu:  'dialogflow',
	},
	endpoint: '${JOVO_WEBHOOK_URL}',
}

As you can see above, an endpoint is automatically added to the file with your unique Jovo Webhook endpoint.

Alternatively, you can reference values from your environment variables like so:

${process.env.KEY}

// Example: Get endpoint from env
endpoint: process.env.ENDPOINT,

These environment variables are usually stored in a .env file or in your cloud environment, for example AWS Lambda.

Platforms

With the Jovo CLI, you can create and deploy project files (e.g. language models) that are specific to each voice platform.

The introduction already shows how a freshly created project.js looks like for both platforms alexaSkill and googleAction. In the following section, you will learn about additional configurations to the voice platform projects.

alexaSkill

The following elements can be added to the alexaSkill object:

{
    alexaSkill: {
        nlu: {
            name: 'alexa',
        },
        skillId: '<your-skill-id>',
        askProfile: '<your-ask-cli-profile>'
    },
}

In the deployment process with the Jovo CLI (jovo deploy), skillId and askProfile are by default taken from the existing config in the /platforms/alexaSkill/.ask folder. Specifying those in the project.js to override the existing platform files is especially useful if you have different versions of the Alexa Skill in separate developer accounts (see Stages).

skill.json Overrides

You can also add information to go into the skill.json. You can basically add or override any element that you can find in the skill manifest:

{
    alexaSkill: {
        manifest: {
            // Add element to override here
        }
    },
}

To add an object, make sure to add the same path it has in the skill.json. The below example (from our Alexa Audioplayer Template) adds the Audioplayer Directive to the Alexa Skill:

{
    alexaSkill: {
        manifest: {
			apis: {
				custom: {
				  interfaces: [
					{
					  type: 'AUDIO_PLAYER',
					}
				  ]
				}
			}
		}
	},
}

googleAction

The deployment communication between the Jovo CLI and the Google Action is going through a Dialogflow agent. This is why project configuration settings are added to a dialogflow object inside the googleAction block:

googleAction: {
    dialogflow: {
        projectId: '<your-project-id>',
        keyFile: './path/to/key-file',
    }
}

Updating a Dialogflow Agent through the Jovo CLI needs an authentication process with a keyFile, which can be downloaded and placed in your project directory. For more information, take a look at the example in our Tutorial: Deploy a Dialogflow Agent with the Jovo CLI.

Similar to additional elements that go into the skill.json for the alexaSkill object, you can also add specific elements to a Dialogflow agent by adding the whole in the the agent.json file, like so:

googleAction: {
    dialogflow: {
        agent: {
            // Add element to override here
        }
    }
}

For example, you can add a header to your webhook like this:

googleAction: {
	dialogflow: {
		agent: {
			webhook: {
				headers: {
					headerKey: 'headerValue',
				}
			}
		}
	}
}

nlu

For every platform, an nlu element is added for additional options regarding natural language understanding tooling. For example:

googleAction: {
    nlu: {
        name: 'dialogflow',
    }
}

Right now, the built-in alexa nlu for Amazon Alexa and dialogflow for Google Assistant are supported.

Different Locales

For the language model conversion, a lang object can be added to the nlu block:

alexaSkill: {
    nlu: {
      name: 'alexa',
      lang: {
        en: [
          'en-US',
        ]
      }
    }
  },

In the above example, the object specifies the following process for the jovo build command:

  • Access an existing Jovo Language Model file en.json in the /models folder
  • Convert it into an Alexa Interaction Model en-US.json in the /platforms/alexaSkill/models folder

Language Model

In the project.js, you can also add or override specific elements of your language model in the models folder.

Learn more about the Jovo Language Model here.

Override the Invocation Name

Changing the invocation name is especially useful for different stages if you want to make sure you know which version of your voice app you're currently talking to.

For example, you can override the language model for the en-US locale like thos:

languageModel: {
	'en-US': {
		invocation: 'my test app dev',
	}
}

NOTE: Currently, the invocation name can only be specified in the models folder for Alexa Skills. For Google Actions, you need to change them in the Actions on Google Console.

Add Intents and Inputs

You can also add whole intents and inputs and any element that you can find in your e.g. en-US.json in the Jovo models folder.

For example, you can add an intent to specific stages like so:

languageModel: {
	'en-US': {
		intents: [
          	{
            	name: 'WhatEnvIntent',
            	samples: [
					'what\'s the stage',
					'what is the environment',
					'what environment is this deployed to'
            	]
          	}
        ]
	}
}

Deployment

Stages

Jovo allows you to define multiple staging environments like dev, test, and prod in your project.js:

{
    stages: {
        dev: {
            
        },
        test: {
            
        },
        prod: {

        }
    }
}

The elements inside a stage are merged into the elements outside the stage with the ability to override configurations. You can also add a defaultStage element (although not necessary):

{
    defaultStage: 'dev',
    stages: {
        dev: {
            //
        }
    }
}

For an overview of staging examples, take a look at our Staging Examples.

A stage is active when one of the following is true (by default, only the elements outside the stages are used):

  • The default stage is set with defaultStage: '<your-stage>'
  • The stage is set in the environment variables with STAGE=<stage>

Host

Sometimes (for example, if you are using an API Gateway), your endpoint might differ from the destination where the code is located (the host).

You can add a host object to specify where the code is hosted. Right now, lambda is supported to deploy your code to AWS Lambda. This is how you can set it up:

host: {
    lambda: {
        arn: '<your-lambda-arn>',
        askProfile: '<your-ask-cli-profile>', // if left out: "default" profile is used
    }
}

Note: If you're only building an Alexa Skill, der is no need to specify a host with the same information that is already in the endpoint. This is mostly important if you also want to host your Google Action on AWS Lambda and make it accessible through an API Gateway.