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:
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:
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. Deployment of project files is currently only supported for Alexa Skills and Google Actions.
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
Find out more about configuration for your Alexa skill here.
googleAction
Find out more about configuration for your Google Action here.
nlu
For every platform, an nlu
element is added for additional options regarding natural language understanding tooling. For example:
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:
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.
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 this:
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:
Deployment
Stages
Jovo allows you to define multiple staging environments like dev
, test
, and prod
in your project.js
:
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):
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:
Note: If you're only building an Alexa Skill, there 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.