Mapping template, Velocity Language & JsonPath: Guide to Api Gateway mapping templates mechanism

Amazon’s API Gateway provides the facilities to map an incoming request’s payload to match the required format of an integration backend. It uses the concept of “models” and “mapping templates” to specify the mapping between the client payload and the server payload.

The API Gateway mapping templates are used to transform an incoming payload into a different format. API Gateway allows you to define input mapping templates, for mapping the incoming request from a client to a server format, and output mapping templates, for mapping the outgoing response from the server to a client format.

Let us all remember that the mappings are defined using the Velocity Template Language combined with JSONPath expressions.

Velocity is a Java-based templating engine that uses a template language to reference objects in Java code. Velocity can be used to generate web XML, HTML or SQL from templates, for example. Here, we will use it to define the mapping between an input and output model.

A Velocity Template Language (VTL) template is composed of statements which begin with the # character and a followed by a directive. Also, it has references, which begin with the $ character. More generally, references begin with $ and are used to get or set something. Directives begin with # and are used to do something.

We can use the statement above to define an HTML template that sets a reference to a value, and retrieves that reference.

  
  #set( $foo = "Velocity" )
  Hello $foo World!
  

For more on VTL, consult the user guide.

In API Gateway mapping templates, the input to your Velocity template is defined by the reference variable $input. To access particular portions of the input JSON document within the template, use JSONPath.

JSONPath provides an XPath-like method of referring to a JSON structure. IN JSONPath, the abstract name $ refers to the root-level object. From this root-level object, you can use dot-notation to select more restrictive portions of the document. 

For example, to reference the playerId of the player model, you would use the following JSONPath expression.

$.playerId

or, you can use a bracket-notation to perform the same selection.

#['playerId']

More complex selections can be done using * for a wildcard, @ to refer to the current node in the document, array indexing, or recursive descent.

Now, we can put together a mapping template for our Player model. This template will just copy the input to the output with new names for field properties. In particular, it will convert an input document of

{
    "id": "AJamal",
    "alias": "Aha Jamal",
    "displayName": "Anwar Jamal Faiz",
    "profilePhotoUrl": "https://api.example.com/player/AJamal/avatar.png"
}

to an output document of

{
    "id": "AJamal",    "alias": "Aha Jamal",    "name": "Anwar Jamal Faiz",    "photo": "https://api.example.com/player/AJamal/avatar.png"}

The mapping to perform the above transformation is specified below, with some inline VTL comments added.

## Set the variable inputRoot to the root JSON document
#set($inputRoot = $input.path('$')) {

    ## Assign output properties to input properties
    "id": "$inputRoot.playerId",
    "alias": "$inputRoot.alias",
    "name": "$inputRoot.displayName",
    "photo": "$inputRoot.profilePhotoUrl"
}

Enjoy!!

Json schema validation for Api Gateway request

Amazon’s API Gateway provides the facilities to map an incoming request’s payload to match the required format of an integration backend. It uses the concept of “models” and “mapping templates” to specify the mapping between the client payload and the server payload.

A model defines the structure of the incoming payload using JSON Schema. The model is an optional, but not required, piece of API Gateway. By providing a model, you make it easier to define the upcoming mapping template that actually does the transformation between the client and server.

For example, we can define a incoming Player model using the following JSON payload.

{
    "id": "AJamal",
    "alias": "Aha Jamal",
    "displayName": "Anwar jamal Faiz",
    "photoUrl": "https://api.example.com/player/AJamal/avatar.png"
}

JSON Schema is a vocabulary allowing you validate JSON documents. The particular JSON payload being validated is called an instance, and the document describing what a valid payload looks like is called the schema.

JSON Schema is used to describe a JSON document, so it helps to understand what, exactly, a JSON document is composed of these primitives:

• objects: {"field1": "value1", "field2": "value2"}
• arrays: ["first", "second", "third"]
• numbers: 42 or 3.1415926
• strings: "Lorem ipsum dolor sit amet"
• booleans: true or false
• null: null

The responsibility of JSON Schema is to describe a JSON document built from the these primitives.

JSON Schema is itself specified using JSON with predefined semantics for keywords and values. Going back to our Player example, we can specify the JSON Schema for the incoming document as follows:

{
    "$schema": "http://json-schema.org/draft-04/schema#",
    "type": "object",
    "properties": {
        "id": { "type": "string" },
        "alias": { "type": "string" },
        "displayName": { "type": "string" },
        "photoUrl": { "type": "string" }
    }
}

This JSON Schema follows the Draft 4 specification and simply declares the types expected for each field in the request. 

For more details consult this guide.