In this example we will create a mashling gateway recipe that conditionally invokes a PetStore API. The recipe will either register or retrieve pet data. The scenario assumes that only pets with a category name of DOG or CAT are supported. If an attempt is made to register any other kind of pet the gateway responds with an error message without hitting the PetStore backend.
Open an editor and define event triggers like the following:
{
"mashling_schema": "1.0",
"gateway": {
"name": "MyProxy",
"version": "1.0.0",
"description": "This is a simple proxy.",
"triggers": [
{
"name": "MyProxy",
"description": "Animals rest trigger - PUT animal details",
"type": "github.com/TIBCOSoftware/mashling/ext/flogo/trigger/gorillamuxtrigger",
"settings": {
"port": "9096"
},
"handlers": [
{
"dispatch": "Retrieve",
"settings": {
"autoIdReply": "false",
"method": "GET",
"path": "/pets/{petId}",
"useReplyHandler": "false"
}
},
{
"dispatch": "Register",
"settings": {
"autoIdReply": "false",
"method": "PUT",
"path": "/pets",
"useReplyHandler": "false"
}
}
]
}
],
There are two handlers created for this HTTP trigger. One for retrieving the pet data through an HTTP GET operation and the other for registering a pet through an HTTP PUT operation. For the details of the handler configuration, see the documentation at https://github.com/TIBCOSoftware/mashling/tree/master/ext/flogo/trigger/gorillamuxtrigger
Add a dispatches section:
"dispatches": [
{
"name": "Retrieve",
"routes": [
{
"steps": [
{
"service": "GetPet",
"input": {
"method": "GET",
"pathParams.id": "${payload.pathParams.petId}"
}
}
],
"responses": [
{
"if": "GetPet.response.body.status != 'available'",
"error": true,
"output": {
"code": 404,
"data": {"error": "Pet is not available."}
}
},
{
"if": "GetPet.response.body.status == 'available'",
"error": false,
"output": {
"code": 200,
"data": "${GetPet.response.body}"
}
}
]
}
]
},
{
"name": "Register",
"routes": [
{
"if": "payload.content.category.name == 'DOG' || payload.content.category.name == 'CAT'",
"steps": [
{
"service": "PutPet",
"input": {
"inputs.content": "${payload.content}"
}
}
],
"responses": [
{
"if": "PutPet.response.outputs.code != 200",
"error": true,
"output": {
"code": 400,
"data": { "error": "Pet is not registered."}
}
},
{
"if": "PutPet.response.outputs.code == 200",
"error": false,
"output": {
"code": 200,
"data": "${PutPet.response.outputs.data}"
}
}
]
},
{
"steps": [
{
"service": "InvalidAnimal",
"input": {
"parameters.content": "${payload.content}"
}
}
],
"responses": [
{
"error": false,
"output": {
"code": 400,
"data": {"error": "${InvalidAnimal.response.result.msg}"}
}
}
]
}
]
}
],
Both Retrieve and Register routes have their own steps and their own response handling defined. Note that Register has a step which is executed only when a dog or cat is registered. If no condition is defined for a set of steps it is considered the default route. Only one route is executed. Thus, in the Register route, the InvalidAnimal service is executed only when the request is for an animal other than a dog or a cat.
Add the services section below, which defines each service reference used in the dispatch section above:
"services": [
{
"name": "GetPet",
"description": "Make GET calls against a remote HTTP service using http endpoint.",
"type": "http",
"settings": {
"url": "http://petstore.swagger.io/v2/pet/:id"
}
},
{
"name": "PutPet",
"description": "Make PUT calls against a remote HTTP service using a Flogo flow.",
"type": "flogoFlow",
"settings": {
"reference": "github.com/TIBCOSoftware/mashling/lib/flow/RestTriggerToRestPutActivity.json"
}
},
{
"name": "InvalidAnimal",
"description": "Message for an invalid animal",
"type": "js",
"settings": {
"script": "result.msg = 'Unsupported animal: ' + parameters.content.category.name;"
}
}
]
}
}
The services section above illustrates 3 different types of service suported. The first service which handles the GET request invokes https endpoint directly. The second service invokes a flogo flow to execute the pet store backend API to handle the PUT request. InvalidAnimal service executes a javascript service to produce a message that the request is invalid. This avoids invoking the backend API if the request is invalid.
Save the recipe file created and validate it with mashling-cli in a terminal:
mashling-cli -validate recipe_tutorial.json
Start the recipe with mashling-gateway:
mashling-gateway -c recipe_tutorial.json
In another terminal, register a pet with
curl -X PUT "http://localhost:9096/pets" -H "Content-Type: application/json" -d '{"category":{"id":1,"name":"DOG"},"id":16,"name":"Olive","photoUrls":["unavailable"],"status":"sold","tags":[{"id":76543,"name":"Olive"}]}'
Now, retrieve the pet data with
curl -X GET http://localhost:9096/pets/16
The pet data just registered should be returned.
Try registering a pet with unsupported type
curl -X PUT "http://localhost:9096/pets" -H "Content-Type: application/json" -d '{"category":{"id":3,"name":"BIRD"},"id":17,"name":"Birdie","photoUrls":["unavailable"],"status":"sold","tags":[{"id":87654,"name":"Birdie"}]}'
An error message is returned like the following:
{
"error": "Unsupported animal: BIRD"
}