Author: Jeetan Madaan
Preface
Let’s try to understand the API Design & Development Cycle using Anypoint Platform, and the target audience is developers & architects in the organization.
Developers & Architects in the mule ecosystem refers to the Mulesoft Training course “Mulesoft.U Development Fundamentals” and this article focuses on “how to approach” an API Design & Development Cycle.
User Story
Create a Restful API that uses APIKit and MongoDB connector to retrieve and create customer documents.
Design with API Designer
When we as a team gets the product specifications, there are design discussions, working sessions, and in that phase, one has to get hold of Design Center and start prototyping the API specs.
Resources: Entities modelled around business domains are mapped with URI paths in a RAML file. Let’s say we have to make system calls to retrieve & create customers, then we will be defining a resource mapped to a /customers URI.
Methods/Verbs: Restful APIs are modeled with Standard Http Verbs, so use GET to retrieve customers and POST to create customers. CRUD APIs should also support Put and Delete.
Types: RAML supports validating the request & responses by defining inline types & library types, and modeled API shows the inline type of Contact.
Query Params, Responses: Use query params to filter collection based resources and responses map to status code output in the desired mime / type with application/json usually the norm for Restful APIs.
Connect with MongoDB
- Create an account on free-tier MongoDB Atlas Cloud using your email address and create a customers collection in CustomerDB.
- Create a user as ‘testuser’ and allow it to read/write any database.
- Install a mongo shell and connect to MongoDB with the connection string.
mongo mongodb+srv://username:password@clusterURL/customerDB
- Create records in the customers collection with the insert command from mongo shell.
Develop with Anypoint Studio
There may be few ways to scaffold a mule application, but for API Development, an initial step is to import / or use the API RAML spec created in the design phase in a mule studio application.
RAML, though being a specification of the API, has support for server side code generators in different languages like NodeJS, Java, Python, Mule.
Apikit: It is a declarative toolkit that parses the RAML API definition, and creates an HTTP Requester to expose the API Resources & APIKit Router that routes the request to auto-generated flow templates. It also creates Error Handling in the main flow that references exception types like [BAD_REQUEST, NOT_FOUND, METHOD_NOT_ALLOWED, NOT_ACCEPTABLE ,UNSUPPORTED_MEDIA_TYPE, NOT_IMPLEMENTED].
Connectors: In an integration scenario, there is always a target system, and multiple enrichment components. Connectors are reusable extensions to mule runtime and imported as a maven dependency in the mule application. It enables you to integrate with third party APIs and databases. Connectors have operations, namespace & connection configuration based on the target application.
Test with Rest Client
Initial Data Setup in Customer Collection has 3 documents.
Retrieve documents from customer collection with restful get call on customer prototype api
Create document in customer collection with restful post call on customer prototype api
