Description
Hello! I love the codegen so far, we use it in our large scale serverless project in the following scenario:
Typescript Backend with TSOA Controllersgenerated OpenApi3 Spec with rich documentation and validationsgenerated frontend typescript client (and other language clients)
Generated code matches the initial structure really well, preserving complex union types (Dog | Cat), understanding interface extension, etc. What it doesn't (obviously) do, is help with validation rules such as min/maxLengths.
It would be non-intrusively helpful if we could optionally preserve the JSDoc structure that we used in the original code to generate OpenAPI rich spec. Or, more closely, generate rich JSDoc from the OpenAPI descriptions.
Our initial annotations (as well as the "target" JSDoc we'd like to have in the generated client).
/**
* @minLength 0
* @maxLength 150
*/
export type Address_Note = string;
which corresponds to the following in the OpenApi.json
"Address_Note": {
"type": "string",
"minLength": 0,
"maxLength": 150
},
If data from OpenAPI spec, such as example, minLength, pattern, description from both Schemas and Operations, could be mapped in the output to JSDoc comments above interfaces, services, properties, etc, we could preserve rich information about the proper data structure and see it in IntelliSense-like way in the IDEs.