How to best implement an output that depends on an input ?

Use Case:

We have an API that has a JSON output. An input parameter allows to filter the returned JSON fields

The only way to implement this as far as I know is to pass the formatting parameter into the server logic which will then have to encapsulate it in the return type with an appropriate codec.

We have the same problem if we want to recover the input url to build output headers (example Link prev/next headers).

It's quite verbose, and does it really have its place in a "logic" method?

Example:

myEndpoint
  .in(EndpointInputs.searchRequest)
  .in(EndpointInputs.jsonFilterRequest)
  .in(EndpointInputs.extractUri)
  .out(
        headers
          .and(statusCode)
          .and(jsonBody[Seq[FilteredJson[T]]])
          .map(PagedSearchResult.mapping[T])
   )
   .serverLogic { case (searchRequest, jsonFilterRequest, uri)
     myRepository.findPageable(searchRequest)
      .map {
        PagedSearchResult(
          maxSize = maxSize,
          uri = uri,
          jsonFilterRequest = jsonFilterRequest
        )
      }
   }

Start by reviewing the endpoint input and output definitions shown in the issue, especially jsonFilterRequest and extractUri. Determine whether the API can expose request-derived values to output encoding and header generation without passing them through server logic. Done means a settled approach for filtered JSON fields and request-based Link headers.