From AST JSON to API Blueprint: References become redundant model data

[rrarunan]

Initial MD:

### Check if a Gist is Starred [GET]
+ Response 200

    [Star][]

Snowcrash Parsed JSON:

    {
              "name": "Check if a Gist is Starred",
              "description": "",
              "method": "GET",
              "parameters": {},
              "headers": {},
              "examples": [
                {
                  "name": "",
                  "description": "",
                  "requests": [],
                  "responses": [
                    {
                      "name": "200",
                      "description": "HAL+JSON representation of Star Resource.\n",
                      "headers": {
                        "Content-Type": {
                          "value": "application/hal+json"
                        },
                        "Link": {
                          "value": "<http:/api.gistfox.com/gists/42/star>;rel=\"self\""
                        }
                      },
                      "body": "{\n    \"_links\": {\n        \"self\": { \"href\": \"/gists/42/star\" },\n    },\n    \"starred\": true\n}\n",
                      "schema": ""
                    }
                  ]
                }
              ]
            }

Matter Compiler -> MD again

### Check if a Gist is Starred [GET]
+ Response 200 (application/hal+json)

    HAL+JSON representation of Star Resource.

    + Headers

            Link: <http:/api.gistfox.com/gists/42/star>;rel="self"

    + Body

            {
                "_links": {
                    "self": { "href": "/gists/42/star" },
                },
                "starred": true
            }

Shouldn't it be a link/reference again? [Star][]

[zdne]

@rrarunan

Good point. The issue is you can't tell from AST. You can do some heuristic and assume that when there is a defined model and a payload 100% matches you can use the model. But then you would be in the opposite problem of having models links everywhere even where they weren't.

With that being said I am not against implementing the heuristics and I will be happy to merge such a PR.

[rrarunan]

Yup, I was thinking the same thing, it would be great when you have the Traits done. Because, the use case for this one was that I was appending random body payload using a simulator in the api blueprint so that my mock server would be dynamic. So, now when compiling back to MD after changing the JSON, I should either generate the body twice or have some kind of heuristic based approach, it could be an option maybe.. I would rather have the links and a smaller MD file than having my body be replicated in several places. Thanks!