feat: support for denormalized response from server

[jimmycallin]

• I have added automatic tests where applicable
• The PR title is suitable as a release note
• The PR contains a description of what has been changed
• The description contains manual test instructions

Changes

Adds support for denormalized response from server

Test

Make sure the response is denormalized, that decoding is skipped as expected, and that there cannot be any cyclical references in the output.

[gismya]

Looks good.

[ffMathy]

Curious. What does this do? Seems interesting.

[jimmycallin]

Curious. What does this do? Seems interesting.

by default our API returns results in a normalized manner, e.g. this query:

[
  {
    "action": "query",
    "expression": "select project.full_name from Task limit 2"
  }
]

would result in this:

[
  {
    "action": "query",
    "data": [
      {
        "__entity_type__": "Task",
        "id": "19d33e8e-8bff-11eb-b5d3-c2ffbce28b68",
        "project": {
          "__entity_type__": "Project",
          "id": "33b1e354-8b3d-11eb-8e99-c2ffbce28b68",
          "full_name": "Sync"
        }
      },
      {
        "__entity_type__": "Task",
        "id": "2b5dda84-8b4f-11eb-b695-c2ffbce28b68",
        "project": {
          "__entity_type__": "Project",
          "id": "33b1e354-8b3d-11eb-8e99-c2ffbce28b68",
          // NOTE - NO FULL_NAME HERE
        }
      }
    ],
    "metadata": {
      "next": {
        "offset": 2
      }
    }
  }
]

We are then in the javascript client recursively merging all entities with same ID to ensure all attributes are available where expected. This can in some cases lead to cyclical object references, meaning that we cannot JSON.stringify them, and things like hashing the result become a problem.

We have just added support for a denormalized response in the api which will return the result in a more expected manner:

[
  {
    "action": "query",
    "data": [
      {
        "__entity_type__": "Task",
        "id": "19d33e8e-8bff-11eb-b5d3-c2ffbce28b68",
        "project": {
          "__entity_type__": "Project",
          "id": "33b1e354-8b3d-11eb-8e99-c2ffbce28b68",
          "full_name": "Sync"
        }
      },
      {
        "__entity_type__": "Task",
        "id": "2b5dda84-8b4f-11eb-b695-c2ffbce28b68",
        "project": {
          "__entity_type__": "Project",
          "id": "33b1e354-8b3d-11eb-8e99-c2ffbce28b68",
          // NOTE - FULL NAME HERE AS EXPECTED
          "full_name": "Sync"
        }
      }
    ],
    "metadata": {
      "next": {
        "offset": 2
      }
    }
  }
]

this in turn means we don't need to recursively merge all entities, and the result is thus serializable.

Hope this explains it!

[ffMathy]

Thank you for that elaboration! We didn't know this, and we'd definitely prefer the non-circular one.

Very nice feature! Thanks for keeping the library up to date regularly! 😍