Hello!

First the problem: I have a login mutation. It returns the user object on successful login. Originally the user object simply contained built in types (string, boolean, ect) so no extra resolvers aside from the login resolver where needed. However I then added another field called workspaces to the user which has its own resolver.

  object :user do
    field(:first_name, :string)
    ...

    field :workspaces, list_of(:workspace) do
      resolve(&ScribeWeb.Resolvers.Workspaces.get_workspace/3)
    end
  end

get_workspace grabs the user_id from the resolution context and uses this to fetch that users workspaces. This works fine if the user is logged in already and the user id is already on the context. However when logging in this is not the case.

The question: Is it possible to modify the resolution context and add the user id in the login resolver, so that when the get_workspace resolver is called it has access to the now logged in users id?

My current workaround fix: I created another user object but this time the workspaces field didn’t have a resolver. Inside the login resolver I now make another call to the db to get the users and update the user struct with these results manually.

  def login(_parent, %{login_input: args}, %{context: context}) do
    %{email: email, password: password} = args

    Scribe.Accounts.authenticate(email, password)
    |> case do
      {:ok, user} = result ->
        workspaces = Scribe.Workspaces.get_workspace(%{user_id: user.id})
        updated_user = %{user | workspaces: workspaces}
        {:ok, updated_user}

      error ->
        error
    end
end

This just seems a little crappy because I now need to maintain two almost identical user objects in my schema, one with a workspace resolver and one without.

Thanks in advance for the help