I’m currently working with Beanie ODM, FastAPI and MongoDB and facing a challenge with optimizing the way references are printed in my JSON responses.

Current Situation: For a Booking object, my current output looks like this:

{
  "_id": "66691ce3f75184ad17b7abd9",
  "account": {
    "id": "6630082da4ecb6802b241748",
    "collection": "accounts"
  },
  "hotel": {
    "id": "6660c3bb318e44905a3cff19",
    "collection": "hotels"
  },
  "arrival_date": "2024-06-12T00:00:00",
  "departure_date": "2024-06-13T00:00:00",
  "language": "en",
  "observations": "",
  "approved": false,
  "created_at": "2024-06-12T03:58:27.887000",
  "updated_at": "2024-06-12T03:58:27.887000"
}

Desired Output: I want to format the output so that it includes the collection prefix directly in the reference fields, like this:

{
  "id": "booking_66691ce3f75184ad17b7abd9",
  "account": "account_6630082da4ecb6802b241748",
  "hotel": "hotel_6660c3bb318e44905a3cff19",
  "arrival_date": "2024-06-12T00:00:00",
  "departure_date": "2024-06-13T00:00:00",
  "language": "en",
  "observations": "",
  "approved": false,
  "created_at": "2024-06-12T03:58:27.887000",
  "updated_at": "2024-06-12T03:58:27.887000"
}

Currently, to achieve this, I am fetching the booking details and then formatting each booking object individually. This approach is not optimal, especially when dealing with a large number of bookings (e.g., 10,000 bookings). It requires fetching all bookings and then iterating through each one to apply the formatting function, which is resource-intensive.

I’m considering exploring custom fields in Beanie or creating a new field class that inherits from the existing Field type (Link) to handle this more efficiently.