🔴🔵 Create a manual intake (without image-based prediction).

Creates an intake for the specified APIUser without using an image-based prediction. This is an alternative to the recognition endpoints.
Multiple dishes can be specified in the intake, with only a valid id for the dish being necessary (A list of valid classes can be obtained by calling /dataset/dishes).
Optionally, quantity, levels and ingredients can be specified for each dish in the intake. If provided, these parameters will be applied in the following order:

1. Ingredients: Use the given list of ingredients for the dish's recipe instead of using the default recipe for the dish id.
2. Levels: Scale the level of salt, sugar and oil in dishes. The weights for ingredients specified in the ingredients field may be altered if you include this parameter.
3. Quantity: If provided the ingredients quantities will be scaled proportionally so that they sum up to add the given quantity.