Get item-level recommendations

Returns pairs of user_id and item_id. Depending on recommender_type, items will either be specifically matched to each user based on item-to-item consumption sequences
(recommendation_type == Personalized) or will follow popularity scores (recommender_type == Popularity).

Recommendations can be filtered to a subset of users using the audience_filters input argument.

Recommendations can also be filtered to a set of item properties using the item_filters input argument.
For example: return only recommendations for brand == X and color == yellow