Rerank a list of item IDs for a given user using a trained ML model.

This endpoint sends the user and items to a trained ranking model, which returns the same items in a personalized order based on predicted relevance.

Flow:

1. Provide user_id and item_ids in the request body
2. The service reranks items using the trained model for the given model_name
3. Response returns the same item IDs in a new order (personalized ranking)

Available Models:

• polymarket-rerank-v1 — Polymarket reranking model trained on trading activity data

Error Handling: Errors are returned with HTTP 200 and an error field in the response body (rather than HTTP error codes). This includes:

• Missing user_id or item_ids
• Unknown model_name
• Service errors (model not found, invalid input, network issues)