{
  "context": "This criteria evaluates how well the engineer uses @gradio/client's session management features to build a session-aware prediction tracker. The focus is on proper utilization of session hashes, the Client.connect() method with session configuration, and session state handling.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Client.connect() usage",
      "description": "Uses Client.connect() method to establish connection to the Gradio application, properly handling the async connection and returning a Client instance.",
      "max_score": 20
    },
    {
      "name": "Custom session hash",
      "description": "Implements custom session hash support by passing session_hash parameter to Client.connect() when provided in config, allowing users to specify their own session identifiers.",
      "max_score": 25
    },
    {
      "name": "Session hash retrieval",
      "description": "Retrieves and exposes the session hash from the connected Client instance (via client.session_hash property or similar) to enable session identification.",
      "max_score": 15
    },
    {
      "name": "predict() method usage",
      "description": "Uses the client.predict() method to make predictions, properly passing endpoint identifier and input data, and handling the returned promise.",
      "max_score": 15
    },
    {
      "name": "State parameter handling",
      "description": "Handles optional state parameter in predictions by properly passing it to the predict() method when provided, enabling stateful interactions.",
      "max_score": 10
    },
    {
      "name": "reconnect() method usage",
      "description": "Uses the client.reconnect() method to verify session validity and returns the connection status ('connected', 'broken', or 'changed').",
      "max_score": 15
    }
  ]
}