{
  "context": "Evaluates how well the solution uses fhir-py to build paired sync and async clients with matching configuration. Focuses on using the library's search and resource helpers for patient retrieval and observation creation across both execution modes.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Sync setup",
      "description": "Creates SyncFHIRClient with the provided base URL, passes the token to the authorization parameter, applies extra_headers, and forwards client_config into requests_config instead of crafting custom HTTP calls.",
      "max_score": 20
    },
    {
      "name": "Async setup",
      "description": "Creates AsyncFHIRClient with the same base URL and headers, forwarding client_config into aiohttp_config, and uses the async client directly rather than wrapping the sync path.",
      "max_score": 20
    },
    {
      "name": "Patient fetch",
      "description": "Uses client.resources('Patient') search helpers (search/get/first/fetch) to retrieve the patient record and derive id/name/birthDate from the returned resource or serialize(), avoiding manual HTTP or hardcoded data.",
      "max_score": 20
    },
    {
      "name": "Observation create",
      "description": "Builds an Observation via client.resource('Observation', ...) or resources('Observation') with subject=client.reference('Patient', id=patient_id), code, and valueQuantity, then persists it with create/save/patch to obtain the server-assigned id.",
      "max_score": 25
    },
    {
      "name": "Async parity",
      "description": "Async methods mirror the sync logic with AsyncFHIRClient operations (resources/search/create/fetch) and await the calls, not by invoking sync methods through executors or blocking adapters, and return the expected id/field shapes.",
      "max_score": 15
    }
  ]
}