{
  "context": "Evaluates whether the solution leverages fhirpy to run paginated encounter searches with forward/reverse includes, keep raw bundles intact, and expose relationships through reference helpers instead of manual HTTP or string handling.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Paginated search",
      "description": "Search uses AsyncFHIRSearchSet (or equivalent) with limit(page_size) and async iteration or fetch_all() to traverse all pages rather than issuing manual page requests.",
      "max_score": 20
    },
    {
      "name": "Forward include iterate",
      "description": "Search adds include(..., iterate=True) for patient/service provider relationships so those related resources arrive alongside encounters.",
      "max_score": 20
    },
    {
      "name": "Reverse include",
      "description": "Search adds revinclude(..., iterate=True) for observations (or similar) referencing the encounter so reverse-linked resources are pulled in without extra queries.",
      "max_score": 20
    },
    {
      "name": "Raw bundle",
      "description": "Results are retrieved via fetch_raw() to preserve Bundle entries (including included/reverse-included resources) instead of using fetch/fetch_all and losing those entries.",
      "max_score": 20
    },
    {
      "name": "Reference objects",
      "description": "Summaries normalize relationships with fhirpy helpers such as resource.to_reference() or client.reference() so patient, service provider, and observation links are FHIRReference instances rather than plain ids/strings.",
      "max_score": 20
    }
  ]
}