{
  "context": "Evaluates whether the solution leverages Luigi's central scheduler and task history features to run a small workflow remotely and report recorded executions. Scoring focuses purely on correct use of the scheduler process, remote task execution, history storage, and retrieval APIs to surface task metadata.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Scheduler init",
      "description": "Starts the Luigi central scheduler using the package's server entrypoint (e.g., luigi.server.run or the luigid CLI) with record_task_history enabled and a valid task history storage path/connection configured.",
      "max_score": 25
    },
    {
      "name": "Remote workflow",
      "description": "Runs the demo dependency chain as Luigi Tasks via the central scheduler using the package's remote execution hooks (e.g., luigi.build with scheduler_url pointing at the scheduler) rather than local scheduling.",
      "max_score": 20
    },
    {
      "name": "History recording",
      "description": "Ensures task completions are persisted through the package's history mechanism (e.g., enabling [task_history] db_connection or DbTaskHistory) so the scheduler UI/DB captures start/end timestamps and statuses.",
      "max_score": 20
    },
    {
      "name": "History fetch",
      "description": "Retrieves recent executions via the scheduler's history RPC/HTTP endpoints (such as /history or /history/by_id) using package clients like luigi.rpc.RemoteScheduler or direct HTTP calls to the scheduler service.",
      "max_score": 20
    },
    {
      "name": "JSON summary",
      "description": "Builds the stdout JSON array directly from the package's history response fields (task id/name, status, start_time, end_time, duration), ordered with most recent tasks first and without custom placeholder data.",
      "max_score": 15
    }
  ]
}