{
  "context": "This criteria evaluates how well the engineer uses ConnectorX to implement a memory-efficient database loading solution. The focus is on leveraging ConnectorX's zero-copy architecture and performance optimization features for memory management and speed comparison.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Uses read_sql function",
      "description": "Implementation uses connectorx.read_sql() to load data from the database in the optimized approach",
      "max_score": 25
    },
    {
      "name": "Uses pandas.read_sql baseline",
      "description": "Implementation uses pandas.read_sql() as the standard/baseline approach for performance comparison",
      "max_score": 15
    },
    {
      "name": "Memory tracking implementation",
      "description": "Implementation correctly tracks memory usage during data loading operations using psutil (e.g., psutil.Process().memory_info()) and returns peak memory in megabytes",
      "max_score": 20
    },
    {
      "name": "Timing measurement accuracy",
      "description": "Implementation accurately measures execution time for both standard and optimized approaches using time.time() or time.perf_counter(), calculating the difference correctly",
      "max_score": 15
    },
    {
      "name": "Returns correct data structures",
      "description": "load_dataset returns a tuple of (pandas.DataFrame, float) and compare_performance returns a dict with the three required keys: 'standard_time_seconds', 'optimized_time_seconds', 'speedup_factor'",
      "max_score": 15
    },
    {
      "name": "Speedup calculation",
      "description": "Implementation correctly calculates speedup_factor as standard_time / optimized_time in the compare_performance function",
      "max_score": 10
    }
  ]
}