{
  "context": "This evaluation assesses how effectively the engineer uses Kedro's data versioning and time-travel capabilities to implement a pipeline with versioned datasets that enable reproducibility and historical data access.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Versioned Dataset Configuration",
      "description": "Uses AbstractVersionedDataset or a versioned dataset implementation (e.g., VersionedPickleDataset, VersionedCSVDataset) to configure versioned datasets in the catalog. Each of the three datasets (cleaned_data, customer_metrics, final_report) should be properly configured with versioning enabled.",
      "max_score": 30
    },
    {
      "name": "DataCatalog Creation",
      "description": "Creates a DataCatalog instance using DataCatalog() or DataCatalog.from_config() with versioned dataset configurations. The catalog should properly manage versioned datasets and enable version tracking.",
      "max_score": 15
    },
    {
      "name": "Pipeline Construction",
      "description": "Uses the Pipeline class and node() function to construct a three-node pipeline that connects preprocess_data -> calculate_metrics -> generate_report with proper input/output dataset naming.",
      "max_score": 15
    },
    {
      "name": "Version-Controlled Execution",
      "description": "Demonstrates running the pipeline with version control using session.run() or runner.run() with save_version parameter to create versioned datasets, or relies on automatic timestamp generation for versioning.",
      "max_score": 15
    },
    {
      "name": "Latest Version Loading",
      "description": "Demonstrates loading the latest version of a dataset using catalog.load() without specifying a version, which should retrieve the most recent versioned data.",
      "max_score": 10
    },
    {
      "name": "Specific Version Loading",
      "description": "Demonstrates loading a specific historical version of a dataset using either the Version namedtuple with catalog.load() or session.run() with load_versions parameter to access time-travel capabilities.",
      "max_score": 15
    }
  ]
}