def load_dataset_from_json()

in evalbench/dataset/dataset.py [0:0]

• 13 lines of code
• 4 McCabe index (conditional complexity)

def load_dataset_from_json(json_file_path, config):
    input_items = []
    all_items = load_json(json_file_path)
    dataset_format = config.get("dataset_format", "evalbench-standard-format")
    if dataset_format == "evalbench-standard-format":
        input_items = load_dataset(all_items, config)
    elif dataset_format == "bird-standard-format":
        input_items = load_dataset_from_bird_format(all_items)
    else:
        raise ValueError("Dataset not in any of the recognised formats")
    totalEntries = sum(len(input_items.get(q, [])) for q in ["dql", "dml", "ddl"])
    logging.info(f"Converted {totalEntries} entries to EvalInput.")

    return input_items