check_dataset_fields.py

#!/usr/bin/env python3
"""
Check the actual fields in the dataset.
"""

from datasets import load_dataset

def main():
    print("Loading dataset...")
    dataset = load_dataset("allenai/Dolci-Instruct-SFT-Tool-Use", split="train")

    print(f"Total samples: {len(dataset)}")
    print(f"\nDataset features: {dataset.features}")
    print(f"\nColumn names: {dataset.column_names}")

    # Check first sample
    print("\n=== First sample ===")
    sample = dataset[0]
    print(f"Keys: {sample.keys()}")
    for key in sample.keys():
        value = sample[key]
        if isinstance(value, str):
            print(f"\n{key}: {value[:200] if len(value) > 200 else value}")
        elif isinstance(value, list):
            print(f"\n{key} (list with {len(value)} items):")
            if value:
                print(f"  First item: {value[0]}")
        else:
            print(f"\n{key}: {value}")

    # Check sample 100
    print("\n\n=== Sample 100 ===")
    sample = dataset[100]
    for key in sample.keys():
        value = sample[key]
        if isinstance(value, str):
            print(f"\n{key}: {value[:200] if len(value) > 200 else value}")
        elif isinstance(value, list):
            print(f"\n{key} (list with {len(value)} items):")
            if value:
                print(f"  First item: {value[0]}")
        else:
            print(f"\n{key}: {value}")

if __name__ == "__main__":
    main()