feat(cli): add Kaggle dataset integration and Croissant metadata parsing

[closestfriend]

Description

Add Kaggle dataset integration and Croissant (ML Commons) metadata parsing to streamline dataset-to-TOON workflows. This enables users to download Kaggle datasets and convert them to TOON format in a single command.

Features

New CLI flags:

• --kaggle - Treat input as Kaggle dataset slug
• --croissant - Parse input as Croissant JSON-LD metadata
• --file / -f - Select specific file from multi-file datasets

Usage examples:

# Download Kaggle dataset and convert to TOON
toon username/dataset-name --kaggle --stats

# Select specific file from dataset
toon username/dataset-name --kaggle --file data.csv

# Parse Croissant metadata to see schema
toon metadata.json --croissant

New Python API:

from toon import download_dataset, parse_croissant, csv_to_records

# Download and process Kaggle dataset
files = download_dataset("username/dataset-name")
csv_file = find_best_csv(files)

# Parse Croissant metadata
info = parse_croissant(metadata)
print(info['schema'])

Implementation

New module toon/kaggle.py provides:

• download_dataset() - Download Kaggle datasets via kaggle CLI
• find_best_csv() - Heuristic selection of main data file
• csv_to_records() - CSV to list[dict] conversion
• parse_croissant() - Extract schema from Croissant JSON-LD
• croissant_to_summary() - Generate human-readable summaries
• is_kaggle_slug() - Detect Kaggle dataset slug format

All imports are optional - gracefully degrades if kaggle package is not installed.

Type of Change

• New feature (non-breaking change which adds functionality)

Testing

• All tests pass
• Added 12 new tests for Kaggle integration
• Tested manually with real Kaggle datasets

Checklist

• Code follows the project's style guidelines
• Self-review completed
• Documentation updated (CLI help, docstrings)
• No new warnings or errors introduced