Parquet & JSONL Packaging

Produce Parquet and/or JSONL output for a dataset, optionally partitioned and compressed.

When to invoke

• User wants a Parquet file for analytics (DuckDB, Spark, pandas at scale, ML training).
• User wants JSONL for streaming or HF Datasets native ingestion.
• Preparing for upload via the hf-dataset-push skill.

Procedure