Bun parquet read write native support

[cjkeilig]

I have a bun application that is used to index large volumes of data for RAG (retrieval augmented generation).
Many vector DBs support importing .parquet files for bulk entry insertion.

Currently, I am using LibertyDSNP parquetjs library which has bad memory over- consumption issues. It would be nice for Bun to have native parquet support in the I/O API.