Experimental s3 and parquet support

[jonmmease]

This PR adds initial support for loading input files from s3, and for parquet support. Includes:

• Use object_store::aws::AmazonS3Builder to support loading json, csv, arrow, and parquet files from s3. Credentials are pulled from environment variables per https://docs.rs/object_store/latest/object_store/aws/struct.AmazonS3Builder.html#method.from_env.
• Support loading parquet files from the local file system and from s3 (no support for http/https yet).

For testing, I added a pixi dependency on minio and added the pixi start minio task that start's minio with pre-determined credentials and loads sample data.

[jonmmease]

I wasn't able to get the minio-based test working in GitHub actions, so flagging behind and environment variable for now