Add DeltaLake source support

[danielbeach]

Add a new method called read_deltalake(). Simply read the Delta Lake table and get the list of parquet files for that Delta Lake table so we can reuse ParquetDataSet. (doing this by calling dt.files() which returns a list of parquet files).

[wangrunji0408]

As far as I know, Delta Lake is not simply a list of parquet files, right? I think we need a dedicated DeltaLakeDataset to handle them properly.

[danielbeach]

As far as I know, Delta Lake is not simply a list of parquet files, right? I think we need a dedicated DeltaLakeDataset to handle them properly.

Yes, but we are using the deltalake package, and specifically an instance of a DeltaTable to give a list of the parquet files that currently make up the current version of the Delta Lake table via files() method they provide. I mean in theory we could add another dependecy of polars or daft or something else to return the Delta Lake as a Dataframe and the dump it to an arrow table, but that seems more indirect.