Read files with polars

[Andilun]

Changes

The PseudoData.from_file function will now use polars when reading parquet formatted files. This should improve performance when reading parquet.

This resolves issue#227

Reason for not using polars for 'csv' files

I have opted not to use Polars for reading CSV files because it would necessitate us to handle a convertion between Polars 'dtypes' and Pandas 'dtype'.

Currently, all the file formats we support reading can either have data types specified with the dtype keyword argument or derive the schema from the file. We should avoid requiring users to define data types differently for CSV files.

dtypes vs dtype

Pandas.read_csv dtypes:dtype or dict of {Hashabledtype}
Polars.read_csv dtype: Mapping[str, PolarsDataType] | Sequence[PolarsDataType] | None = None

QOL changes:

• Add example of how to ovveride service url and token

[mmwinther]

Looks good 👍