Hi team. How could I manipulate an S3 IO manager to save partitioned files in a hive structure

year=2023/month=08/day=23/raw.parquet

?

is the file path determined within the asset function like from the current date, or info in the data? or is it something you want to set as configuration?

this is my asset:

@asset(
    compute_kind="python",
    name="raw_matrix_accounts_receivable",
    key_prefix=["omie", "matrix"],
    io_manager_key="parquet_s3_io_manager",
    partitions_def=DailyPartitionsDefinition(start_date="2023/06/22", fmt="%Y/%m/%d", timezone=TZ, hour_offset=11),
)
def raw_matrix_accounts_receivable(context: OpExecutionContext, matrix_omie_client: OmieAPI) -> pd.DataFrame:
    data: list = matrix_omie_client.list_all_accounts_receivable()
    df = pd.json_normalize(data)
    return df

I then have an IO manager similar to the

PickledObjectS3IOManager(UPathIOManager)

here. I want to edit the method that determines the path so that it saves partitioned assets in a hive structure. Currently, it saves as

2023/08/23/raw.parquet

in the example in my first message.