By default, soorgeon generates products as pickle files since we don't know what data type the outputs are. However, pickle files aren't a good choice. Maybe we could have some options that based on the name of the variable, we could use a different format, things like:
if prefix is df, then use parquet/csv
if prefix is X, Y, x, or y, then use numpy binary format
By default, soorgeon generates products as pickle files since we don't know what data type the outputs are. However, pickle files aren't a good choice. Maybe we could have some options that based on the name of the variable, we could use a different format, things like: