The read_file and write_file methods required the caller to pass an extension that was lower-case and included a leading period in order to determine whether the contents were parquet or csv. This design leaks expectations of the method to the caller (and relies on custom file path parsing to get the extension, which is usually not ideal.) It is also subject to potential errors where the caller passes the incorrect extension.
This revision simplifies the signature to take the file path and then all logic for determining how to read or write is encapsulated in the read_file/write_file methods.
The read_file and write_file methods required the caller to pass an extension that was lower-case and included a leading period in order to determine whether the contents were parquet or csv. This design leaks expectations of the method to the caller (and relies on custom file path parsing to get the extension, which is usually not ideal.) It is also subject to potential errors where the caller passes the incorrect extension.
This revision simplifies the signature to take the file path and then all logic for determining how to read or write is encapsulated in the read_file/write_file methods.