Update parquet structure and add database submodule

[stephprince]

Fix #154. Some updates to the parquet structure to help with querying and some general refactoring to make it easier to generate the database files locally without using the app.

Creating the new submodule makes it harder to see the changes so here is a summary:

• pulled out dataclasses and parquet processing from app, moved to different module to run locally more easily outside of the app
• added version filtering before concatenation to avoid issues with older results structures
• added "machine_id" column to machine.parquet file
• add new columns for parameter_case fields ("parameter_case_name", "parameter_case_https_url", "parameter_case_object_name", "parameter_case_slice_range").

Should environment parquet file also have a version column in case we add additional fields at any point?

[CodyCBakerPhD]

@stephprince Feel free to change whatever you need - my first attempt was just to get something roughly in the ballpark of some kind of usable shape