Calculate the five categories for each state

[natemcintosh]

This pull request includes changes to the cfa_rt_postprocessing/main_functions.py file to add functionality for calculating and saving categories from sample data. The most important changes include the addition of a new function for calculating categories and the integration of this function into the existing workflow.

New functionality:

• Added a new function calculate_categories to calculate five categories for each geo_value, disease, and reference_date from sample data using duckdb. This function returns a DataFrame with the calculated categories.

Workflow integration:

• Integrated the calculate_categories function into the merge_and_render_anomaly function to calculate categories from final samples and save the results as both parquet and CSV files.

I figured that small PRs are better than big ones.

[natemcintosh]

I would like to add some tests for this. The general strategy I'm thinking of would be to take a sample of ~100 rows from a samples.parquet file, alter the values, and save it in a test folder. Then test that we get the general categories we expect.

I think that will work, but it feels kind of clunky. Does anyone have any other suggestions?