Switch MKI sorting from quicksort to mergesort

[dfsnow]

While working on ccao-data/data-architecture#422, I discovered that mki() doesn't produce stable results i.e. it returns slightly different values for each run. This is due to the default pandas.sort_values() function using quicksort, which is not stable and returns different index orders for ties.

Per this SO post, This PR switches the sort method to mergesort, which is stable.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.74%. Comparing base (534d741) to head (6ed1aac).

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #19   +/-   ##
=======================================
  Coverage   98.74%   98.74%           
=======================================
  Files          12       12           
  Lines         478      478           
=======================================
  Hits          472      472           
  Misses          6        6           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[wrridgeway]

Thanks for catching this

[jeancochrane]

Nice!

[jeancochrane]

Quick follow-up question: We're not using quicksort anywhere else, are we?

[dfsnow]

A quick ripgrep of all our repos shows the following. I'm not sure if any of these would be impactful. Perhaps only the comps one is a concern @jeancochrane

projects/model-sales-val/glue/flagging_script_glue/flagging.py
485:        original_df.sort_values("meta_sale_date")
628:        dups.sort_values("meta_sale_date")
657:        original_df.sort_values("meta_sale_date")

models/model-res-avm/python/comps.py
42:    sorted_comparison_df = comparison_df.sort_values(

models/model-land-avm/notebooks/data_exploration.ipynb
286:    "price_sorted = cop.sort_values(by=['sale_price']).loc[:, 'sale_price']"

projects/scratch-res-avm/python/comps.py
42:    sorted_comparison_df = comparison_df.sort_values(

packages/assesspy/assesspy/formulas.py
179:    df = df.sort_values(by="sp", kind="mergesort")  # for stable sort results

packages/assesspy/docs/source/notebooks/example-ratio-study.ipynb
782:    "gini_data = combined[['sale_price', 'assessed']].sort_values(by='sale_price')\n",