Update ratio_stats sales chasing detection methods

[dfsnow]

We disabled the sales chasing detection in the reporting.ratio_stats table when it was converted to a Python model in #422. The reasons to disable it were enumerated in that PR, but to recap:

• The detect_chasing() function specifically did not play well with the Spark environment for some reason (probably using ECDF()).
• The results of the chasing detection were not heavily used anywhere.
• The detect_chasing() function combines two methods using an AND operator. At least one of those methods was highly sensitive to/created lots of false positives for areas with extremely homogenous housing stock.

However, the chasing detection output is valuable. We should update the ratio_stats Python model to reimplement chasing with improved methods. "Improved" here could mean many things, e.g.:

• Simply changing some of the parameters set by default for the original methods.
• Trying to refine or tweak the existing methods
• Creating entirely new methods, incorporating them into assesspy, and then integrating them into the Python model

[dfsnow]

This issue is closed as of #521. The updated sales chasing detection methods added to assesspy 2.0.0 are less sensitive and work within dbt Python models.