We move the race calling from get_national_summary to get_aggregate_predictions and get_aggregate_prediction_intervals. This means that we apply the race calls to the contest level predictions also (ie. state predictions or house district predictions) instead of purely on the national summary predictions.
This is more conceptually correct, but will also let us show the our contest level predictions after race calls have already happened if we want.
Unit tests were also updated to take this into account.
Description
We move the race calling from
get_national_summary
toget_aggregate_predictions
andget_aggregate_prediction_intervals
. This means that we apply the race calls to the contest level predictions also (ie. state predictions or house district predictions) instead of purely on the national summary predictions.This is more conceptually correct, but will also let us show the our contest level predictions after race calls have already happened if we want.
Unit tests were also updated to take this into account.
Jira Ticket
https://arcpublishing.atlassian.net/browse/ELEX-4455
Test Steps
This now allows a force call for Republicans (and you can see the predictions and intervals all be below zero for the state as a result).
If you want to test on aggregate model, then you can just add more contests using this dictionary (either in the CLI or using the testbed).
also
pytest