SEAJang commented 5 months ago

Pull Request Key Information

Reason for change

Stationary deployment methods were not re-implemented with the original code rework, only a skeleton existed.

Stationary deployment methods have been re-introduced Stationary methods now use rolling averages based on a short or a long window

To reimplement the stationary deployment method and improve upon it by adding a feature to use the rolling averages as thresholds.

All unit tests pass - added in new unit test to ensure rolling averages function works as intended. results.txt

Manual end to end tests checked the following:

Manually ran a mobile method parameterized to act like a stationary method and compared it to a stationary deployed method; the results matched.
Verified that the daily deployment cost is as expected, despite a known bug with upfront costs.
Deployment cost is added to the total even if measurements didn't happen due to weather.
Deployment cost is not considered only if the deployment month/year was invalid.
Confirmed that the deployment year/month functionality with stationary works.
Checked that deployment numbers, when accounting for weather, made sense. Previously, the count increased even if deployment didn't happen due to weather.
Checked that the survey time/time between sites was not a factor for stationary deployment

SEAJang commented 5 months ago

All unit tests pass. results.txt

Tested partial deployment with stationary functionality using the following infrastructure file. facilities_Perm.csv

SEAJang commented 5 months ago

All new unit tests pass results.txt

SEAJang commented 5 months ago

SEAJang commented 5 months ago

Manual test results of testing stationary deployment with Upfront cost, and partial deployment at 17/50 total sites.