Sam has found some edge cases, so here are the updates to fix those.
add_threshold_lines()
Removed the sensor mins from being plotted because they were either all 0 or not useful.
Filter for seasonal thresholds first and then check how many seasonal thresholds there are. The logic before would break if there was a season mismatch. This fixed that edge case.
generate_initial_weekly_plots()
The important update here is on L105. Arranging by day instead of weekday fixed the inspect some call DT mismatch that Sam noticed, with the side effect that the days on the secondary axis will now sometimes not be 1-7, but some other order. Until this gets fixed, just check the weekday that is associated with your inspect call matches what day you actually want to look at.
2_verification_method
Sometimes the data is in a weird structure that the previous bind_rows() call wouldn’t handle. This if else checks the structure of update and cleans it so it’s easier to handle.
Misc
plotter()
These are some changes that I had made when I was using it briefly. This does not need to be included in this commit, so feel free to not approve this.
Sam has found some edge cases, so here are the updates to fix those.
add_threshold_lines()generate_initial_weekly_plots()2_verification_methodbind_rows()call wouldn’t handle. This if else checks the structure ofupdateand cleans it so it’s easier to handle.Misc
plotter()