Consider API changes for the next release

Cutting a major release is the time to make breaking changes to the API so we should consider what (if any) we should make. One that comes to mind is renaming the soiling.srr_analysis class to use CamelCase.

[mdeceglie]

Sorry for the vagueness. The new functionality would allow all precipitation events greater than some specified threshold to be counted as cleaning events.

To be consistent with the way precip_clean_only works, we would do it without modifying the soiling_interval_summary dataframe. precip_clean_only currently works by only allowing soiling recoveries on cleaning events that coincided with precipitation thresholds, but the slopes are still found independently for the sub-parts of the interval separated by detected cleaning events that did not coincide with rain.

But thinking this through it may be overall more consistent to handle both cases earlier in the algorithm, before slopes are fit for each interval.

[matt14muller]

I gave Kevin the few lines of code I used to combine detected shifts and rain for using as intervals. What we talked about is inserting another option for the user besided precip_clean_only…detected_clean_and_precip or something like that and the user could select this option. The Theil-sen fits would then be calculated on these intervals (this is what I did in my version) for the soiling map.