Input data overhaul

[scotthavens]

A big issue with SMRF and gridded data is that it must loaded all upfront which does not make sense for a year long run. Refactored the data loading to be more organized and have an easier call in SMRF.

• The HRRR data can be loaded either upfront or each timestep through a new config option
• When running threaded, the data is loaded into a data queue that the threads get from
• All gridded data is used in the same way for thermal radiation, meaning that the clear sky thermal is modeled then adjusted for cloud factor and not distributing thermal radiation from the gridded dataset.
• Remove MySQL for #127
• Removed the last_storm_basin metric what calculated the time since last storm for the whole basin, this is not being used anymore

[scotthavens]

@etrujil and @Hedrick-ARS, here is a good overhaul to the input data for SMRF to make it more efficient. This also has removed the MySQL capability since we haven't used it in a few years.

[jomey]

Some of my comments apply to multiple data loading classes and I did not repeat the message there. Easier to discuss suggestions in one place.

[scotthavens]

Some of my comments apply to multiple data loading classes and I did not repeat the message there. Easier to discuss suggestions in one place.

Got it, a lot of the classes are similar so I've applied the comments to all classes that were applicable.