A few new things:

• I went through and replaced eccentricity_factor with irradiance_factor throughout the code. This is a breaking change, sort of, except that support for user specification of this parameter was never fully implemented, so I don't think it will break anything in practice. And it will avoid future confusion, because this factor is not the eccentricity, but rather the number multiplied by the solar constant to get the current solar irradiance (defaults to 1, but changes throughout the year due to the eccentric orbit).
• Changed the defaults in the climlab.radiation._Radiation_SW parent class to set a default value for coszen and infer insolation from coszen, rather than the other way around. This doesn't seem to break anything.
• Some more modular reorganization of the code in climlab.solar.insolation. Working toward having clean ways to call methods that return not just insolation, but the coszen and irradiance_factor values that are needed to drive radiation processes. Still a work in progress.

Have now reorganized the code more and written documentation. There are new methods

climlab.solar.insolation.instant_insolation_factors()

and

climlab.solar.insolation.daily_insolation_factors()

that are called similarly to the existing climlab.solar.insolation.daily_insolation(), but return coszen and irradiance_factor.

So we preserve all existing functionality but make it simple to get the parameters we need to pass to radiation modules.

WIP: supporting multiple time-averaging methods for the daily zenith angle. The function climlab.solar.insolation.daily_insolation_factors() accepts an input flag weighting where we can specify which time averaging method to use.

Still to do:

• Code to scale the irradiance_factor for cases where we use something other than simple time averaging for coszen
• Tests for that functionality

Then in a further PR I will work on adapting the Insolation process objects to handle all this new functionality.