Will Holmgren

It may be worth benchmarking the retrieval speeds for each data source before trying to improve the raw surfrad fetch. But removing the existing surfrad fetch/read functions is not on...

It's not as simple as putting an `if aoi_lt_90.any()` around `_gti_dirint_lt_90`. `_gti_dirint_gte_90` requires the `best_kt_prime` determined in `_gti_dirint_lt_90`. I don't think we can make this work without going beyond what's...

>library updates don't require a pvlib re-release if the two are decoupled +1. But that feels like a breaking change even if it's not a real API issue, so I'll...

NEP-29 says we can increase the minimum pandas to 1.1 and get rid of the test silliness.

To be clear, I only care about naming for the public API. Use whatever you want within functions so long as it's readable. > The other style seems to be...

> it's worth going through that typing business within the pvlib function. I think we're leaning towards the other direction: avoid anything to do with types in functions like this....

> That said, I was trying to think of the next person who comes along and is confused about why their code doesn't work. I had encountered this before and...

I don't have strong feelings about it, but at the moment I prefer * a single `mlfm.py` module that includes graphics * add matplotlib to [optional dependencies](https://github.com/pvlib/pvlib-python/blob/ed05fdd2e6f4b95646e9223cdd3a4db46a160962/setup.py#L55-L58) * [mark](https://github.com/pvlib/pvlib-python/blob/ed05fdd2e6f4b95646e9223cdd3a4db46a160962/pvlib/tests/test_solarposition.py#L185-L186) tests...

@adriesse thank you for proposing this interesting idea. My initial reactions were similar to @mikofski's: 1. Can we use functions and if necessary wrap them in a class? But I...

@mikofski I like those functions. I'm less sure about the API for the setters... looks pretty complicated for a setter. Another thought is a frozen data class with constructors for...