wkearn / daylightplots.jl Goto Github PK

Right now we use the tlims hack to calculate the data range, which we need to compute the sunrise/sunset times. This means that using DaylightPlots with user-defined time series types requires extending tlims with a new method. This isn't too hard, but makes it less useful out of the box. Perhaps there is a better way to figure out the data range.

Right now, this is hardcoded to fillcolor=[:yellow :grey], fillalpha=0.3. Someone could conceivably want to use a different color.

The solar calculator is almost line-for-line ported from the NOAA calculator. I have not really benchmarked this. It seems like it does all right when computing a few months worth of data at a time. Any more than that, and the plot gets crowded anyway, so maybe the performance is fine.

Also, all the function names are in the Javascript CamelCase style, and those could be made to look more Julian.

I just noticed that if you try to add something to the plot that isn't covered up by the recipe (like a markershape), then it also adds this to the daylight series, which results in weird behavior: