I don’t think the solution to the rain sensor is in software this time. I think it’s due to wind gusts flexing the pole with enough acceleration to flip the rain bucket. If true, this would be electrically indistinguishable from legitimate rain and so can’t easily be resolved in software. I’ve released the code already in a previous post. I just haven’t posted it on github yet, but I’ll do that as soon as I get some time.
I believe we’re going to need a physical modification in order to fix the false rain issue. Either stiffen the pole, or add guy wires, or move the rain sensor lower on the pole, or something else. Basically, the top of the pole (where the rain sensor is) is whipping around violently enough to cause the rain bucket inside to move. Each time it moves far enough to trip its sensor is equivalent to ~0.3mm of rain filling and tipping the bucket, and the software is interpreting it as such.