changes from #224 re-applied from the latest main branch
found a bug in the way rh_u and rh_l were clipped (#220 )
adjusted upper limit of rh to 100% which is the instrument specification when RH is with regards to water (rh_cor, with regards to ice can be >100 and has an upper limit of 150 a.t.m.)
msg_lat and msg_lon are not precise enough and linear regression (even noise-resilient ones like RANSAC) were shown to fail at different places (#152). So they are removed from the output. They are still scraped from the emails by the tx-handling scripts (which I am not familiar with and didn't want to mess up with)
changes from #224 re-applied from the latest main branch