Closed dirkseidensticker closed 5 years ago
Unfortunately, your very elegant solution resulted in an build error in travis. If you accept (or modify and correct) my solution, I am more than happy to merge!
Your solution works perfectly 🎱 Sorry, it took me so long for testing. And many thanks. I learned quite something going through your code, adapting my quick-and-dirty hack.
The base plots of calibrated dates are disturbed as soon as the calibrated range hit's 1950 and the probability is not 0 at this time; e.g.
results in:
I suggest to set the last value of the raw_probabilities to 0 if a calibrated date 'hit's' 1950.5. With that fix (?) output looks like:
Alternatively one might think of adding a value at
year
1951(?) and set its probability to 0.@MartinHinz What do you think?