This is an updated version of the original pull request (the old one can be deleted). I fixed a couple of bugs mentioned in the previous PR (e.g. some adaptation period issues, local sea level rise and negative values) and also timed the protect component. After it compiled, It averaged out to 0.0003 seconds for a one-segment run for 20 periods, so I think even if we add 12,000 segments and a few additional components it should in theory only add a few seconds to FUND. Definitely a lot better than we originally anticipated!
I still need to work on data integration and country to segment mapping. I also need to add the no-adaptation and retreat cases, which I can do now or after a review of this code. Finally, everything needs to be compared to Delavane's data.
Let me know if you have any thoughts on coding style or the issues I've flagged so far!
@davidanthoff, @FrankErrickson, @lrennels
This is an updated version of the original pull request (the old one can be deleted). I fixed a couple of bugs mentioned in the previous PR (e.g. some adaptation period issues, local sea level rise and negative values) and also timed the protect component. After it compiled, It averaged out to 0.0003 seconds for a one-segment run for 20 periods, so I think even if we add 12,000 segments and a few additional components it should in theory only add a few seconds to FUND. Definitely a lot better than we originally anticipated!
I still need to work on data integration and country to segment mapping. I also need to add the no-adaptation and retreat cases, which I can do now or after a review of this code. Finally, everything needs to be compared to Delavane's data.
Let me know if you have any thoughts on coding style or the issues I've flagged so far!