Open ptrscll opened 4 months ago
Hello, this is leeyabot
!
The current pull request's outputs do not differ from the old-new version of Hector (3.1.1 (d931a004)).
R squared | NRMSE | |
---|---|---|
CO2_concentration | 1.000 | 0.000 |
global_tas | 1.000 | 0.000 |
RF_CO2 | 1.000 | 0.000 |
RF_tot | 1.000 | 0.000 |
This table indicates the R^2 and normalized root mean squared error values for each variable. An R^2 of 1 and NRMSE of 0 indicate that the new values are identical to the old values.
Below are several plots comparing the performance of Hector between old and new versions:
The current pull request's outputs do not differ from the release version of Hector (3.2.0 (1ef0b4d)).
R squared | NRMSE | |
---|---|---|
CO2_concentration | 1.000 | 0.000 |
global_tas | 1.000 | 0.000 |
RF_tot | 1.000 | 0.000 |
This table indicates the R^2 and normalized root mean squared error values for each variable. An R^2 of 1 and NRMSE of 0 indicate that the new values are identical to the release values.
Below are several plots comparing the performance of Hector between the new version and the release version:
This PR adds code to make leeyabot also pull data from the
hector-run-archive
repository and make comparisons between that data and the new version of Hector. Additionally, the leeyabot graphs are now included within the resulting html file, and there is a new box plot summarizing the distribution of differences between the new and old data. This box plot (and the other graphs) are created for both the comparison between the new Hector data and thehector-run-archive
data and the comparison between the new Hector data and the data generated for the old-new R unit test inhector_comp.csv
.The leeyabot yml file is currently unchanged, so the yml file still only checks whether the output for the new version of Hector matches the data in
hector_comp.csv