I noticed this in the tests and though I'd mention it in case it turns out to be an issue.
The online CI tests run on linux, mac and windows. All the tests are fine, except for the CCEI tests on windows.
The snapshot tests is failing, meaning that it thinks the values output from one of the calculations has changed.
It's super odd however, because it's only in the values of two of the models, and they're hugely different!
You can see the comparison, looking at the last column (cmd_CNRM-ESM2-1-ssp585) and comparing the rows (i.e. the first row is the old data for row 1, the second is the new data for row 2; then the 3rd is the old data for row 2 etc.). The values are an order of magnitude off, but nothing else has changed.
The same for model cmd_CanESM5-ssp585 below.
I don't think this is an issue (given that Mac and Linux are both good), but it's worth remembering this discrepancy, especially if you recreate the CCEIs on a windows computer.
One quick test would be to run the tests in the test-00_ccei.R file on your (presumably?) windows computer and see what happens.
I noticed this in the tests and though I'd mention it in case it turns out to be an issue.
The online CI tests run on linux, mac and windows. All the tests are fine, except for the CCEI tests on windows.
The snapshot tests is failing, meaning that it thinks the values output from one of the calculations has changed.
It's super odd however, because it's only in the values of two of the models, and they're hugely different!
You can see the comparison, looking at the last column (cmd_CNRM-ESM2-1-ssp585) and comparing the rows (i.e. the first row is the old data for row 1, the second is the new data for row 2; then the 3rd is the old data for row 2 etc.). The values are an order of magnitude off, but nothing else has changed.
The same for model cmd_CanESM5-ssp585 below.
I don't think this is an issue (given that Mac and Linux are both good), but it's worth remembering this discrepancy, especially if you recreate the CCEIs on a windows computer.
One quick test would be to run the tests in the test-00_ccei.R file on your (presumably?) windows computer and see what happens.