Conversation
| # argv already somehow? | ||
| cats_dir = "/Users/acostare/Documents/CATS-CaliforniaTestSystem/" | ||
| include(joinpath(cats_dir, "Sienna/build_CATS.jl")) | ||
| sys = build_CATS_system(first_order = true) |
There was a problem hiding this comment.
[JuliaFormatter] reported by reviewdog 🐶
| sys = build_CATS_system(first_order = true) | |
| sys = build_CATS_system(; first_order = true) |
| output_dir = mktempdir(; cleanup = true) | ||
| status = build!(model; output_dir = joinpath(output_dir, "CATS_UC")) | ||
| end | ||
| build_ok = status == IOM.ModelBuildStatus.BUILT |
There was a problem hiding this comment.
[JuliaFormatter] reported by reviewdog 🐶
| build_ok = status == IOM.ModelBuildStatus.BUILT | |
| build_ok = status == IOM.ModelBuildStatus.BUILT |
| try | ||
| # TODO Should probably be passed through argv but I think Github CI uses | ||
| # argv already somehow? | ||
| cats_dir = "/Users/acostare/Documents/CATS-CaliforniaTestSystem/" |
There was a problem hiding this comment.
There's a way to tell GitHub CI actions "clone this repo locally as setup." But then it'd also have to read the time series data from the CSV, which takes ~15 minutes. Not good.
A more robust solution would be to add the CATS system to PSB's system catalog, so that we can use it across the Sienna ecosystem via using PSB; sys = PSB.build_system(FooSystemCategory, "CATS_California"). There was discussion of doing that a while back...alternatively, we convert the time series CSV into something cross-platform (is JDL2 portable between machines?) and upload it somewhere.
There was a problem hiding this comment.
CATS is too large to try to run in the free CI-CD. This is a performance test geared towards offline testing IMO
|
Oh there's also one comment that got lost somehow: I see that you're having issues with the CATS system solving successfully. Check the device formulations and make sure they're reasonable, some sort of |
No description provided.