| OHDSI Home | Forums | Wiki | Github |
Hi Everyone,
I’m trying to get a patient-level-prediction package from ATLAS to run (via the CodetoRun file), and I think I’m almost there, but I can’t seem to get around the attached error. I found this 2-year-old post with the same issue, but no resolution. Any help would be greatly appreciated.
Hi @Bradenday_Jonah_Brad, this is possibly a versioning issue. can you share what version of the plp package you are using and what version of atlas you used to create the package?
Hi @RossW I’m using plp version 6.4.0, and I generated the package using ATLAS version 2.13.0.
Unfortunately the integration with Atlas hasn’t been maintained in a while and I would be surprised if it works.
But the error you’re seeing is indeed a versioning issue, in PLP 6.X the splitSettings are inside the modelDesign object while I believe it used to be separate before. So when you give the modelDesign and splitSettings to runMultiplePlp it doesn’t know what to do with the splitSettings, hence the error unused argument.
Having said that, I think there should be a renv.lock file as part of the output when you create the package with Atlas, that one should reference a PLP version that matches with the input objects you get from atlas.
Do you see a renv.lock in the same folder as the codeToRun file?
Egill
HI @egillax Thanks for the explanation. There’s no renv.lock file in the directory containing CodetoRun, though there is a createLockfiles R file (not sure if this is relevant).
There’s no renv.lock file in the directory containing CodetoRun, though there is a createLockfiles R file (not sure if this is relevant).
Yes I think that’s correct. The code seems to be from here I think: GitHub - OHDSI/SkeletonPredictionStudy: A skeleton package to use as a template for studies developing prediction models
But depending on what you were trying/wanting to do I could point you towards some code which should be more modern.
Egill
@egillax for now, I’m just trying to test ATLAS functionality to eventually make training materials for our lab. If the patient-level-prediction feature of ATLAS is no longer fully supported, that’s okay. We can always promote the R package instead.
