search

Open-Systems-Pharmacology / QualificationRunner

Qualification runner in charge of managing a qualification workflow
Other
0 stars 1 forks source link

Warning: Error in TimeProfile plot 31. The formula 1 for unit µmol/l evaluates to a non-finite value (NaN, Inf, -Inf) #68

Open PavelBal opened 4 years ago

PavelBal commented 4 years ago

I receive this error when running a qualification plan. Any idea what this is supposed to mean? The plot is supposed to be created from the "AllPlots"-Section of the qualificiation plan, so just re-creating a plot from PK-Sim. I have no idea where to look.

Yuri05 commented 4 years ago

This error message comes from the Reporting Engine, not from the QualificationRunner. https://github.com/Open-Systems-Pharmacology/Matlab-Toolbox-Core/blob/099d1c9528f82e5aa77c0fa73821033d672d904b/code/auxiliaries/getUnitFactor.m#L75

Seems to be an issue with the units conversion (missing mol weight in an observed data set?). You could set the breakpoint in the Matlab Code to figure out what happens.

Which version of RE are ayou using? Can you try the latest one (not released yet)? https://ci.appveyor.com/project/open-systems-pharmacology-ci/reporting-engine/build/artifacts

PavelBal commented 4 years ago

Yes, it is the missing MW, thank you @Yuri05 . But how can I add the missing MW to the data??

The "molecule" thas has been selected during import is not present in the project, and I cannot change the molecule in the "Observed Data" view in PK-Sim:

image

PavelBal commented 4 years ago

Wait, I just see - there is the field "Molecular weight" which is filled out.

PavelBal commented 4 years ago

Ok, obviously the Reporting Engine does not care about the MW stored in Observed Data, it always tries to find the respective compound in the simulation (in a very shady way):

https://github.com/Open-Systems-Pharmacology/Reporting-Engine/blob/c4031220e9ae589ffa4757461cd8fd9010052a77/src/code/qualification/getMolecularWeightForPath.m#L35

Any advice what to do except for re-importing the data?

Yuri05 commented 4 years ago

obviously the Reporting Engine does not care about the MW stored in Observed Data, it always tries to find the respective compound in the simulation

So your observed data was created for a compound X which is not in the simulation? Or is it unrelated to any compound?

msevestre commented 4 years ago

@PavelBal The meta data are not exported to the reporting engine. Only the CSV Values

Any advice what to do except for re-importing the data?

Do not expect meta data to be used

PavelBal commented 4 years ago

So your observed data was created for a compound X which is not in the simulation? Or is it unrelated to any compound?

It is created for compound X which is not in the simulation - or at least not with this name.

Yuri05 commented 4 years ago

Well, as a workaround you could create compound with the name X and your target mol weight and add it into your simulations (without any administration). Alternatively: wait for the reimplementation in R ;)

msevestre commented 4 years ago

@Yuri05 EVen reimplementation in R will not use meta data as long as we only export csv

PavelBal commented 4 years ago

Well, as a workaround you could create compound with the name X and your target mol weight and add it into your simulations (without any administration).

Yepp, did exactly this...

EVen reimplementation in R will not use meta data as long as we only export csv But the snapshots have these data, so why not use it?

msevestre commented 4 years ago

As long as we only export csv :) The proof is in the pudding