loading a model built in V1, in V3

[ks-liv]

Hi there, I have a model that I built using PM V 1.5.0 and want to load it into V 3.1.2. Using PM_load() - and I have also tried using legacy code PMload() - I get an error that the output file cannot be found

Error in PM_load():
✖ No Pmetrics output file found in ..
Run rlang::last_trace() to see where the error occurred.

rlang::last_trace()
<error/rlang_error>
Error in PM_load():
✖ No Pmetrics output file found in ..

---

Backtrace:
▆

1. └─Pmetrics::PM_load(390)
   Run rlang::last_trace(drop = FALSE) to see 2 hidden frames.

...despite the output folder and files being there. Can you please advise?

Many thanks!

[mhovd]

It seems you are running PM_load() without supplying the run number of the path to the outputs. If you made a run in e.g. a Runs folder, make sure to have that folder as your working directory, then use PM_load(run_number), where run_number is the number of your run.

Let me know if that doesn't work out!

[ks-liv]

Yes sorry I left that out of my comment but I did supply the run number in my code! Sent from Outlook for Android<https://aka.ms/AAb9ysg>

…

________________________________ From: Markus Hovd ***@***.***> Sent: Monday, April 27, 2026 9:03:54 AM To: LAPKB/Pmetrics ***@***.***> Cc: ks-liv ***@***.***>; Author ***@***.***> Subject: Re: [LAPKB/Pmetrics] loading a model built in V1, in V3 (Issue #346) [https://avatars.githubusercontent.com/u/66058642?s=20&v=4]mhovd left a comment (LAPKB/Pmetrics#346)<#346 (comment)> It seems you are running PM_load() without supplying the run number of the path to the outputs. If you made a run in e.g. a Runs folder, make sure to have that folder as your working directory, then use PM_load(run_number), where run_number is the number of your run. Let me know if that doesn't work out! — Reply to this email directly, view it on GitHub<#346 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/CCVJFV4ENLDFGL2ZSJXCOC34X4H6VAVCNFSM6AAAAACYHM2OQKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DGMRVGI2DGMBZGI>. Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS<https://github.com/notifications/mobile/ios/CCVJFV2UUFRKH3JPSOSLDPT4X4H6VA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMZSGUZDIMZQHEZKM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KUZTPN52GK4S7NFXXG> and Android<https://github.com/notifications/mobile/android/CCVJFV57RY3Z373K52ZVTZ34X4H6VA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMZSGUZDIMZQHEZKM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2K4ZTPN52GK4S7MFXGI4TPNFSA>. Download it today! You are receiving this because you authored the thread.Message ID: ***@***.***>

[mhovd]

Pmetrics 1.5 is extremely different from 3.1, so I am not certain that loading such old objects is even supported.

Unfortunately, I think I your best option is to run the model again in Pmetrics 3.1.2.

@mnneely , any other suggestions?

[mnneely]

Yes, I'm afraid we don't have automatic support to load 1.5 results in 3+. Is it possible to rerun the model? You likely can still use the same model file although I can't guarantee it. If you need help getting it into 3+ version, let us know and we'd be glad to do that. The data format should still be fine.

[ks-liv]

Thanks both - I re-ran the model in V3 as suggested; it converged but the results page did not pop up, and I get this:

thread '' (27313210) panicked at /private/var/folders/tb/y368xp_x10s3ty1b_mtl5mxr0000gn/T/RtmpTTlBoT/R.INSTALL183c401554d/Pmetrics/src/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/extendr-api-0.9.0/src/robj/into_robj.rs:73:25:
Failed to write predictions

In the Runs/outputs folder, there are only 5 files: covs.csv, cycles.csv, PMout.Rdata, settings.json, theta.csv

It was definitely working in V1 so I'm not sure what the issue is?

Many thanks again

[mhovd]

Apparently, the simulator that powers Pmetrics crashed while attempting to write the predictions, and as such the rest your output was not generated either. What kind of model are you using? I would be very interested in seeing if I am able to reproduce that error, so that I can prevent others from experiencing it in the future.

[ks-liv]

would you mind emailing me at katstott@liverpool.ac.uk and i can send the model and data file in confidence?

[mhovd]

I will e-mail you shortly, from markus.hovd@farmasi.uio.no

[mhovd]

I managed to find the root of the issue - your model is what we call non-stiff. By default, Pmetrics uses the BDF method to solve differential equations. However, for non-stiff problems like yours, the BDF solver can have problems with the step size, and a Runge-Kutta solver is recommended instead.

Luckily, we recently implemented several different solvers, all available using the solver = ... argument in PM_model$new().

All you have to do is add this little line

modPK <- PM_model$new(
  pri = list( ... ),
  sec = function(){ ... },
  eqn = function(){ ... },
  out = function(){ ... },
  
  err = list( ... ), 
  solver = "TSIT45" # <- Add this
)

And your model will run as expected.

I will make sure to make the error message more informative, so that anyone encountering this issue in the future will know what to do.

[ks-liv]

Thank you! That works. Much appreciated!

…

On Thu, Apr 30, 2026 at 11:29 AM Markus Hovd ***@***.***> wrote: *mhovd* left a comment (LAPKB/Pmetrics#346) <#346 (comment)> I managed to find the root of the issue - your model is what we call *non-stiff*. By default, Pmetrics uses the BDF method to solve differential equations. However, for non-stiff problems like yours, the BDF solver can have problems with the step size, and a Runge-Kutta solver is recommended instead. Luckily, we recently implemented several different solvers, all available using the solver = ... argument in PM_model$new(). All you have to do is add this little line modPK <- PM_model$new( pri = list( ... ), sec = function(){ ... }, eqn = function(){ ... }, out = function(){ ... }, err = list( ... ), solver = "TSIT45" # <- Add this ) And your model will run as expected. I will make sure to make the error message more informative, so that anyone encountering this issue in the future will know what to do. — Reply to this email directly, view it on GitHub <#346 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/CCVJFV4WY3HQBHN6CWUEOBL4YMTJFAVCNFSM6AAAAACYHM2OQKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DGNJRGY4DQMRTGI> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you authored the thread.Message ID: ***@***.***>