liesel model definition fails since recently

[stefgehrig]

After a while and some updating on my local Windows machine (of R, Python, their libraries, including liesel), the rliesel::liesel() function is not called successfully in two scripts where it used to work fine previously. I don't know yet whether this a weird local issue in my environment / a coding error, or indeed some general incompatability of software versions. I'd be interested in whether you can reproduce the problem.

This is the code that fails since recently:

1. R chunk to simulate some data

library(tidyverse)
library(rliesel)
library(reticulate)
library(mgcv)

# simulate data
set.seed(123)
df <- tibble(
  center = sample(c("1", "2", "3", "4", "5"), size = 1e3, replace = TRUE),
  E_center = 2.25 + 0.05 * as.numeric(center),
  SD_center = as.numeric(center) / 20,
  bmi = rlnorm(1e3, log(25), 1/5) ,
  y = rlnorm(1e3, 0.15 * bmi - 0.00225 * bmi^2 + E_center, SD_center)
) %>% mutate(center = factor(center))

2. R chunk to set up the model that fails:

model <- liesel(
  response = df$y,
  distribution = "LogNormal",
  predictors = list(
    loc = predictor( ~ s(bmi, bs = "ps", k = 10) + s(center, bs = "re"), inverse_link = "Identity"),
    scale = predictor( ~ s(center, bs = "re"), inverse_link = "Exp") 
  ),
  data = df
)

This is my config:

• R:

R version 4.4.0 (2024-04-24 ucrt)
Platform: x86_64-w64-mingw32/x64
Running under: Windows 11 x64 (build 22631)
other attached packages:
[1] mgcv_1.9-1 nlme_3.1-164 reticulate_1.36.1 rliesel_0.0.1 lubridate_1.9.3 forcats_1.0.0 stringr_1.5.1 dplyr_1.1.4 purrr_1.0.2 readr_2.1.5 tidyr_1.3.1 tibble_3.2.1 ggplot2_3.5.1
[14] tidyverse_2.0.0
loaded via a namespace (and not attached):
[1] utf8_1.2.4 generics_0.1.3 xml2_1.3.6 stringi_1.8.3 lattice_0.22-6 hms_1.1.3 digest_0.6.35 magrittr_2.0.3 evaluate_0.23 grid_4.4.0 timechange_0.3.0 fastmap_1.1.1 jsonlite_1.8.8
[14] Matrix_1.7-0 fansi_1.0.6 viridisLite_0.4.2 scales_1.3.0 cli_3.6.2 rlang_1.1.3 munsell_0.5.1 splines_4.4.0 withr_3.0.0 tools_4.4.0 tzdb_0.4.0 colorspace_2.1-0 kableExtra_1.4.0
[27] png_0.1-8 vctrs_0.6.5 R6_2.5.1 lifecycle_1.0.4 pkgconfig_2.0.3 pillar_1.9.0 gtable_0.3.5 Rcpp_1.0.12 glue_1.7.0 systemfonts_1.0.6 xfun_0.43 tidyselect_1.2.1 rstudioapi_0.16.0
[40] knitr_1.46 htmltools_0.5.8.1 rmarkdown_2.26 svglite_2.1.3 compiler_4.4.0

• Python:

jax 0.4.26
liesel 0.2.9
pandas 2.2.2
plotnine 0.13.5
tensorflow_probability 0.24.0
Python 3.12.3 (tags/v3.12.3:f6650f9, Apr 9 2024, 14:05:25) [MSC v.1938 64 bit (AMD64)]
Windows-11-10.0.22631-SP0

The messages, warnings and error I get from the last chunk are:

Installed Liesel version 0.2.9 is compatible, continuing to set up model
liesel.model.nodes - WARNING - Calc(name="") was not updated during initialization, because the following exception occured: RuntimeError('Error while updating Calc(name="").'). See debug log for the full traceback.
liesel.model.nodes - WARNING - Calc(name="") was not updated during initialization, because the following exception occured: RuntimeError('Error while updating Calc(name="").'). See debug log for the full traceback.
liesel.model.nodes - WARNING - Calc(name="_model_log_lik") was not updated during initialization, because the following exception occured: RuntimeError('Error while updating Calc(name="_model_log_lik").'). See debug log for the full traceback.
liesel.model.nodes - WARNING - Calc(name="_model_log_prior") was not updated during initialization, because the following exception occured: RuntimeError('Error while updating Calc(name="_model_log_prior").'). See debug log for the full traceback.
liesel.model.nodes - WARNING - Calc(name="_model_log_prob") was not updated during initialization, because the following exception occured: RuntimeError('Error while updating Calc(name="_model_log_prob").'). See debug log for the full traceback.
Error in py_call_impl(callable, call_args$unnamed, call_args$named) :
RuntimeError: Error while updating Calc(name="scale_value").

And the traceback is:

── Python Exception Message ──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
Traceback (most recent call last):
File "C:\Users\stefa\AppData\Local\Programs\Python\Python312\Lib\site-packages\liesel\model\nodes.py", line 705, in update
self._value = self.function(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\stefa\AppData\Local\Programs\Python\Python312\Lib\site-packages\liesel\model\legacy.py", line 102, in fn
return bijector().forward(x) if not inverse else bijector().inverse(x)
^^^^^^^^^^
TypeError: 'NoneType' object is not callable
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "C:\Users\stefa\AppData\Local\Programs\Python\Python312\Lib\site-packages\liesel\model\model.py", line 454, in build_model
model = Model(nodes_and_vars, grow=False, copy=copy)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\stefa\AppData\Local\Programs\Python\Python312\Lib\site-packages\liesel\model\model.py", line 1032, in init
node.update()
File "C:\Users\stefa\AppData\Local\Programs\Python\Python312\Lib\site-packages\liesel\model\nodes.py", line 707, in update
raise RuntimeError(f"Error while updating {self}.") from e
RuntimeError: Error while updating Calc(name="scale_value").
── R Traceback ─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────── 1. └─rliesel::liesel(...) 2. └─mb$build_model() 3. └─reticulate:::py_call_impl(callable, call_args$unnamed, call_args$named)

[hriebl]

Thank you for your bug report, @stefgehrig, and sorry for the delay! The problem was caused by a reticulate update and should now be fixed. Could you confirm that things are working again for you with the latest commit from main?

[stefgehrig]

Thanks @hriebl and @jobrachem !! The issues is succesfully fixed also on my end.