Merge branch 'main' into dev

2 files changed, 108 insertions, 38 deletions

diff --git a/log/test.log b/log/test.log
index c872cb1d..6d5bc470 100644
--- a/log/test.log
+++ b/log/test.log
@@ -1,50 +1,117 @@
 ℹ Testing mkin
 ✔ | F W S  OK | Context
 ✔ |         5 | AIC calculation
-✔ |         5 | Analytical solutions for coupled models [3.1s]
+✔ |         5 | Analytical solutions for coupled models [1.5s]
 ✔ |         5 | Calculation of Akaike weights
 ✔ |         3 | Export dataset for reading into CAKE
-✔ |         6 | Use of precompiled symbols in mkinpredict [5.3s]
-✔ |        12 | Confidence intervals and p-values
-✔ |     1  12 | Dimethenamid data from 2018 [31.5s]
-✔ |        14 | Error model fitting [5.6s]
+✔ |         6 | Use of precompiled symbols in mkinpredict [3.1s]
+✔ |        12 | Confidence intervals and p-values [0.4s]
+✔ |     1  12 | Dimethenamid data from 2018 [13.2s]
+────────────────────────────────────────────────────────────────────────────────
+Skip ('test_dmta.R:88:3'): Different backends get consistent results for SFO-SFO3+, dimethenamid data
+Reason: Fitting this ODE model with saemix takes about 5 minutes on my new system
+────────────────────────────────────────────────────────────────────────────────
+✔ |        14 | Error model fitting [2.5s]
 ✔ |         5 | Time step normalisation
-✔ |         4 | Calculation of FOCUS chi2 error levels
-✔ |        14 | Results for FOCUS D established in expertise for UBA (Ranke 2014)
-✔ |         4 | Test fitting the decline of metabolites from their maximum
-✔ |         1 | Fitting the logistic model
-✔ |        10 | Batch fitting and diagnosing hierarchical kinetic models [43.1s]
-✔ |     2  16 | Nonlinear mixed-effects models [441.3s]
+✔ |         4 | Calculation of FOCUS chi2 error levels [0.3s]
+✔ |        14 | Results for FOCUS D established in expertise for UBA (Ranke 2014) [0.4s]
+✔ |         4 | Test fitting the decline of metabolites from their maximum [0.2s]
+✔ |         1 | Fitting the logistic model [0.1s]
+✔ |        10 | Batch fitting and diagnosing hierarchical kinetic models [19.0s]
+✖ | 1   2  15 | Nonlinear mixed-effects models [149.2s]
+────────────────────────────────────────────────────────────────────────────────
+Failure ('test_mixed.R:21:3'): Print methods work
+Results have changed from known value recorded in 'print_dfop_saem_1.txt'.
+
+old[13:23] vs new[13:23]
+  ""
+  "Fitted parameters:"
+  "          estimate lower  upper"
+- "parent_0     99.92 98.77 101.06"
++ "parent_0     99.96 98.82 101.11"
+- "log_k1       -2.72 -2.95  -2.50"
++ "log_k1       -2.71 -2.94  -2.49"
+- "log_k2       -4.14 -4.27  -4.01"
++ "log_k2       -4.14 -4.26  -4.01"
+- "g_qlogis     -0.35 -0.53  -0.16"
++ "g_qlogis     -0.36 -0.54  -0.17"
+- "a.1           0.92  0.68   1.16"
++ "a.1           0.93  0.69   1.17"
+  "b.1           0.05  0.04   0.06"
+  "SD.log_k1     0.37  0.23   0.51"
+and 1 more ...
+
+Skip ('test_mixed.R:80:3'): saemix results are reproducible for biphasic fits
+Reason: Fitting with saemix takes around 10 minutes when using deSolve
+
+Skip ('test_mixed.R:133:3'): SFO-SFO saemix specific analytical solution work
+Reason: This is seldom used, so save some time
+────────────────────────────────────────────────────────────────────────────────
 ✔ |         3 | Test dataset classes mkinds and mkindsg
-✔ |        10 | Special cases of mkinfit calls
-✔ |         3 | mkinfit features
+✔ |        10 | Special cases of mkinfit calls [0.3s]
+✔ |         3 | mkinfit features [0.5s]
 ✔ |         8 | mkinmod model generation and printing
-✔ |         3 | Model predictions with mkinpredict
-✔ |        12 | Multistart method for saem.mmkin models [71.6s]
-✔ |        16 | Evaluations according to 2015 NAFTA guidance [2.7s]
-✔ |         9 | Nonlinear mixed-effects models with nlme [10.1s]
-✔ |        15 | Plotting [12.1s]
+✔ |         3 | Model predictions with mkinpredict [0.1s]
+✖ | 3       9 | Multistart method for saem.mmkin models [23.2s]
+────────────────────────────────────────────────────────────────────────────────
+Failure ('test_multistart.R:44:3'): multistart works for saem.mmkin models
+Snapshot of `testcase` to 'multistart/mixed-model-fit-for-saem-object-with-mkin-transformations.svg' has changed
+Run `testthat::snapshot_review('multistart/')` to review changes
+Backtrace:
+ 1. vdiffr::expect_doppelganger(...)
+      at test_multistart.R:44:2
+ 3. testthat::expect_snapshot_file(...)
+
+Failure ('test_multistart.R:55:3'): multistart works for saem.mmkin models
+Snapshot of `testcase` to 'multistart/llhist-for-dfop-sfo-fit.svg' has changed
+Run `testthat::snapshot_review('multistart/')` to review changes
+Backtrace:
+ 1. vdiffr::expect_doppelganger("llhist for dfop sfo fit", llhist_dfop_sfo)
+      at test_multistart.R:55:2
+ 3. testthat::expect_snapshot_file(...)
+
+Failure ('test_multistart.R:56:3'): multistart works for saem.mmkin models
+Snapshot of `testcase` to 'multistart/parplot-for-dfop-sfo-fit.svg' has changed
+Run `testthat::snapshot_review('multistart/')` to review changes
+Backtrace:
+ 1. vdiffr::expect_doppelganger("parplot for dfop sfo fit", parplot_dfop_sfo)
+      at test_multistart.R:56:2
+ 3. testthat::expect_snapshot_file(...)
+────────────────────────────────────────────────────────────────────────────────
+✔ |        16 | Evaluations according to 2015 NAFTA guidance [1.6s]
+✔ |         9 | Nonlinear mixed-effects models with nlme [3.8s]
+✖ | 1      14 | Plotting [4.7s]
+────────────────────────────────────────────────────────────────────────────────
+Failure ('test_plot.R:55:3'): Plotting mkinfit, mmkin and mixed model objects is reproducible
+Snapshot of `testcase` to 'plot/mixed-model-fit-for-nlme-object.svg' has changed
+Run `testthat::snapshot_review('plot/')` to review changes
+Backtrace:
+ 1. vdiffr::expect_doppelganger(...)
+      at test_plot.R:55:2
+ 3. testthat::expect_snapshot_file(...)
+────────────────────────────────────────────────────────────────────────────────
 ✔ |         4 | Residuals extracted from mkinfit models
-✔ |     1  36 | saemix parent models [81.8s]
-✔ |         2 | Complex test case from Schaefer et al. (2007) Piacenza paper [1.5s]
+✔ |     1  36 | saemix parent models [31.4s]
+────────────────────────────────────────────────────────────────────────────────
+Skip ('test_saemix_parent.R:143:3'): We can also use mkin solution methods for saem
+Reason: This still takes almost 2.5 minutes although we do not solve ODEs
+────────────────────────────────────────────────────────────────────────────────
+✔ |         2 | Complex test case from Schaefer et al. (2007) Piacenza paper [0.5s]
 ✔ |        11 | Processing of residue series
-✔ |        10 | Fitting the SFORB model [3.8s]
+✔ |        10 | Fitting the SFORB model [1.7s]
 ✔ |         1 | Summaries of old mkinfit objects
 ✔ |         5 | Summary
-✔ |         4 | Results for synthetic data established in expertise for UBA (Ranke 2014) [2.1s]
-✔ |         9 | Hypothesis tests [7.3s]
-✔ |         4 | Calculation of maximum time weighted average concentrations (TWAs) [1.5s]
+✔ |         4 | Results for synthetic data established in expertise for UBA (Ranke 2014) [0.8s]
+✔ |         9 | Hypothesis tests [2.9s]
+✔ |         4 | Calculation of maximum time weighted average concentrations (TWAs) [0.7s]
 
 ══ Results ═════════════════════════════════════════════════════════════════════
-Duration: 731.0 s
-
-── Skipped tests (4) ───────────────────────────────────────────────────────────
-• Fitting this ODE model with saemix takes about 15 minutes on my system (1):
-  'test_dmta.R:88:3'
-• Fitting with saemix takes around 10 minutes when using deSolve (1):
-  'test_mixed.R:80:3'
-• This is seldom used, so save some time (1): 'test_mixed.R:133:3'
-• This still takes almost 2.5 minutes although we do not solve ODEs (1):
-  'test_saemix_parent.R:143:3'
-
-[ FAIL 0 | WARN 0 | SKIP 4 | PASS 281 ]
+Duration: 262.6 s
+
+── Skipped tests  ──────────────────────────────────────────────────────────────
+• Fitting this ODE model with saemix takes about 5 minutes on my new system (1)
+• Fitting with saemix takes around 10 minutes when using deSolve (1)
+• This is seldom used, so save some time (1)
+• This still takes almost 2.5 minutes although we do not solve ODEs (1)
+
+[ FAIL 5 | WARN 0 | SKIP 4 | PASS 276 ]
diff --git a/tests/testthat/test_dmta.R b/tests/testthat/test_dmta.R
index cd6de341..3941cae1 100644
--- a/tests/testthat/test_dmta.R
+++ b/tests/testthat/test_dmta.R
@@ -85,9 +85,12 @@ test_that("Different backends get consistent results for SFO-SFO3+, dimethenamid
     "Iteration 5, LME step.*not converge")
   ints_nlme_mets <- intervals(nlme_sfo_sfo3p_tc, which = "fixed")
 
-  skip("Fitting this ODE model with saemix takes about 15 minutes on my system")
+  skip("Fitting this ODE model with saemix takes about 5 minutes on my new system")
+  # August 2023: Please refer to the mkin vignette for the currently
+  # recommended way to fit the DMTA pathway data with saemix
+
   # As DFOP is overparameterised and leads to instabilities and errors, we
-  # need to use SFO.
+  # need to use SFO with nlme
   # saem_saemix_sfo_sfo3p_tc <- saem(dmta_sfo_sfo3p_tc)
   # The fit above, using SFO for the parent leads to low values of DMTA_0
   # (confidence interval from 84.4 to 92.8) which is not consistent with what