diff options
| -rw-r--r-- | NEWS.md | 2 | ||||
| -rw-r--r-- | docs/news/index.html | 3 | ||||
| -rw-r--r-- | docs/pkgdown.yml | 2 | ||||
| -rw-r--r-- | log/test.log | 37 | ||||
| -rw-r--r-- | tests/testthat/setup_script.R | 22 | ||||
| -rw-r--r-- | tests/testthat/test_saemix_parent.R | 2 |
6 files changed, 40 insertions, 28 deletions
@@ -1,5 +1,7 @@ # mkin 1.2.1 (unreleased) +- '{data,R}/ds_mixed.rda': Include the test data in the package instead of generating it in 'tests/testthat/setup_script.R'. Refactor the generating code to make it consistent and update tests. + - 'tests/testthat/setup_script.R': Excluded another ill-defined random effect for the DFOP fit with 'saem', in an attempt to avoid a platform dependence that surfaced on Fedora systems on the CRAN check farm - 'tests/testthat/test_mixed.R': Round parameters found by saemix to two significant digits before printing, to also help to avoid platform dependence of tests diff --git a/docs/news/index.html b/docs/news/index.html index 6b9a7d51..f6883766 100644 --- a/docs/news/index.html +++ b/docs/news/index.html @@ -89,7 +89,8 @@ <div class="section level2"> <h2 class="page-header" data-toc-text="1.2.1" id="mkin-121-unreleased">mkin 1.2.1 (unreleased)<a class="anchor" aria-label="anchor" href="#mkin-121-unreleased"></a></h2> -<ul><li><p>‘tests/testthat/setup_script.R’: Excluded another ill-defined random effect for the DFOP fit with ‘saem’, in an attempt to avoid a platform dependence that surfaced on Fedora systems on the CRAN check farm</p></li> +<ul><li><p>‘{data,R}/ds_mixed.rda’: Include the test data in the package instead of generating it in ‘tests/testthat/setup_script.R’. Refactor the generating code to make it consistent and update tests.</p></li> +<li><p>‘tests/testthat/setup_script.R’: Excluded another ill-defined random effect for the DFOP fit with ‘saem’, in an attempt to avoid a platform dependence that surfaced on Fedora systems on the CRAN check farm</p></li> <li><p>‘tests/testthat/test_mixed.R’: Round parameters found by saemix to two significant digits before printing, to also help to avoid platform dependence of tests</p></li> <li><p>‘R/saem.R’: Fix a bug that prevented that ‘error.ini’ is passed to ‘saemix_model’, and set default to c(1, 1) to avoid changing test results</p></li> <li><p>‘R/parplot.R’: Show initial values for error model parameters</p></li> diff --git a/docs/pkgdown.yml b/docs/pkgdown.yml index b6c8f1cc..7cf87069 100644 --- a/docs/pkgdown.yml +++ b/docs/pkgdown.yml @@ -13,7 +13,7 @@ articles: dimethenamid_2018: web_only/dimethenamid_2018.html multistart: web_only/multistart.html saem_benchmarks: web_only/saem_benchmarks.html -last_built: 2022-11-18T18:18Z +last_built: 2022-11-18T18:28Z urls: reference: https://pkgdown.jrwb.de/mkin/reference article: https://pkgdown.jrwb.de/mkin/articles diff --git a/log/test.log b/log/test.log index d1de270e..8fdaec48 100644 --- a/log/test.log +++ b/log/test.log @@ -1,53 +1,53 @@ ℹ Testing mkin ✔ | F W S OK | Context ✔ | 5 | AIC calculation -✔ | 5 | Analytical solutions for coupled models [3.3s] +✔ | 5 | Analytical solutions for coupled models [3.4s] ✔ | 5 | Calculation of Akaike weights ✔ | 3 | Export dataset for reading into CAKE -✔ | 12 | Confidence intervals and p-values [1.0s] -✔ | 1 12 | Dimethenamid data from 2018 [32.6s] +✔ | 12 | Confidence intervals and p-values [1.2s] +✔ | 1 12 | Dimethenamid data from 2018 [34.0s] ──────────────────────────────────────────────────────────────────────────────── Skip ('test_dmta.R:98'): Different backends get consistent results for SFO-SFO3+, dimethenamid data Reason: Fitting this ODE model with saemix takes about 15 minutes on my system ──────────────────────────────────────────────────────────────────────────────── -✔ | 14 | Error model fitting [5.1s] +✔ | 14 | Error model fitting [5.4s] ✔ | 5 | Time step normalisation ✔ | 4 | Calculation of FOCUS chi2 error levels [0.6s] -✔ | 14 | Results for FOCUS D established in expertise for UBA (Ranke 2014) [0.8s] +✔ | 14 | Results for FOCUS D established in expertise for UBA (Ranke 2014) [0.9s] ✔ | 4 | Test fitting the decline of metabolites from their maximum [0.4s] -✔ | 1 | Fitting the logistic model [0.2s] -✔ | 10 | Batch fitting and diagnosing hierarchical kinetic models [24.6s] +✔ | 1 | Fitting the logistic model [0.3s] +✔ | 10 | Batch fitting and diagnosing hierarchical kinetic models [25.8s] ✔ | 1 13 | Nonlinear mixed-effects models [0.4s] ──────────────────────────────────────────────────────────────────────────────── Skip ('test_mixed.R:79'): saemix results are reproducible for biphasic fits Reason: Fitting with saemix takes around 10 minutes when using deSolve ──────────────────────────────────────────────────────────────────────────────── ✔ | 3 | Test dataset classes mkinds and mkindsg -✔ | 10 | Special cases of mkinfit calls [0.5s] +✔ | 10 | Special cases of mkinfit calls [0.6s] ✔ | 3 | mkinfit features [0.7s] ✔ | 8 | mkinmod model generation and printing [0.2s] ✔ | 3 | Model predictions with mkinpredict [0.4s] -✔ | 9 | Multistart method for saem.mmkin models [37.3s] -✔ | 16 | Evaluations according to 2015 NAFTA guidance [2.5s] +✔ | 9 | Multistart method for saem.mmkin models [38.2s] +✔ | 16 | Evaluations according to 2015 NAFTA guidance [2.4s] ✔ | 9 | Nonlinear mixed-effects models with nlme [9.2s] -✔ | 16 | Plotting [10.1s] +✔ | 16 | Plotting [10.5s] ✔ | 4 | Residuals extracted from mkinfit models -✔ | 1 36 | saemix parent models [71.4s] +✔ | 1 36 | saemix parent models [75.0s] ──────────────────────────────────────────────────────────────────────────────── Skip ('test_saemix_parent.R:143'): 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 [1.4s] +✔ | 2 | Complex test case from Schaefer et al. (2007) Piacenza paper [1.6s] ✔ | 11 | Processing of residue series -✔ | 10 | Fitting the SFORB model [3.9s] +✔ | 10 | Fitting the SFORB model [4.0s] ✔ | 1 | Summaries of old mkinfit objects ✔ | 5 | Summary [0.2s] -✔ | 4 | Results for synthetic data established in expertise for UBA (Ranke 2014) [2.3s] -✔ | 9 | Hypothesis tests [8.3s] -✔ | 4 | Calculation of maximum time weighted average concentrations (TWAs) [2.2s] +✔ | 4 | Results for synthetic data established in expertise for UBA (Ranke 2014) [2.5s] +✔ | 9 | Hypothesis tests [9.0s] +✔ | 4 | Calculation of maximum time weighted average concentrations (TWAs) [2.3s] ══ Results ═════════════════════════════════════════════════════════════════════ -Duration: 220.0 s +Duration: 229.5 s ── Skipped tests ────────────────────────────────────────────────────────────── • Fitting this ODE model with saemix takes about 15 minutes on my system (1) @@ -55,3 +55,4 @@ Duration: 220.0 s • This still takes almost 2.5 minutes although we do not solve ODEs (1) [ FAIL 0 | WARN 0 | SKIP 3 | PASS 270 ] +Error while shutting down parallel: unable to terminate some child processes diff --git a/tests/testthat/setup_script.R b/tests/testthat/setup_script.R index c554800d..2448886d 100644 --- a/tests/testthat/setup_script.R +++ b/tests/testthat/setup_script.R @@ -2,17 +2,23 @@ require(mkin) require(testthat) # Per default (on my box where I set NOT_CRAN in .Rprofile) use all cores minus one +# Otherwise (CRAN check systems) use the allowed maximum of two cores if (identical(Sys.getenv("NOT_CRAN"), "true")) { n_cores <- parallel::detectCores() - 1 } else { - n_cores <- 1 + n_cores <- 2 } # Use the two available cores on travis if (Sys.getenv("TRAVIS") != "") n_cores = 2 -# On Windows we would need to make a cluster first -if (Sys.info()["sysname"] == "Windows") n_cores = 1 +# On Windows we need to make a cluster, or use one core +if (Sys.info()["sysname"] == "Windows") { + cl <- parallel::makePSOCKcluster(n_cores) + n_cores = 1 +} else { + cl <- parallel::makeForkCluster(n_cores) +} # Very simple example fits f_1_mkin_trans <- mkinfit("SFO", FOCUS_2006_A, quiet = TRUE) @@ -24,7 +30,7 @@ models <- c("SFO", "FOMC", "DFOP", "HS") fits <- suppressWarnings( # FOCUS A FOMC was, it seems, in testthat output mmkin(models, list(FOCUS_A = FOCUS_2006_A, FOCUS_C = FOCUS_2006_C, FOCUS_D = FOCUS_2006_D), - quiet = TRUE, cores = n_cores)) + quiet = TRUE, cluster = cl)) # One metabolite SFO_SFO <- mkinmod(parent = mkinsub("SFO", to = "m1"), @@ -82,13 +88,14 @@ fit_tc_1 <- mkinfit(m_synth_SFO_lin, SFO_lin_a, error_model = "tc", quiet = TRUE error_model_algorithm = "threestep") # Mixed model fits -mmkin_sfo_1 <- mmkin("SFO", ds_sfo, quiet = TRUE, error_model = "tc", cores = n_cores) -mmkin_dfop_1 <- mmkin("DFOP", ds_dfop, quiet = TRUE, cores = n_cores, +mmkin_sfo_1 <- mmkin("SFO", ds_sfo, quiet = TRUE, error_model = "tc", cluster = cl) +mmkin_dfop_1 <- mmkin("DFOP", ds_dfop, quiet = TRUE, cluster = cl, error_model = "tc") DFOP_SFO <- mkinmod(parent = mkinsub("DFOP", "m1"), m1 = mkinsub("SFO"), quiet = TRUE) -mmkin_dfop_sfo <- mmkin(list("DFOP-SFO" = DFOP_SFO), ds_dfop_sfo, quiet = TRUE, cores = n_cores, +mmkin_dfop_sfo <- mmkin(list("DFOP-SFO" = DFOP_SFO), ds_dfop_sfo, quiet = TRUE, + cluster = cl, control = list(eval.max = 500, iter.max = 400), error_model = "tc") @@ -105,3 +112,4 @@ dfop_saem_1 <- saem(mmkin_dfop_1, quiet = TRUE, transformations = "mkin", saem_dfop_sfo_m <- saem(mmkin_dfop_sfo, transformations = "mkin", quiet = TRUE) saem_dfop_sfo_s <- saem(mmkin_dfop_sfo, transformations = "saemix", quiet = TRUE) +parallel::stopCluster(cl) diff --git a/tests/testthat/test_saemix_parent.R b/tests/testthat/test_saemix_parent.R index 31605931..7fbecd0c 100644 --- a/tests/testthat/test_saemix_parent.R +++ b/tests/testthat/test_saemix_parent.R @@ -15,7 +15,7 @@ test_that("Parent fits using saemix are correctly implemented", { expect_silent(print(illparms(sfo_saem_1_reduced))) # We cannot currently do the fit with completely fixed initial values - mmkin_sfo_2 <- update(mmkin_sfo_1, fixed_initials = c(parent = 100)) + mmkin_sfo_2 <- update(mmkin_sfo_1, fixed_initials = c(parent = 100), cluster = NULL, cores = n_cores) sfo_saem_3 <- expect_error(saem(mmkin_sfo_2, quiet = TRUE), "at least two parameters") # We get an error if we do not supply a suitable model specification |