Release fixes (#179)

* make 2x2 error consistent between bayes and freq * bayesian log-linear reg. uses first level as the reference level * more user friendly error message * remove base:: and use nlevels Co-authored-by: Don van den Bergh <[email protected]> --------- Co-authored-by: Don van den Bergh <[email protected]>
jasp-stats · Nov 29, 2023 · 79d1735 · 79d1735
1 parent fe71901
commit 79d1735
Show file tree

Hide file tree

Showing 3 changed files with 50 additions and 34 deletions.
diff --git a/R/contingencytablesbayesian.R b/R/contingencytablesbayesian.R
@@ -559,8 +559,8 @@ ContingencyTablesBayesianInternal <- function(jaspResults, dataset = NULL, optio
 
     if (!identical(dim(counts.matrix), as.integer(c(2,2)))) {
       message <- gettext("Odds ratio restricted to 2 x 2 tables")
-      analysisContainer[["contTabBasLogOdds"]]$addFootnote(message)
-
+      analysisContainer[["contTabBasLogOdds"]]$setError(message)
+      return()
     } else if ( options$samplingModel == "hypergeometric") {
       row[["value[oddsRatio]"]] <- NaN
       row[["low[oddsRatio]"]]   <- row[["up[oddsRatio]"]] <- "" #Really? What order is this evaluated in?

diff --git a/R/regressionloglinearbayesian.R b/R/regressionloglinearbayesian.R
@@ -34,17 +34,28 @@ RegressionLogLinearBayesianInternal <- function(jaspResults, dataset = NULL, opt
 
 # Preprocessing functions
 .basRegLogLinReadData <- function(dataset, options) {
-  if (!is.null(dataset))
-    return(dataset)
-  else {
+  if (is.null(dataset)) {
     counts <- factors <- NULL
     if(options$count != "")
       counts <- options$count
     if(length(options$modelTerms) > 0)
       factors <- options$factors
-    return(.readDataSetToEnd(columns.as.factor = factors,
-                             columns.as.numeric = counts))
+    dataset <- .readDataSetToEnd(columns.as.factor = factors,
+                                 columns.as.numeric = counts)
+
+    # conting uses the last level as the reference level,
+    # but elsewhere we use the first level instead.
+    # So, we shift the first level to be the last level to keep the output consistent
+    for (fac in factors) {
+      var <- dataset[[fac]]
+      if (nlevels(var) > 1) {
+        lev <- levels(var)
+        dataset[[fac]] <- factor(var, levels = c(lev[-1], lev[1]))
+      }
+    }
   }
+
+  return(dataset)
 }
 
 .basRegLogLinCheckErrors <- function(dataset, options) {
@@ -164,7 +175,7 @@ RegressionLogLinearBayesianInternal <- function(jaspResults, dataset = NULL, opt
       if(grepl(pattern = "the leading minor of order [0-9]+ is not positive definite", x = msg)) {
         msg <- gettext("Cannot compute the results; a numerical error occurred during sampling. Try to change (e.g., simplify) the model or adjust priors.")
       }
-      stop(gettextf("R Package error: %s", msg))
+      stop(gettextf("R package 'conting' error: %s <br> It is possible that a numerical error occured during sampling. Try to change (e.g., simplify) the model, adjust priors, change seed, or change the number of MCMC samples.", msg))
     }
   }
 

diff --git a/tests/testthat/test-regressionloglinearbayesian.R b/tests/testthat/test-regressionloglinearbayesian.R
@@ -45,23 +45,22 @@ test_that("General summary statistics table matches", {
   results <- jaspTools::runAnalysis("RegressionLogLinearBayesian", "test.csv", options)
   table <- results[["results"]][["Container"]][["collection"]][["Container_SummaryTable"]][["data"]]
   jaspTools::expect_equal_tables(table,
-    list("(Intercept)", 1, 2.28941355597128, 0.0114477565469203, 2.12177466183418,
-         2.45495750281602, "contBinom = 0", 1, 0.0303566571915708, 0.00356265251723887,
-         -0.0378564465631073, 0.144271031109649, "facFive = 1", 1, -0.00552214172853799,
-         0.00720162344071968, -0.130345670875483, 0.138580850953902,
-         "facFive = 2", 1, -0.00240692298068409, 0.00784162472732633,
-         -0.121705317175139, 0.165537141029262, "facFive = 3", 1, 0.00293548076176623,
-         0.00745020526106448, -0.0900898563442242, 0.187558384342406,
-         "facFive = 4", 1, 0.00138448207875663, 0.00779485819569601,
-         -0.122842441315205, 0.155851155312602, "contBinom = 0*facFive = 1",
-         0.571428571428571, 0.00197519410432235, 0.00569094808602158,
-         -0.125500914290167, 0.103584632765005, "contBinom = 0*facFive = 2",
-         0.571428571428571, -0.0270367686155361, 0.00624820819787772,
-         -0.180746332353481, 0.0610871410555307, "contBinom = 0*facFive = 3",
-         0.571428571428571, 0.00343309539396836, 0.00598953855547496,
-         -0.145370964229942, 0.0973020746027855, "contBinom = 0*facFive = 4",
-         0.571428571428571, -0.0218844778500034, 0.00501003480774965,
-         -0.131413707533984, 0.101000867077934)
+                                 list("(Intercept)", 2.12746929269443, 2.28118363693666, 1, 0.00959232762655234,
+                                      2.43233357540468, "contBinom = 1", -0.1140552657318, -0.0335221198542702,
+                                      1, 0.002756395445979, 0.0475020520551498, "facFive = 2", -0.195453489777316,
+                                      -0.00282655167800188, 1, 0.0102248921960175, 0.155047320962659,
+                                      "facFive = 3", -0.164264352414958, 0.000154400018643484, 1,
+                                      0.00940048687056697, 0.161656677341903, "facFive = 4", -0.14973534001884,
+                                      0.00369626302015217, 1, 0.00974587350979484, 0.189539723263921,
+                                      "facFive = 5", -0.155273680656904, 0.0016839574489477, 1, 0.0108516685524461,
+                                      0.166831941585999, "contBinom = 1*facFive = 2", -0.0819863860420921,
+                                      0.0507141965203517, 0.447023809523808, 0.0117408712962367, 0.225608774669159,
+                                      "contBinom = 1*facFive = 3", -0.172855306187846, -0.0133761570465655,
+                                      0.447023809523808, 0.00900801924647208, 0.144539535306823, "contBinom = 1*facFive = 4",
+                                      -0.0818337842044022, 0.0427001178126396, 0.447023809523808,
+                                      0.00995671151898941, 0.241519319460282, "contBinom = 1*facFive = 5",
+                                      -0.289531236785577, -0.0761428540661306, 0.447023809523808,
+                                      0.0132782059648042, 0.0542974169718703)
   )
 })
 
@@ -81,15 +80,21 @@ test_that("Submodel summary statistics table matches", {
   results <- jaspTools::runAnalysis("RegressionLogLinearBayesian", "test.csv", options)
   table <- results[["results"]][["Container"]][["collection"]][["Container_SubSummaryTable"]][["data"]]
   jaspTools::expect_equal_tables(table,
-    list("(Intercept)", 2.29560729883006, 0.00945972825329099, 2.12809954463567,
-         2.52220705393406, "contBinom = 0", 0.045757962353209, 0.00487119274553831,
-         -0.0378564465631073, 0.235518792759572, "facFive = 1", -0.00411654631578624,
-         0.00919225838347475, -0.189678327954162, 0.225524507296169,
-         "facFive = 2", -0.00755884938209783, 0.0107688538955323, -0.230464286621307,
-         0.203303292163532, "facFive = 3", 0.00693535371456428, 0.00958840576151784,
-         -0.151633708350598, 0.242411127225971, "facFive = 4", 0.00306711588492548,
-         0.0107120223942474, -0.234264000793185, 0.230287649489829)
-  )
+                                 list("(Intercept)", 2.12746929269443, 2.27823739725555, 0.00954231418605421,
+                                      2.47945144261042, "contBinom = 1", -0.152571436550678, -0.0308665527935553,
+                                      0.00251737809993112, 0.0428060084103773, "facFive = 2", -0.172121681790293,
+                                      0.00224525307377201, 0.00878383066077853, 0.230919007122397,
+                                      "facFive = 3", -0.196710458257331, -0.0044781058298241, 0.00835331749889998,
+                                      0.208709120970735, "facFive = 4", -0.241047231509551, 0.00449593617720325,
+                                      0.00801996356268645, 0.162572205036627, "facFive = 5", -0.226987657684912,
+                                      0.000215113407450374, 0.011947132025284, 0.232769962304245,
+                                      "contBinom = 1*facFive = 2", -0.153533409328021, 0.0507141965203517,
+                                      0.0117408712962367, 0.307061040408055, "contBinom = 1*facFive = 3",
+                                      -0.213739201091917, -0.0133761570465655, 0.00900801924647208,
+                                      0.184451333453313, "contBinom = 1*facFive = 4", -0.148024819514679,
+                                      0.0427001178126396, 0.00995671151898941, 0.248503021545585,
+                                      "contBinom = 1*facFive = 5", -0.344769999413831, -0.0761428540661306,
+                                      0.0132782059648042, 0.0669949682186052))
 })
 
 test_that("Analysis handles errors - Infinity", {