Remove readDataSetToEnd

R/HayesModels.R

@@ -78,31 +78,11 @@
   }
 
   if (number == 4) {
-    processRelationships <- list(
-      list(
-        processDependent = "Y",
-        processIndependent = "X",
-        processType = "mediators",
-        processVariable = "M"
-      )
-    )
+    processRelationships <- .generateProcessRelationships_4(k)
   }
 
   if (number == 5) {
-    processRelationships <- list(
-      list(
-        processDependent = "Y",
-        processIndependent = "X",
-        processType = "mediators",
-        processVariable = "M"
-      ),
-      list(
-        processDependent = "Y",
-        processIndependent = "X",
-        processType = "moderators",
-        processVariable = "W"
-      )
-    )
+    processRelationships <- .generateProcessRelationships_5(k)
   }
 
   if (number == 6) {
@@ -1568,6 +1548,42 @@
   return(.procEncodeProcessRelationships(processRelationships))
 } # end of function on hardcoded models
 
+.generateProcessRelationships_4 <- function(k) {
+  k <- max(k, 1)
+  processRelationships <- lapply(1:k, function(i) {
+    list(
+      processDependent = "Y",
+      processIndependent = "X",
+      processType = "mediators",
+      processVariable = paste0("M", i)
+    )
+  })
+  return(processRelationships)
+}
+
+.generateProcessRelationships_5 <- function(k) {
+  k <- max(k, 1)
+  processRelationships <- lapply(1:k, function(i) {
+    list(
+      processDependent = "Y",
+      processIndependent = "X",
+      processType = "mediators",
+      processVariable = paste0("M", i)
+    )
+  })
+
+  processRelationships <- append(processRelationships,
+    list(list(
+      processDependent = "Y",
+      processIndependent = "X",
+      processType = "moderators",
+      processVariable = "W"
+    ))
+  )
+
+  return(processRelationships)
+}
+
 # model 6
 .generateProcessRelationships_6 <- function(k) {
 

R/bayesianProcess.R

@@ -27,8 +27,6 @@ BayesianProcess <- function(jaspResults, dataset = NULL, options) {
     return()
   }
   options$naAction <- "listwise"
-  # Read dataset
-  dataset <- .procReadData(options)
   # Check for errors in dataset
   .procErrorHandling(dataset, options)
   # Create a container for each model

R/classicProcess.R

@@ -26,8 +26,6 @@ ClassicProcess <- function(jaspResults, dataset = NULL, options) {
     .procModelSummaryTable(jaspResults, options, NULL)
     return()
   }
-  # Read dataset
-  dataset <- .procReadData(options)
   # Check for errors in dataset
   .procErrorHandling(dataset, options)
   # Create a container for each model
@@ -83,7 +81,7 @@ ClassicProcess <- function(jaspResults, dataset = NULL, options) {
     "meanCenteredModeration", "standardizedModelEstimates", "errorCalculationMethod", "bootstrapCiType",
     "mcmcBurnin", "mcmcSamples", "mcmcChains", "seed", "setSeed", "nuPriorMu",
     "nuPriorSigma", "betaPriorMu", "betaPriorSigma", "psiPriorAlpha",
-    "psiPriorBeta", "rhoPriorAlpha", "rhoPriorBeta"
+    "psiPriorBeta", "rhoPriorAlpha", "rhoPriorBeta", "moderationProbes"
   ))
 }
 
@@ -394,10 +392,10 @@ ClassicProcess <- function(jaspResults, dataset = NULL, options) {
   # Get encoding
   encoding <- .procVarEncoding()
 
-  independent <- modelOptions[["modelNumberIndependent"]][["value"]]
-  mediators   <- modelOptions[["modelNumberMediators"  ]][["value"]]
-  modW        <- modelOptions[["modelNumberModeratorW" ]][["value"]]
-  modZ        <- modelOptions[["modelNumberModeratorZ" ]][["value"]]
+  independent <- modelOptions[["modelNumberIndependent"]]
+  mediators   <- modelOptions[["modelNumberMediators"  ]]
+  modW        <- modelOptions[["modelNumberModeratorW" ]]
+  modZ        <- modelOptions[["modelNumberModeratorZ" ]]
 
   # Apply the JASP coding to X, W, Z, and M if the user has specified the variables and 'vars' still contains the dummy version
   if (independent != "" && encoding$X %in% vars) {
@@ -468,16 +466,6 @@ ClassicProcess <- function(jaspResults, dataset = NULL, options) {
   return(graph)
 }
 
-.procReadData <- function(options) {
-  # Read in selected variables from dataset
-  dataset <- .readDataSetToEnd(
-    columns.as.numeric = c(options[['dependent']], options[['covariates']]),
-    columns.as.factor = options[['factors']]
-  )
-
-  return(dataset)
-}
-
 .procAddFactorDummyIntVars <- function(jaspResults, dataset, options) {
   modelsContainer <- jaspResults[["modelsContainer"]]
 
@@ -651,7 +639,7 @@ ClassicProcess <- function(jaspResults, dataset = NULL, options) {
   if (length(options[["covariates"]]) > 1) {
     .hasErrors(dataset, "run",
       type = "varCovData",
-      varCovData.target = options[["covariates"]],
+      varCovData.target = c(options[["dependent"]], options[["covariates"]]),
       varCovData.corFun = stats::cov,
       varCovData.corArgs = list(use = "complete.obs"),
       exitAnalysisIfErrors = TRUE
@@ -1314,6 +1302,18 @@ ClassicProcess <- function(jaspResults, dataset = NULL, options) {
   return(fittedModel)
 }
 
+.procCheckImpliedVarianceRatios <- function(fittedModel) {
+  impliedVariances <- diag(fittedModel@implied[["cov"]][[1]])
+
+  return(any(outer(impliedVariances, impliedVariances, FUN = "/") > 1000))
+}
+
+.procCheckImpliedVarianceMagnitudes <- function(fittedModel) {
+  impliedVariances <- diag(fittedModel@implied[["cov"]][[1]])
+
+  return(any(impliedVariances > 1000000))
+}
+
 .procResultsFitModel <- function(container, dataset, options, modelOptions) {
   # Check if graph has error message
   if (!.procCheckGraph(container[["graph"]]$object) && jaspBase::isTryError(container[["graph"]]$object)) {
@@ -1346,6 +1346,15 @@ ClassicProcess <- function(jaspResults, dataset = NULL, options) {
     return(.procLavaanMsg(fittedModel))
   }
 
+  if (doFit && is.null(lavaan::vcov(fittedModel))) {
+    checkImpliedVarianceRatios <- .procCheckImpliedVarianceRatios(fittedModel)
+    checkImpliedVarianceMagnitudes <- .procCheckImpliedVarianceMagnitudes(fittedModel)
+    return(.procNotIdentifiedMsg(
+      impliedVarianceRatios = checkImpliedVarianceRatios,
+      impliedVarianceMagnitudes = checkImpliedVarianceMagnitudes
+    ))
+  }
+
   if (doFit) {
     if (options$errorCalculationMethod == "bootstrap") {
       fittedModel <- .procBootstrap(fittedModel, options$bootstrapSamples)
@@ -1481,10 +1490,10 @@ ClassicProcess <- function(jaspResults, dataset = NULL, options) {
 
   # Create model options dummy object
   modelOptions <- list(
-    modelNumberIndependent = list(value = independent),
-    modelNumberMediators = list(value = mediators),
-    modelNumberModeratorW = list(value = modW),
-    modelNumberModeratorZ = list(value = modZ)
+    modelNumberIndependent = independent,
+    modelNumberMediators = mediators,
+    modelNumberModeratorW = modW,
+    modelNumberModeratorZ = modZ
   )
 
   # Check which hard-coded model matches user model
@@ -1511,7 +1520,7 @@ ClassicProcess <- function(jaspResults, dataset = NULL, options) {
   }
 }
 
-.procModelSummaryTable <- function(jaspResults, options, modelsContainer) {
+.procModelSummaryTable <- function(jaspResults, options, modelsContainer, errors) {
   if (!is.null(jaspResults[["modelSummaryTable"]])) return()
 
   modelNumbers <- lapply(options[["processModels"]], function(mod) {
@@ -1525,7 +1534,7 @@ ClassicProcess <- function(jaspResults, dataset = NULL, options) {
   modelNames <- sapply(options[["processModels"]], function(mod) mod[["name"]])
 
   procResults <- lapply(options[["processModels"]], function(mod) modelsContainer[[mod[["name"]]]][["fittedModel"]]$object)
-  
+
   if (length(procResults) == 0) return()
 
   # Remove invalid models
@@ -2775,6 +2784,17 @@ ClassicProcess <- function(jaspResults, dataset = NULL, options) {
 
 .procConvergenceMsg <- function() gettext("Model did not converge.")
 
+.procNotIdentifiedMsg <- function(impliedVarianceRatios = FALSE, impliedVarianceMagnitudes = FALSE) {
+  reasons <- c()
+  if (impliedVarianceRatios) {
+    reasons <- c(reasons, gettext("<li>Some implied model variances are > 1,000,000</li>"))
+  }
+  if (impliedVarianceMagnitudes) {
+    reasons <- c(reasons, gettext("<li>Some implied model variances are at least a factor 1000 times larger than others</li>"))
+  }
+  gettextf("The information matrix could not be inverted. This may be a symptom that the model is not identified. Possible reasons: <ul>%s</ul>", paste(reasons, collapse = ""))
+}
+
 .procModelIncompleteMsg <- function() gettext("At least one model is incomplete or no model is specified. Please add at least one model and complete specified models.")
 
 .procInformationCriteriaMsg <- function() gettext("The AIC, BIC and additional information criteria are only available with ML-type estimators.")

inst/Description.qml

@@ -12,6 +12,7 @@ Description
 	maintainer	: "JASP Team <[email protected]>"
 	website		: "https://jasp-stats.org"
 	license		: "GPL (>= 2)"
+	preloadData : true
 
 	GroupTitle
 	{

inst/qml/BayesianProcess.qml

@@ -55,7 +55,6 @@ Form
                     Common.InputVariables {
 						visible: inputType.value == "inputVariables"
 						adjustedWidth: models.width - 2 * jaspTheme.contentMargin
-						colWidth: (models.width - 3 * 40 * preferencesModel.uiScale) / 4
 					}
 
                     Common.InputModelNumber {

inst/qml/ClassicProcess.qml

@@ -26,13 +26,11 @@ Form
 {
 	Common.VariablesForm {}
 
-
 	Section
 	{
 		title: qsTr("Models")
 		columns: 1
 
-
 		TabView
 		{
 			id:				models
@@ -44,6 +42,7 @@ Form
 			content: Group
 			{
 				childControlsArea.anchors.leftMargin: jaspTheme.contentMargin
+
 				Common.InputType
 				{
 					id: inputType
@@ -56,7 +55,6 @@ Form
 				{
 					visible: inputType.value == "inputVariables"
 					adjustedWidth: models.width - 2 * jaspTheme.contentMargin
-					colWidth: (models.width - 3 * 40 * preferencesModel.uiScale) / 4
 				}
 
 				Common.InputModelNumber
@@ -230,6 +228,7 @@ Form
 
 		Common.PathPlotOptions {}
 	}
+
 	Section
 	{
 		id: advanced

inst/qml/common/InputModelNumber.qml

@@ -52,7 +52,7 @@ Group
 		{
 			name: 				"modelNumberMediators"
 			title: 				qsTr("Mediators M")
-			allowedColumns: 	["scale", "ordinal"]
+			allowedColumns: 	["scale", "nominal"]
 		}
 		// TODO
 		// AssignedVariablesList

inst/qml/common/InputVariables.qml

@@ -23,38 +23,10 @@ import JASP.Controls
 
 Group
 {
-	property int adjustedWidth: parent.width
-
 	id: 		modelsGroup
 
-	property int colWidth: parent.width / 4 - jaspTheme.contentMargin
-	property int labelWidth: modelsGroup.colWidth + jaspTheme.contentMargin
-
-	RowLayout
-	{
-		Layout.margins: 	jaspTheme.contentMargin
-		Label
-		{
-			text: qsTr("From")
-			Layout.preferredWidth: modelsGroup.labelWidth
-		}
-		Label
-		{
-
-			text: qsTr("To")
-			Layout.preferredWidth: modelsGroup.labelWidth
-		}
-		Label
-		{
-			text: qsTr("Process Type")
-			Layout.preferredWidth: modelsGroup.labelWidth
-		}
-		Label
-		{
-			text: qsTr("Process Variable")
-			Layout.preferredWidth: modelsGroup.labelWidth
-		}
-	}
+	property int adjustedWidth: parent.width
+	property int colWidth: (adjustedWidth / 4)  - 4 * jaspTheme.contentMargin
 
 	ComponentsList
 	{
@@ -63,18 +35,20 @@ Group
 		preferredWidth: 		adjustedWidth
 		itemRectangle.color: 	jaspTheme.controlBackgroundColor
 		minimumItems:           1
+		headerLabels:			[qsTr("From "), qsTr("To"), qsTr("Process Type"), qsTr("Process Variable")]
 		rowComponent: 			RowLayout
 		{
 			id: 		rowComp
 			enabled: 	rowIndex === relations.count - 1
+			spacing:	jaspTheme.contentMargin
 
 			Layout.columnSpan: 4
 			DropDown
 			{
 				id: 				    procIndep
 				name: 				    'processIndependent'
 				source: 			    ['covariates', 'factors']
-				controlMinWidth: 	    modelsGroup.colWidth
+				fieldWidth:			    modelsGroup.colWidth
 				addEmptyValue: 		    true
 				onCurrentValueChanged:
 				{
@@ -93,7 +67,7 @@ Group
 				id: 				procDep
 				name: 				'processDependent'
 				source: 			["dependent", "processVariable"] //, {name: "processRelationships.processVariable", use: "discardIndex=" + (relations.count - 1)}]
-				controlMinWidth: 	modelsGroup.colWidth
+				fieldWidth:			modelsGroup.colWidth
 				addEmptyValue: 		true
 				onCurrentValueChanged:
 				{
@@ -118,7 +92,7 @@ Group
 					{ label: qsTr("Confounder"), 	value: 'confounders'	},
 					{ label: qsTr("Direct"), 		value: 'directs'		}
 				]
-				controlMinWidth: 	modelsGroup.colWidth
+				fieldWidth:			modelsGroup.colWidth
 				addEmptyValue: 		true
 				onCurrentValueChanged:
 				{
@@ -134,7 +108,7 @@ Group
 				name: 				    'processVariable'
 				enabled: 			    procType.currentValue != "directs"
 				source: 			    procType.currentValue == 'mediators' ? ['covariates'] : ['covariates', 'factors']
-				controlMinWidth: 	    modelsGroup.colWidth
+				fieldWidth:				modelsGroup.colWidth
 				addEmptyValue: 		    true
 				onCurrentValueChanged:
 				{