-
-
Notifications
You must be signed in to change notification settings - Fork 210
Making a QML interface
QML (Qt Modeling Language) is a user interface markup language that JASP uses to show the analysis input panel. In this panel the user can specify what options should be set to what values and thereby change the tables and plots that the analysis computes. QML is a very flexible language that allows us to easily generate checkboxes, dropdowns and other common interface components. It also gives the possiblility to insert JavaScript expressions to dynamically alter the interface based on actions a user makes. To create a more uniform layout and make it easier to add new analyses we have provided a number of standardized components. These components should satify most analysis creators, although you can always add your own.
Table of Contents:
The components can roughly be divided in three classes. One that deals with general inputs (e.g., checkboxes), one that deals with assigning variables and one that groups components together. Each will be covered in the following section. Some remarks about these components:
- They are all QML items, so they automatically get generic QML properties like
enabled
orvisible
. - The properties of the components can be set directly with a value. You can use also a JavaScript expression so that the value will depend on other components (see the section on Connecting Multiple Components).
- In several examples you may encounter
qsTr()
, it is important that this function wraps all text that will be shown in the interface. It will provide the possibility to translate JASP in the future. - The components described below may generally be nested to an arbitrary level (e.g., a checkbox in a group in a checkbox).
These components are quite common in questionnaires and input forms on websites, they include the checkbox, radiobutton, dropdown, slider and textfields where text may be entered.
Check button that can be toggled on and off. If some components are nested inside a CheckBox, they are automatically enabled or disabled when the CheckBox is set or not. Properties
-
name
: string identifier (in your R code you will be able to retrieve the value of the checkbox through this identifier) -
label
: text that will be shown to the left of the checkbox -
checked
: [optional, default:false
] boolean specifying if it should be set per default on or not. -
childrenOnSameRow
: [optional, default:false
] boolean specifying if components (e.g., other checkboxes, textfields) nested within the checkbox should be shown on the same row or, alternatively, below the checkbox -
columns
: [optional, default:1
] integer specifying how many columns the nested components should occupy (e.g., when set to 3, with three nested checkboxes, these three checkboxes will appear side by side on the same horizontal row)
Examples
CheckBox { name: "pearson"; label: qsTr("Pearson"); checked: true }
CheckBox
{
name: "homogeneityCorrections"
label: qsTr("Homogeneity corrections")
columns: 3
CheckBox { name: "homogeneityNone"; label: qsTr("None") ; checked: true }
CheckBox { name: "homogeneityBrown"; label: qsTr("Brown-Forsythe") ; checked: true }
CheckBox { name: "homogeneityWelch"; label: qsTr("Welch") ; checked: true }
}
A RadioButton is used with other RadioButton's inside a RadioButtonGroup
. Contrary to CheckBox, only 1 RadioButton can be checked at the same time.
RadioButtonGroup properties:
-
name
: string identifier (in your R code you will be able to retrieve the value of the checked radiobutton through this identifier) -
title
: text that is shown above the set of radiobuttons -
radioButtonsOnSameRow
: [optional, default:false
] per default, the RadioButton's are placed vertically under the title. Set this property to true to place the buttons horizontally. -
columns
: [optional, default:1
] integer specifying how many columns should be used to display the RadioButton's
RadioButton properties:
-
value
: the value of theRadioButtonGroup
that is send to R when this radiobutton is checked -
label
: the text that is shown to the right of the radiobutton -
checked
: [optional, default:false
] boolean specifying if it should contain a checkmark (and thus be the default); one radiobutton should have this property set totrue
-
childrenOnSameRow
: [optional, default:false
] boolean specifying if components (e.g., checkboxes, textfields) nested within the radiobutton should be shown on the same row or, alternatively, below the radiobutton -
columns
: [optional, default:1
] integer specifying how many columns the nested components should occupy (e.g., when set to 3, with three nested checkboxes, these three checkboxes will appear side by side on the same horizontal row)
Examples
RadioButtonGroup
{
name: "countProp"
title: qsTr("Display")
RadioButton { value: "descCounts"; label: qsTr("Counts"); checked: true }
RadioButton { value: "descProps"; label: qsTr("Proportions") }
}
RadioButtonGroup
{
name: "steppingMethodCriteriaType"
title: qsTr("Stepping Method Criteria")
RadioButton
{
value: "usePValue"; label: qsTr("Use p value"); checked: true
columns: 2
DoubleField { name: "steppingMethodCriteriaPEntry"; label: qsTr("Entry"); fieldWidth: 60; defaultValue: 0.05; max: 1; decimals: 3 }
DoubleField { name: "steppingMethodCriteriaPRemoval"; label: qsTr("Removal"); fieldWidth: 60; defaultValue: 0.1; max: 1; decimals: 3 }
}
RadioButton
{
value: "useFValue"; label: qsTr("Use F value")
columns: 2
DoubleField { name: "steppingMethodCriteriaFEntry"; label: qsTr("Entry"); fieldWidth: 60; defaultValue: 3.84; decimals: 3 }
DoubleField { name: "steppingMethodCriteriaFRemoval"; label: qsTr("Removal"); fieldWidth: 60; defaultValue: 2.71; decimals: 3 }
}
}
Properties
-
name
: string identifier (in your R code you will be able to retrieve the value of the dropdown through this identifier) -
label
: [optional, default:""
] text that will be shown above the dropdown -
values
: array of (named) values that is shown in the dropdown list, in the case of an unnamed array the value and the label are the same, but when a named array is provided the label shown to the user can be different from the value send to R (see example below) -
indexDefaultValue
: [optional, default:1
] integer specifying the index of the dropdown item that is selected by default, take note that this is zero-based -
source
: ifvalues
is empty, you can use the name of a VariablesList (see Variable Specification) to set the source of the values of the DropDown.
Examples
DropDown
{
name: "orthogonalSelector"
values: ["none", "varimax", "quartimax", "bentlerT", "equamax", "varimin"]
}
DropDown
{
name: "sumOfSquares"
indexDefaultValue: 2
label: qsTr("Sum of squares")
values:
[
{ label: "Type \u2160"; value: "type1"}
{ label: "Type \u2161"; value: "type2"}
{ label: "Type \u2162"; value: "type3"}
]
}
Slider is used to select a value by sliding a handle along a track. Properties
-
name
: string identifier (in your R code you will be able to retrieve the value of the slider through this identifier) -
label
: [optional, default:""
] text that will be shown above the slider -
value
: [optional, default:0.5
] default value -
min
: [optional, default:0
] minimum value -
max
: [optional, default:1
] maximum value -
vertical
: [optional, default: true] set whether the slider should be displayed vertically or horizontally
Examples
Slider
{
name: "highlight"
label: qsTr("Highlight")
value: 0.4
}
Properties
-
name
: string identifier (in your R code you will be able to retrieve the value of the field through this identifier) -
label
: [optional, default:""
] text that will be shown to the left of the field -
afterLabel
: [optional, default:""
] text that will be shown to the right of the field -
defaultValue
: [optional, default:0
] numeric specifying the default value shown; can be an integer or a decimal value, just be sure to use a dot and not a comma -
negativeValues
: [optional, default:false
] specifies if the user should be able to input negative values (works only when themin
argument is omitted) -
min
: [optional, default:-Infinity
] numeric specifying the minimum value a user can enter -
max
: [optional, default:Infinity
] numeric specifying the maximum value a user can enter -
decimals
: [optional, default:3
] integer specifying how many decimals the user can enter -
fieldWidth
: [optional, default:40
] in pixels how wide should the field be
Examples
DoubleField
{
name: "eigenValuesBox"
label: qsTr("Eigenvalues above")
defaultValue: 1
decimals: 1
}
DoubleField
{
name: "priorFixedEffects"
label: qsTr("r scale fixed effects")
defaultValue: 0.5
fieldWidth: 50
max: 2
decimals: 1
}
Properties
-
name
: string identifier (in your R code you will be able to retrieve the value of the field through this identifier) -
label
: [optional, default:""
] text that will be shown to the left of the field -
afterLabel
: [optional, default:""
] text that will be shown to the right of the field -
defaultValue
: [optional, default:0
] integer specifying the default value shown -
negativeValues
: [optional, default:false
] specifies if the user should be able to input negative values (works only when themin
argument is omitted) -
min
: [optional, default:0
] integer specifying the minimum value a user can enter -
max
: [optional, default:Infinity
] integer specifying the maximum value a user can enter -
fieldWidth
: [optional, default:40
] in pixels how wide should the field be
Examples
IntegerField
{
name: "fixedSamplesNumber"
label: qsTr("No. samples")
defaultValue: 10000
fieldWidth: 60
}
IntegerField
{
name: "percentileValuesEqualGroupsNo"
min: 1
max: 1000
defaultValue: 4
afterLabel: qsTr(" equal groups")
}
Properties
-
name
: string identifier (in your R code you will be able to retrieve the value of the field through this identifier) -
label
: text that will be shown to the left of the field -
defaultValue
: [optional, default:50
] integer specifying the default percentage shown -
fieldWidth
: [optional, default:40
] in pixels how wide should the field be -
decimals
: [optional, default:0
] integer specifying how many decimals the user can enter
Examples
PercentField { name: "estimatesPlotsCI"; label: qsTr("Confidence interval"); defaultValue: 95 }
Properties:
-
name
: string identifier (in your R code you will be able to retrieve the value of the field through this identifier) -
label
: [optional, default:""
] text that will be shown to the left of the field -
afterLabel
: [optional, default:""
] text that will be shown to the right of the field -
defaultValue
: [optional, default:""
] default text before the user enters anything -
placeholderText
: [optional, default:""
] text shown as a placeholder until a user enters something, will not be send to R if left unchanged by the user (mutually exclusive withdefaultValue
) -
fieldWidth
: [optional, default:40
] in pixels how wide should the field be
Examples
TextField { name: "labelYAxis"; label: qsTr("Label y-axis"); fieldWidth: 200 }
Most analyses are performed on variables. JASP offers a few ways of visualizing these variables in the input form. The general idea is that you get a field of available variables (taken from the dataset) and then one or more fields that the variables can be dragged to. Variable fields should be wrapped in a VariablesForm
. This makes it possible to automatically align multiple variable fields and add assign-buttons.
Properties
-
name
: identifier of the variables list, this is never send to R -
label
: [optional, default:""
] text that will be shown above the variable field -
source
: [optional] can be set to thename
(or a list of names) of anAssignedVariablesList
, when omitted variables are taken from the dataset (example usage: you have a secondAvailableVariablesList
for plotting the variables that are assigned to anAssignedVariablesList
; see also the example) -
width
: [optional, default: 2/5 of the VariablesForm width] in pixels how wide should the field be
Note: height
should be defined on VariablesForm
itself.
Examples
VariablesForm
{
AvailableVariablesList { name: "allVariables" }
AssignedVariablesList { name: "fixedFactors"; label: qsTr("Fixed Factors"); allowedColumns: ["ordinal", "nominal"] }
}
VariablesForm
{
height: 200
AvailableVariablesList { name: "postHocTestsAvailable"; source: "fixedFactors" }
AssignedVariablesList { name: "postHocTestsVariables" }
}
Properties
-
name
: identifier of the particular variable field (in your R code you will be able to retrieve the assigned variable(s) through this identifier) -
label
: [optional, default:""
] text that will be shown above the variable field -
allowedColumns
: [optional, default:["scale", "ordinal", "nominal"]
] array specifying the allowed column types -
singleVariable
: [optional, default:false
] boolean specifying if this field will accept a maximum of one variable -
listViewType
: [optional] string that specifies the type ofAssignedVariablesList
, when omitted we get a normal list, options are"Layers"
(see Contingency Tables),"Pairs"
(see Paired Samples T-Test),"Interaction"
(see ANOVA) and"RepeatedMeasures"
(see Repeated Measures ANOVA) -
ExtraControlColumn
: [optional] addititional general input component (e.g.,"CheckBox"
) which is added to each variable and can be used to add extra control (e.g., you can specify the status of the variable within a linear model, see example below) -
width
: [optional, default: 2/5 of the VariablesForm width] in pixels how wide should the field be -
height
: [optional] in pixels how heigh should the field be. Per default, it is set so that all AssignedVariablesList's fit the VariablesForm. If you set the height for 1 AssignedVariablesList, it will try to set height of the other AssignedVariablesLists's so that they all fit the heigth of the VariablesForm.
Examples
VariablesForm
{
AvailableVariablesList { name: "allVariables" }
AssignedVariablesList
{
name: "dependent"
label: qsTr("Dependent Variable")
allowedColumns: ["scale"]
singleVariable: true
}
}
}
VariablesForm
{
AvailableVariablesList { name: "allVariables" }
AssignedVariablesList
{
name: "modelTerms"
label: qsTr("Model terms")
listViewType: "Interaction"
ExtraControlColumn
{
type: "CheckBox"
name: "isNuisance"
title: "Add to null model"
}
}
}
Properties
-
name
: identifier of the particular variable field (in your R code you will be able to retrieve the assigned variable(s) through this identifier) -
label
: [optional, default:""
] text that will be shown above the variable field -
width
: [optional, default: ±230
] in pixels how wide should the field be -
height
: [optional, default:350
] in pixels how heigh should the field be
Examples
VariablesForm
{
RepeatedMeasuresFactorsList
{
name: "repeatedMeasuresFactors"
label: qsTr("Repeated Measures Factors")
height: 180
}
AssignedVariablesList
{
name: "repeatedMeasuresCells"
label: qsTr("Repeated Measures Cells")
allowedColumns: ["scale"]
listViewType: "MeasuresCells"
source: "repeatedMeasuresFactors"
height: 140
}
}
In order to add more structure to the input panel you can group components together. There are 2 levels of grouping: Section and Group. A Section can hide a group of components under a button. A Group is a smaller logical unit.
Properties
-
title
: text shown above the grouped input components -
columns
: [optional, default:1
] integer specifying how many columns the grouped components should occupy
Examples
Group
{
title: qsTr("Output")
CheckBox
{
name: "effects"; label: qsTr("Effects")
RadioButtonGroup
{
name: "effectsType"
RadioButton { value: "allModels"; label: qsTr("Across all models"); checked: true }
RadioButton { value: "matchedModels"; label: qsTr("Across matched models") }
}
}
CheckBox { name: "descriptives"; label: qsTr("Descriptives") }
}
A Section sets components (or groups of components) under a button. By clicking the button, you can hide or display these components. Properties
-
title
: text shown in the button that controls the collapse of an entire section -
columns
: [optional, default:2
] integer specifying how many columns the grouped components should occupy
Examples
Section
{
title: qsTr("Advanced Options")
RadioButtonGroup
{
title: qsTr("Missing Values")
name: "missingValues"
RadioButton { value: "excludeCasesListwise"; label: qsTr("Exclude cases listwise"); checked: true }
RadioButton { value: "excludeCasesPairwise"; label: qsTr("Exclude cases pairwise") }
}
Group
{
title: qsTr("Confidence Interval")
enabled: chronbach.checked
CheckBox
{
name: "confAlpha"
label: qsTr("Cronbach's α analytical")
PercentField { name: "confAlphaLevel"; label: qsTr("Confidence"); defaultValue: 95 }
}
}
}
By default JASP creates a two column grid that is filled row-wise. So each component you place in your QML file will be added to the input panel left to right, top to bottom. We will demonstrate what this means by the use of a table.
Imagine a simple table with two columns and two rows. We have three components (CheckBox A, B and C) and each must go in their own cell. When these components are added row-wise we get A in the top left cell, then next to it B and on its own row C:
<column 1> | <column 2> | |
---|---|---|
<row 1> | CheckBox A | CheckBox B |
<row 2> | CheckBox C |
The above works great, but it definitely helps that the checkboxes are the same size. We could also add a RadioButtonGroup that is much taller than a simple checkbox. If we again imagine three components (but this time CheckBox A and B and RadioButtonGroup A) that go in the table in the same way:
<column 1> | <column 2> | |
---|---|---|
<row 1> | CheckBox A | RadioButtonGroup A |
RadioButtonGroup A | ||
<row 2> | CheckBox B |
We notice that there is a lot of space between CheckBox A and CheckBox B. The reason for this is that CheckBox B is placed in the next available row. Fortunately we can remedy this with the property Layout.rowSpan
; all we need to do is set Layout.rowSpan: 2
in RadioButtonGroup A. This tells the engine that the component spans two rows and therefore CheckBox B will fit snuggly under CheckBox A:
<column 1> | <column 2> | |
---|---|---|
<row 1> | CheckBox A | RadioButtonGroup A |
<row 2> | CheckBox B | RadioButtonGroup A |
Now imagine we really wanted to put the RadioButtonGroup in the top left cell and the two checkboxes below it on the same height as one another:
<column 1> | <column 2> | |
---|---|---|
<row 1> | RadioButtonGroup A | |
RadioButtonGroup A | ||
<row 2> | CheckBox A | CheckBox B |
However, if we simply added the components we would end up with this situation:
<column 1> | <column 2> | |
---|---|---|
<row 1> | RadioButtonGroup A | CheckBox A |
RadioButtonGroup A | ||
<row 2> | CheckBox B |
To accomplish this we can set Layout.columnSpan: 2
on RadioButtonGroup A. This results in the following situation:
<column 1> | <column 2> | |
---|---|---|
<row 1> | RadioButtonGroup A | RadioButtonGroup A |
RadioButtonGroup A | RadioButtonGroup A | |
<row 2> | CheckBox A | CheckBox B |
In AvailableVariablesList
we showed that through its source
property, we can establish a link between two components. There are other ways of connecting components and they are not limited to a VariableList
(although they are a little different, only a VariableList
uses the name
property). QML allows us to do this is by adding an id
to one component and then referencing this id
in a different component. Note that the id
can be any string starting with a lowercase, but it must be unique.
An example is enabling a checkbox if one of two different checkboxes is checked.
Implementation
CheckBox { name: "checkboxA"; label: qsTr("Some label"), id: checkA}
CheckBox { name: "checkboxB"; label: qsTr("Some label"), id: checkB}
CheckBox { name: "checkboxC"; label: qsTr("Some Label too"); enabled: checkA.checked || checkB.checked }
Here we make use of a JavaScript expression to evaluate if either CheckBox A or CheckBox B has been checked and if this is the case we enable CheckBox C. This JavaScript expression will be automatically updated each time that the checked values of checkA and checkB changes.
Another example would be setting the visibility of some textfield to invisible if a checkbox is not checked.
Implementation
CheckBox
{
name: "checkboxA"
label: qsTr("Some label")
id: checkA
TextField { name: "textfieldA"; afterlabel: qsTr("x"); visible: checkA.checked}
}
Any property can be set with an expression. A title of an ExpanderButton for example:
Implementation
DropDown
{
id: estimator
name: "estimator"
label: qsTr("Estimator")
values: ["EBICglasso", "cor", "pcor", "IsingFit", "IsingSampler", "huge", "adalasso", "mgm"]
}
ExpanderButton
{
title: qsTr("Analysis Options - ") + estimator.currentText
....
}
We'll create a simple analysis input panel to show the workflow. Read first the Guide to adding a module in JASP. In this way, when you edit your QML file, the changes made will be automatically seen in JASP. You can play with the components and their properties, and see immediately the output in JASP.
We can begin actual work on the QML file, first we have to tell the engine where to find our resources. To do so, we add a number of imports to the top of our file.
Code
import QtQuick 2.11
import QtQuick.Layouts 1.3
import JASP.Controls 1.0
import JASP.Widgets 1.0
At this point we add a Form
which will hold all our input components:
Code
import QtQuick 2.11
import QtQuick.Layouts 1.3
import JASP.Controls 1.0
import JASP.Widgets 1.0
Form
{
}
It's now a matter of mixing and matching the previously shown components to create a form to our liking. Of course, if something isn't quite possible, you can also use QML features that were not covered here.
Code
import QtQuick 2.11
import QtQuick.Layouts 1.3
import JASP.Controls 1.0
import JASP.Widgets 1.0
Form
{
VariablesForm
{
AssignedVariablesList { name: "variables"; allowedColumns: ["ordinal", "nominal"] }
}
RadioButtonGroup
{
Layout.rowSpan: 2
name: "hypothesis"
title: qsTr("Hypothesis")
RadioButton { value: "notEqualToTestValue"; label: qsTr("≠ Test value"); checked: true }
RadioButton { value: "greaterThanTestValue"; label: qsTr("> Test value") }
RadioButton { value: "lessThanTestValue"; label: qsTr("< Test value") }
}
DoubleField { name: "testValue"; label: qsTr("Test value: "); defaultValue: 0.5 ; max: 1; decimals: 2}
Group
{
title: qsTr("Plots")
CheckBox
{
name: "descriptivesPlots"
label: qsTr("Descriptive plots")
PercentField { name: "descriptivesPlotsConfidenceInterval"; label: qsTr("Confidence Interval"); defaultValue: 95 }
}
}
}