/*

* Project Name : Visual Python

* Description : GUI-based Python code generator

* File Name : Anova.js

* Author : Black Logic

* Note : ANOVA

* License : GNU GPLv3 with Visual Python special exception

* Date : 2023. 05. 24

* Change Date :

*/

//============================================================================

// [CLASS] Anova

//============================================================================

define([

__VP_TEXT_LOADER__('vp_base/html/m_stats/anova.html'),

'vp_base/js/com/com_util',

'vp_base/js/com/com_Const',

'vp_base/js/com/com_String',

'vp_base/js/com/com_generatorV2',

'vp_base/js/com/component/PopupComponent',

'vp_base/js/com/component/DataSelector',

'vp_base/js/m_apps/Subset'

], function(nmHTML, com_util, com_Const, com_String, com_generator, PopupComponent, DataSelector, Subset) {

/**

* Anova

*/

class Anova extends PopupComponent {

_init() {

super._init();

/** Write codes executed before rendering */

this.config.sizeLevel = 2;

this.config.checkModules = ['pd', 'np', 'stats', 'vp_confidence_interval', 'vp_sem'];

this.config.docs = 'https://docs.scipy.org/doc/scipy/reference/';

this.state = {

testType: 'one-way',

data: '',

dataType: '',

depVar: '',

factor: '',

factorA: '',

factorB: '',

covariate: '',

sigLevel: 0.05,

// Post hoc analysis option

tukeyHSD: true,

tukey: false,

scheffe: false,

duncan: false,

bonferroni: false,

// Display option

statistics: true,

boxplot: true,

equalVariance: true,

interPlot: true,

...this.state

};

this.columnBindList = ['depVar', 'factor', 'factorA', 'factorB', 'covariate'];

this.tmpInstallCode = []; // install codes

this.subsetEditor = {};

}

_bindEvent() {

super._bindEvent();

/** Implement binding events */

var that = this;

$(this.wrapSelector('#testType')).on('change', function() {

let testType = $(this).val();

that.state.testType = testType;

$(that.wrapSelector('.vp-st-option')).hide();

$(that.wrapSelector('.vp-st-option.' + testType)).show();

that.tmpInstallCode = [];

that.hideInstallButton();

if (testType === 'one-way' || testType === 'two-way') {

if (that.state.tukey || that.state.scheffe || that.state.duncan) {

// Add installation code

if (vpConfig.extensionType === 'lite') {

that.tmpInstallCode = ["%pip install scikit-posthocs"];

} else {

that.tmpInstallCode = ["!pip install scikit-posthocs"];

}

that.showInstallButton();

}

} else if (testType === 'ancova') {

// Add installation code : # pip install pingouin

if (vpConfig.extensionType === 'lite') {

that.tmpInstallCode = ["%pip install pingouin"];

} else {

that.tmpInstallCode = ["!pip install pingouin"];

}

that.showInstallButton();

}

});

$(this.wrapSelector('#data')).on('change', function() {

if (that.state.dataType === 'Series') {

// Series

that.columnBindList.forEach(id => {

$(that.wrapSelector('#' + id)).html('');

$(that.wrapSelector('#' + id)).prop('disabled', true);

});

} else {

// DataFrame

that.columnBindList.forEach(id => {

$(that.wrapSelector('#' + id)).prop('disabled', false);

});

com_generator.vp_bindColumnSource(that, 'data', that.columnBindList, 'select', false, false);

}

});

$(this.wrapSelector('.vp-st-posthoc-box .vp-state')).on('change', function() {

let id = $(this)[0].id;

let checked = $(this).prop('checked') === true;

that.state[id] = checked;

let { testType, tukey, scheffe, duncan } = that.state;

if (testType === 'one-way' || testType === 'two-way') {

if (tukey || scheffe || duncan) {

// Add installation code

if (vpConfig.extensionType === 'lite') {

that.tmpInstallCode = ["%pip install scikit-posthocs"];

} else {

that.tmpInstallCode = ["!pip install scikit-posthocs"];

}

that.showInstallButton();

} else {

that.hideInstallButton();

}

}

});

$(this.wrapSelector(''))

}

templateForBody() {

let page = $(nmHTML);

let that = this;

let dataSelector = new DataSelector({

pageThis: this, id: 'data', placeholder: 'Select data', required: true, boxClasses: 'vp-flex-gap5',

allowDataType: ['DataFrame'], withPopup: false,

finish: function(data, type) {

that.state.data = data;

that.state.dataType = type;

$(that.wrapSelector('#data')).trigger('change');

},

select: function(data, type) {

that.state.data = data;

that.state.dataType = type;

$(that.wrapSelector('#data')).trigger('change');

}

});

$(page).find('#data').replaceWith(dataSelector.toTagString());

return page;

}

render() {

super.render();

let that = this;

// render Subset

this.subsetEditor['data'] = new Subset({

pandasObject: '',

config: { name: 'Subset', category: this.name } },

{

useAsModule: true,

useInputColumns: true,

targetSelector: this.wrapSelector('#data'),

pageThis: this,

finish: function(code, state) {

that.state.data = code;

$(that.wrapSelector('#data')).val(code);

that.state.dataType = state.returnType;

$(that.wrapSelector('#data')).trigger('change');

}

});

// bind column if data exist

if (this.state.data !== '') {

com_generator.vp_bindColumnSource(this, 'data', this.columnBindList, 'select', false, false);

}

// control display option

$(this.wrapSelector('.vp-st-option')).hide();

$(this.wrapSelector('.vp-st-option.' + this.state.testType)).show();

}

generateInstallCode() {

return this.tmpInstallCode;

}

generateCode() {

let {

testType, data, depVar, factor, factorA, factorB, covariate, sigLevel,

// Post hoc analysis option

tukeyHSD, tukey, scheffe, duncan, bonferroni,

// Display option

statistics, boxplot, equalVariance, interPlot

} = this.state;

// get only text without '' or ""

let depVarText = $(this.wrapSelector('#depVar option:selected')).text();

let factorText = $(this.wrapSelector('#factor option:selected')).text();

let factorAText = $(this.wrapSelector('#factorA option:selected')).text();

let factorBText = $(this.wrapSelector('#factorB option:selected')).text();

let covariateText = $(this.wrapSelector('#covariate option:selected')).text();

let codeList = [];

let code = new com_String();

// test type label

let testTypeLabel = $(this.wrapSelector('#testType option:selected')).text();

code.appendFormatLine("# {0}", testTypeLabel);

code.appendFormat("vp_df = {0}.dropna().copy()", data);

switch (testType) {

case 'one-way':

// 1. One-way ANOVA

code.appendLine();

code.appendLine();

code.appendLine("from IPython.display import display, Markdown");

code.appendLine("_df = pd.DataFrame()");

code.appendFormatLine("for k, v in dict(list(vp_df.groupby({0})[{1}])).items():", factor, depVar);

code.appendLine(" _df_t = v.reset_index(drop=True)");

code.appendLine(" _df_t.name = k");

code.append(" _df = pd.concat([_df, _df_t], axis=1)");

// Display - Statistics

if (statistics === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Statistics");

code.appendLine("display(Markdown('### Statistics'))");

code.appendLine("display(pd.DataFrame(data={'Count':_df.count(),'Mean':_df.mean(numeric_only=True),'Std. Deviation':_df.std(numeric_only=True),'Min':_df.min(),'Max':_df.max(),");

code.appendLine(" 'Std. Error Mean':_df.apply(vp_sem),'Confidence interval':0.95,");

code.append(" 'Lower':_df.apply(vp_confidence_interval).T[0],'Upper':_df.apply(vp_confidence_interval).T[1] }))");

}

// Display - Boxplot

if (boxplot === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Boxplot");

code.appendLine("import seaborn as sns");

code.appendLine("import warnings");

code.appendLine("with warnings.catch_warnings():");

code.appendLine(" warnings.simplefilter(action='ignore', category=Warning)");

code.appendLine(" sns.boxplot(data=_df)");

code.append(" plt.show()");

}

// Display - Equal Variance

if (equalVariance === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Equal Variance test (Levene)");

code.appendLine("from scipy import stats");

code.appendLine("_lst = []");

code.appendLine("_df.apply(lambda x: _lst.append(x.dropna()))");

code.appendLine("_res = stats.levene(*_lst, center='mean')");

code.appendLine("display(Markdown('### Equal Variance test (Levene)'))");

code.append("display(pd.DataFrame(data={'Statistic':_res.statistic,'p-value':_res.pvalue}, index=['Equal Variance test (Levene)']))");

}

code.appendLine();

code.appendLine();

code.appendLine("# One-way ANOVA");

code.appendLine("import statsmodels.formula.api as smf");

code.appendLine("from statsmodels.stats.anova import anova_lm");

code.appendFormatLine("_model = smf.ols('{0} ~ C({1})', vp_df)", depVarText, factorText);

code.appendLine("_result = _model.fit()");

code.appendLine("_dfr = anova_lm(_result)");

code.appendLine("_dfr.loc['Total','df'] = _dfr['df'].sum()");

code.appendLine("_dfr.loc['Total','sum_sq'] = _dfr['sum_sq'].sum()");

code.appendLine("display(Markdown('### One-way ANOVA'))");

code.append("display(_dfr)");

// Post hoc analysis - Tukey HSD

if (tukeyHSD === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Post-hoc: Tukey HSD");

code.appendLine("from statsmodels.sandbox.stats.multicomp import MultiComparison");

code.appendFormatLine("_res = MultiComparison(vp_df[{0}], vp_df[{1}]).tukeyhsd(alpha={2})", depVar, factor, sigLevel);

code.appendLine("display(Markdown('### Post-hoc: Tukey HSD'))");

code.append("display(_res.summary())");

}

// Post hoc analysis - Bonferroni

if (bonferroni === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Post-hoc: Bonferroni");

code.appendLine("from statsmodels.sandbox.stats.multicomp import MultiComparison");

code.appendFormatLine("_res = MultiComparison(vp_df[{0}], vp_df[{1}]).allpairtest(stats.ttest_ind,alpha={2},method='bonf')", depVar, factor, sigLevel);

code.appendLine("display(Markdown('### Post-hoc: Bonferroni'))");

code.append("display(_res[0])");

}

if (tukey === true || scheffe === true || duncan === true) {

// Post hoc analysis - Tukey

if (tukey === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Post-hoc: Tukey");

code.appendLine("import scikit_posthocs as sph");

code.appendLine("display(Markdown('### Post-hoc: Tukey'))");

code.appendFormat("display(sph.posthoc_tukey(vp_df, val_col={0}, group_col={1}))", depVar, factor);

}

// Post hoc analysis - Scheffe

if (scheffe === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Post-hoc: Scheffe");

code.appendLine("import scikit_posthocs as sph");

code.appendLine("display(Markdown('### Post-hoc: Scheffe'))");

code.appendFormat("display(sph.posthoc_scheffe(vp_df, val_col={0}, group_col={1}))", depVar, factor);

}

// Post hoc analysis - duncan

if (duncan === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Post-hoc: Duncan");

code.appendLine("import scikit_posthocs as sph");

code.appendLine("display(Markdown('### Post-hoc: Duncan'))");

code.appendFormat("display(sph.posthoc_dunn(vp_df, val_col={0}, group_col={1}))", depVar, factor);

}

}

break;

case 'two-way':

// 1. Two-way ANOVA

code.appendLine();

code.appendLine();

code.appendLine("from IPython.display import display, Markdown");

code.appendLine("_df = pd.DataFrame()");

code.appendFormatLine("for k, v in dict(list(vp_df.groupby([{0},{1}])[{2}])).items():", factorB, factorA, depVar);

code.appendLine(" _df_t = v.reset_index(drop=True)");

code.appendLine(" _df_t.name = k");

code.appendLine(" _df = pd.concat([_df, _df_t], axis=1)");

code.append(" _df.columns = [[x[0] for x in _df.columns],[x[1] for x in _df.columns]]");

// Display - Statistics

if (statistics === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Statistics");

code.appendLine("display(Markdown('### Statistics'))");

code.appendLine("display(pd.DataFrame(data={'Count':_df.count(),'Mean':_df.mean(numeric_only=True),'Std. Deviation':_df.std(numeric_only=True),'Min':_df.min(),'Max':_df.max(),");

code.appendLine(" 'Std. Error Mean':_df.apply(vp_sem),'Confidence interval':0.95,");

code.append(" 'Lower':_df.apply(vp_confidence_interval).T[0],'Upper':_df.apply(vp_confidence_interval).T[1] }))");

}

// Display - Boxplot

if (boxplot === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Boxplot");

code.appendLine("import seaborn as sns");

code.appendLine("import warnings");

code.appendLine("with warnings.catch_warnings():");

code.appendLine(" warnings.simplefilter(action='ignore', category=Warning)");

code.appendLine(" sns.boxplot(data=_df)");

code.append(" plt.show()");

}

// Display - Equal Variance test

if (equalVariance === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Equal Variance test (Levene)");

code.appendLine("from scipy import stats");

code.appendLine("_lst = []");

code.appendLine("_df.apply(lambda x: _lst.append(x.dropna()))");

code.appendLine("_res = stats.levene(*_lst, center='mean')");

code.appendLine("display(Markdown('### Equal Variance test (Levene)'))");

code.append("display(pd.DataFrame(data={'Statistic':_res.statistic,'p-value':_res.pvalue}, index=['Equal Variance test (Levene)']))");

}

code.appendLine();

code.appendLine();

code.appendLine("# Two-way ANOVA");

code.appendLine("import statsmodels.formula.api as smf");

code.appendLine("from statsmodels.stats.anova import anova_lm");

code.appendFormatLine("_model = smf.ols('{0} ~ C({1}) + C({2}) + C({3}):C({4})', vp_df)", depVarText, factorAText, factorBText, factorAText, factorBText);

code.appendLine("_result = _model.fit()");

code.appendLine("_dfr = anova_lm(_result)");

code.appendLine("_dfr.loc['Total','df'] = _dfr['df'].sum()");

code.appendLine("_dfr.loc['Total','sum_sq'] = _dfr['sum_sq'].sum()");

code.appendLine("display(Markdown('### Two-way ANOVA'))");

code.append("display(_dfr)");

// Display - Interaction plot

if (interPlot === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Interaction plot");

code.appendLine("from statsmodels.graphics.factorplots import interaction_plot");

code.appendLine("import warnings");

code.appendLine("with warnings.catch_warnings():");

code.appendLine(" warnings.simplefilter(action='ignore', category=Warning)");

code.appendFormatLine(" fig = interaction_plot(x=vp_df[{0}], trace=vp_df[{1}], response=vp_df[{2}])", factorA, factorB, depVar);

code.append(" plt.show()");

}

// Post hoc analysis - Tukey HSD

if (tukeyHSD === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Post-hoc: Tukey HSD");

code.appendLine("from statsmodels.sandbox.stats.multicomp import MultiComparison");

code.appendFormatLine("_res = MultiComparison(vp_df[{0}], vp_df[{1}]).tukeyhsd(alpha={2})", depVar, factorA, sigLevel);

code.appendLine("display(Markdown('### Post-hoc: Tukey HSD'))");

code.append("display(_res.summary())");

}

// Post hoc analysis - Bonferroni

if (bonferroni === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Post-hoc: Bonferroni");

code.appendLine("from statsmodels.sandbox.stats.multicomp import MultiComparison");

code.appendFormatLine("_res = MultiComparison(vp_df[{0}], vp_df[{1}]).allpairtest(stats.ttest_ind,alpha={2},method='bonf')", depVar, factorA, sigLevel);

code.appendLine("display(Markdown('### Post-hoc: Bonferroni'))");

code.append("display(_res[0])");

}

if (tukey === true || scheffe === true || duncan === true) {

// Post hoc analysis - Tukey

if (tukey === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Post-hoc: Tukey");

code.appendLine("import scikit_posthocs as sph");

code.appendLine("display(Markdown('### Post-hoc: Tukey'))");

code.appendFormat("display(sph.posthoc_tukey(vp_df, val_col={0}, group_col={1}))", depVar, factorA);

}

// Post hoc analysis - Scheffe

if (scheffe === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Post-hoc: Scheffe");

code.appendLine("import scikit_posthocs as sph");

code.appendLine("display(Markdown('### Post-hoc: Scheffe'))");

code.appendFormat("display(sph.posthoc_scheffe(vp_df, val_col={0}, group_col={1}))", depVar, factorA);

}

// Post hoc analysis - Duncan

if (duncan === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Post-hoc: Duncan");

code.appendLine("import scikit_posthocs as sph");

code.appendLine("display(Markdown('### Post-hoc: Duncan'))");

code.appendFormat("display(sph.posthoc_dunn(vp_df, val_col={0}, group_col={1}))", depVar, factorA);

}

}

break;

case 'ancova':

// 1. ANCOVA

code.appendLine();

code.appendLine();

code.appendLine("from IPython.display import display, Markdown");

code.appendLine("_df = pd.DataFrame()");

code.appendFormatLine("for k, v in dict(list(vp_df.groupby({0})[{1}])).items():", factor, depVar);

code.appendLine(" _df_t = v.reset_index(drop=True)");

code.appendLine(" _df_t.name = k");

code.append(" _df = pd.concat([_df, _df_t], axis=1)");

// Display - Statistics

if (statistics === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Statistics");

code.appendLine("display(Markdown('### Statistics'))");

code.appendLine("display(pd.DataFrame(data={'Count':_df.count(),'Mean':_df.mean(numeric_only=True),'Std. Deviation':_df.std(numeric_only=True),'Min':_df.min(),'Max':_df.max(),");

code.appendLine(" 'Std. Error Mean':_df.apply(vp_sem),'Confidence interval':0.95,");

code.append(" 'Lower':_df.apply(vp_confidence_interval).T[0],'Upper':_df.apply(vp_confidence_interval).T[1] }))");

}

// Display - Boxplot

if (boxplot === true) {

code.appendLine();

code.appendLine();

code.appendLine("# Boxplot");

code.appendLine("import seaborn as sns");

code.appendLine("import warnings");

code.appendLine("with warnings.catch_warnings():");

code.appendLine(" warnings.simplefilter(action='ignore', category=Warning)");

code.appendLine(" sns.boxplot(data=_df)");

code.append(" plt.show()");

}

code.appendLine();

code.appendLine();

code.appendLine("# ANCOVA - Analysis of covariance");

code.appendLine("import pingouin as pg");

code.appendLine("display(Markdown('### ANCOVA - Analysis of covariance'))");

code.appendFormat("display(pg.ancova(data=vp_df, dv={0}, between={1}, covar={2}))", depVar, factor, covariate);

break;

}

codeList.push(code.toString());

return codeList;

}

}

return Anova;

});