""" ****************************************************************************************** Written by Jim Rowe (U of Sheffield) Started: 2026-02-19 @BotanicalJim james.rowe at sheffield.ac.uk ****************************************************************************************** """ from ij import IJ from ij.measure import ResultsTable from fiji.util.gui import GenericDialogPlus from java.awt import GridLayout, Font, Color from ij import WindowManager as WM from datetime import datetime import os import json from os.path import exists from jarray import array from java.util import ArrayList, Random from weka.core import SerializationHelper, Attribute, Instances, DenseInstance from weka.classifiers.trees import RandomForest def errorDialog(message): """Outputs a given error for end users""" gd = GenericDialogPlus("Error") gd.addMessage(message) gd.showDialog() return def folderSelectDialog(): """Select training file or folder of training files""" imps = WM.getImageTitles() gd = GenericDialogPlus("Select training file or folder of training files") gd.addDirectoryOrFileField("Choose path", "") gd.addMessage('Choose parameters for random forest classifier') gd.addSlider("Number of trees", 50, 1000, 200, 10) gd.addSlider("Number of features", 1, 20, 5, 1) gd.addSlider("Depth of trees", 1, 20, 2, 1) gd.showDialog() if gd.wasCanceled(): IJ.exit() sliders=gd.getSliders() trees = sliders.get(0).getValue()*10 features = sliders.get(1).getValue() treeDepth = sliders.get(2).getValue() print trees, features, treeDepth return gd.getNextString(),trees, features, treeDepth def chooseColumns(columns, headingList): """Select which parameters to train the model with""" gd = GenericDialogPlus("Choose parameters for training") for i in range(len(headingList)): gd.addCheckbox(headingList[i],columns[i]) gd.addCheckbox("Save settings as default", 0) gd.setLayout(GridLayout(0,3)) gd.showDialog() checkBoxes=gd.getCheckboxes() choices=[] for i in range(len(checkBoxes)-1): boolchoice=gd.getNextBoolean() if boolchoice == True: choices.append(i) columns[i]=boolchoice saveSettings= gd.getNextBoolean() if saveSettings== True: with open('FN2_training_defaults.json', 'w') as f: json.dump(columns, f) if gd.wasCanceled(): IJ.exit() return choices, columns def filteredResultsTable(rtc, columnChoices): "Returns a resultsTable of just the columns to be trained" rtc2=ResultsTable() for j in range(rtc.size()): rtc2.addRow() for i in columnChoices: rtc2.addValue(rtc.getColumnHeading(i), rtc.getColumn(i)[j]) return rtc2 def concatResultsTable(rt1, rtc): "Concatenates results tables" rtcheadingList=rtc.getColumnHeadings().split() #morphResults.show("huh") rt1headingsList = rt1.getColumnHeadings().split() #combine both data tables and display the training data file for j in range(rtc.size()): rt1.addRow() for i in range(len(rtcheadingList)): rt1.addValue(rtcheadingList[i], rtc.getColumn(i)[j]) #print(j) return rt1 def convertTableToInstances(rt2): "Converts a results table to instances to be passes to the classifier" rt2headingsList = rt2.getColumnHeadings().split() n_attributes = len(rt2headingsList) classList= [str(i) for i in range(12)] attributes = ArrayList([Attribute(i) for i in rt2headingsList[:-1]]) attributes.add(Attribute("class", classList)) samples=[] sample=[] for i in range(rt2.size()): sample=[] if rt2.getValue(rt2headingsList[-1],i) <10 : for j in rt2headingsList: sample.append(rt2.getValue(j, i)) samples.append(array(sample,'d')) #print sample #print len(attributes) #print n_attributes input_data = Instances("training", attributes, len(samples)) input_data.setClassIndex(len(attributes) -1) # the last one is the class for vector in samples: input_data.add(DenseInstance(1.0, vector)) return input_data, attributes try: from net.haesleinhuepf.clijx import CLIJx #from net.haesleinhuepf.clij2 import CLIJx except: errorDialog("""This plugin requires clij2 to function. To install please follow these instructions: 1. Click Help>Update> Manage update sites 2. Make sure the "clij" and "clij2" update sites are selected. 3. Click Close> Apply changes. 4. Close and reopen ImageJ""") clij2 = CLIJx.getInstance() # *****************************body of code starts**************************************** if __name__ == "__main__": clij2.clear() filePath, trees, features, treeDepth =folderSelectDialog() rt=ResultsTable() if os.path.isfile(filePath): rt1=ResultsTable.open(filePath) else: fileList=os.listdir(filePath) for i in fileList: if i[-4:]==".csv": rt1=ResultsTable.open(filePath+ "/" +i) concatResultsTable(rt, rt1) rtheadingsList = rt.getColumnHeadings().split() columns = [False, False, False, False, False, False, False, True, True, True, False, False, False, False, False, True, False, False, False, False, False, False, False, False, False, False, False, True, True, True, True, False, False, False, False, True, True, True, False, False, False, False, True, True, True, False, False, False, False, False, False, False, False, False, True, False, False, False, False, False, False, True, True, True, True, True, True, True, True, True, True, True, True, True, True, True, True, True] if exists('FN2_training_defaults.json'): print('exists') sf=file('FN2_training_defaults.json', 'rb') columns=json.load(sf) sf.close() #columns = [1]*len(rtheadingsList) columnChoices, columns= chooseColumns(columns, rtheadingsList) print columns rt2= filteredResultsTable(rt, columnChoices) #rt2.show('filtered') training_data, attributes= convertTableToInstances(rt2) print len(training_data) date= datetime.now().strftime("%Y-%m-%d-%H.%M") #clij2.trainWekaFromTable(rt2, 'LABEL_ID', filePath + "/" + date + " classifier.model", 500, 10,10) classifier = RandomForest() classifier.setCalcOutOfBag(True) classifier.setComputeAttributeImportance(True) classifier.setNumIterations(trees) classifier.setNumFeatures(features) classifier.setMaxDepth(treeDepth) classifier.buildClassifier(training_data) SerializationHelper.write(filePath + "/" + date + " classifier.model", classifier) with open(filePath+"/"+date+' classifier_fileheaders.json', 'w') as f: json.dump(columns, f) oob_eval = classifier.getOutOfBagEvaluationObject() #print('Out of bag error:' + str(oob_eval)) IJ.log("\n=== Out of bag error summary ===") IJ.log(oob_eval.toSummaryString()) IJ.log(oob_eval.toMatrixString()) importances=classifier.computeAverageImpurityDecreasePerAttribute(None) chosenHeadings = [rtheadingsList[item] for item in columnChoices] importancesdict=dict(zip(chosenHeadings, importances)) sortimportances=sorted(importancesdict.items(), key=lambda item: item[1], reverse=True) IJ.log("\n=== Parameter/attribute importance in model ===") for i in sortimportances: IJ.log(str(i)[2:-1]) oob_error = 100*classifier.measureOutOfBagError() # Returns error rate as a decimal (e.g., 0.045 = 4.5%) gd2 = GenericDialogPlus("Model results") gd2.addMessage('Error rate is '+ str(oob_error)+ """ %. If this error is too high, try increasing the number of trees (e.g. 500), the depth of trees (e.g. 10) and the Depth of Trees (e.g. 5). You can also try altering which parameters/ attributes the training considers. Check the log t0 see which attributes are most important and a detailed analysis of model error""") gd2.showDialog()