""" ****************************************************************************************** Written by Jim Rowe (U of Sheffield) in collaboration with Alexander Jones (SLCU, Cambridge) Started: 2021-01-30 @BotanicalJim james.rowe at sheffield.ac.uk Version 2 ****************************************************************************************** """ # *******************************import libraries***************************************** from ij import IJ, ImagePlus, ImageStack, CompositeImage from fiji.util.gui import GenericDialogPlus from ij import WindowManager as WM from java.awt.event import ActionListener,AdjustmentListener, ItemListener from net.haesleinhuepf.clij2 import CLIJ2 from ij.process import ImageProcessor, ShortProcessor,StackStatistics, ImageConverter from java.awt import GridLayout, Font, Color from ij.measure import ResultsTable from ij.gui import Overlay from ij.gui import TextRoi, Roi from java.lang import Thread from datetime import datetime import os import json from weka.core import SerializationHelper, Attribute, Instances, DenseInstance from weka.classifiers.trees import RandomForest from sys import path from java.lang.System import getProperty path.append(getProperty('fiji.dir') + '\\plugins\\FRETENATOR2') import FN2_Apply_Weka_to_labels as clsfr import FN2_Label_ROI_map_to_training_data as trningdata import FN2_Train_label_classifier as trnclsfr # *******************************classes***************************************** class previewLabelerAndListeners(ActionListener, AdjustmentListener): '''Class which unique function is to handle the button clics''' def __init__(self,imp1, slider1,slider2,slider3, slider4, gd): self.height=imp2.getHeight() self.width=imp2.getWidth() self.depth=imp2.getStackSize() self.stats=StackStatistics(imp2) self.src=clij2.push(imp2) self.cal = imp2.getCalibration() self.size=int(self.stats.max)+1 #print 'size' , int(self.stats.max) self.labelValues=[1]*int(self.size) self.labelValues[0]=0 self.frames=imp2.getNFrames() self.sizedepth=imp2.getStackSize()/frames self.dimensions = [self.width, self.height, self.sizedepth] self.nucleiLabels=range(self.size) self.results=ResultsTable() self.renderPreview(1) #self.results.show("results") self.labelDict={0:[0], 1: sorted(self.nucleiLabels[1:]), 2:[], 3:[], 4:[], 5:[], 6:[], 7:[], 8:[], 9:[], 10:[]} self.identifiers=[] self.xs=[] self.ys=[] self.zs=[] self.bbz=[] self.bbze=[] self.vol=[] self.errors=[] for i in range(int(self.size)): try: self.xs.append(int(self.results.getValue("CENTROID_X",i))) self.ys.append(int(self.results.getValue("CENTROID_Y",i))) self.zs.append(int(self.results.getValue("CENTROID_Z",i))) self.bbz.append(self.results.getValue("BOUNDING_BOX_Z",i)+1) self.bbze.append(self.results.getValue("BOUNDING_BOX_END_Z",i)+1) self.identifiers.append(self.results.getValue("IDENTIFIER", i)) self.vol.append(self.results.getValue("PIXEL_COUNT", i)) except: print 'error'+str( i) self.identifiers.append(self.results.getValue("IDENTIFIER", i)) self.xs.append(-100) self.ys.append(-100) self.zs.append(-100) self.bbz.append(1) self.bbze.append(1) self.vol.append(0) self.errors.append(i) self.nucLoc= map(lambda i : self.ys[i] * width +self.xs[i], range(len(self.xs))) """labelPreviewImp - label image preview; maxZPreviewImp - maxZ label preview; maxYPreviewImp - maxY label preview""" self.slider1=slider1 self.slider2=slider2 self.slider3=slider3 self.slider4=slider4 self.gd=gd self.current=self.labelPreviewImp.getCurrentSlice() def renderPreview(self,runStats): try: self.labelPreviewImp.close() self.maxZPreviewImp.close() self.maxYPreviewImp.close() except: print "imps already closed" fp= ShortProcessor(len(self.labelValues), 1, self.labelValues, None) labelerImp= ImagePlus("labeler", fp) src2=clij2.push(labelerImp) dst=clij2.create(self.src) labelerImp.close() clij2.replaceIntensities(self.src, src2, dst) self.labelPreviewImp=clij2.pull(dst) previewDisplaySettings(self.labelPreviewImp, "label preview", 100, self.cal) try: self.labelPreviewImp.setSlice(self.current) except: pass if runStats: clij2.statisticsOfBackgroundAndLabelledPixels(dst, self.src, self.results) dst2=clij2.create(width, height, 1) clij2.maximumZProjection(dst,dst2) self.maxZPreviewImp=clij2.pull(dst2) previewDisplaySettings(self.maxZPreviewImp, "maxZ label preview", 50, self.cal) dst3=clij2.create(width, depth, 1) clij2.maximumYProjection(dst,dst3) self.maxYPreviewImp=clij2.pull(dst3) previewDisplaySettings(self.maxYPreviewImp, "maxY label preview", 50, self.cal) dst3.close() dst.close() dst2.close() src2.close() labelWindow = self.labelPreviewImp.getWindow() x=labelWindow.getLocation().x y=labelWindow.getLocation().y maxZPreviewWindow=self.maxZPreviewImp.getWindow() maxZPreviewWindow.setLocation(x, y+height+50) maxYPreviewWindow=self.maxYPreviewImp.getWindow() maxYPreviewWindow.setLocation(x+width/2, y+height+50) print labelWindow def actionPerformed(self, event): """event: actionlistener does stuff on buttonpress""" Source = event.getSource() # returns the Button object self.current=self.labelPreviewImp.getCurrentSlice() if Source.label == "Set top": self.top = self.labelPreviewImp.getCurrentSlice() self.bottom = self.slider2.getValue() self.slider1.setValue(int(self.top)) if (self.bottom < self.top) : self.bottom=self.top self.slider2.setValue(int(self.top)) return if Source.label == "Set bottom": self.top = self.slider1.getValue() self.bottom = self.labelPreviewImp.getCurrentSlice() self.slider2.setValue(int(self.bottom)) if self.bottom < self.top: self.slider1.setValue(int(self.bottom)) self.top = self.bottom print "B" return if Source.label == "Whole stack": self.top = 1 self.bottom = depth self.slider1.setValue(int(self.top)) self.slider2.setValue(int(self.bottom)) return if Source.label == "Just slice": print "current slice" self.top = self.labelPreviewImp.getCurrentSlice() self.bottom = self.labelPreviewImp.getCurrentSlice() self.slider2.setValue(int(self.top)) self.slider1.setValue(int(self.bottom)) return if Source.label == "Use pretrained classifier": labelFilePath , modelFilePath, headersFilePath=clsfr.fileSelectDialog() rt=clsfr.createResultsTable(imp2,self.src, self.dimensions) with open(headersFilePath, 'r') as config_file: columns = json.load(config_file) rtheadingsList = rt.getColumnHeadings().split() columnChoices = [] for i in range(len(columns)): if columns[i] == True: columnChoices.append(i) #print columnChoices rt2= clsfr.filteredResultsTable(rt, columnChoices) input_data, attributes= clsfr.convertTableToInstances(rt2) labeledrois = clsfr.classify(modelFilePath, input_data, attributes ) self.labelDict={0:[],1:[], 2:[], 3:[], 4:[], 5:[], 6:[], 7:[], 8:[], 9:[], 10:[], 11:[]} for i,v in enumerate(labeledrois): self.labelDict[v].append(i) for key in self.labelDict: for value in self.labelDict[int(key)]: self.labelValues[value]=int(key) print self.labelDict print self.labelValues self.src=clij2.push(imp2) self.renderPreview(0) return if Source.label == "Use existing classified image": ROIimageName= fileSelectDialog() print(ROIimageName) ROIimp= WM.getImage(ROIimageName) self.labelDict=importROIImage(ROIimp, self.src) for key in self.labelDict: for value in self.labelDict[int(key)]: self.labelValues[value]=int(key) self.renderPreview(0) return if Source.label == "Train classifier from this image": rtc2 = trningdata.createTrainingResultsTable(imp2, self.labelPreviewImp) rtc2.show('test') filePath=trnclsfr.folderSelectDialog() #rt.show("data") rtheadingsList = rtc2.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] columnChoices, columns= trnclsfr.chooseColumns(columns, rtheadingsList) print columns rtc3= trnclsfr.filteredResultsTable(rtc2, columnChoices) date= datetime.now().strftime("%Y-%m-%d-%H.%M") rtc2.save(filePath + "/" + date + "labels.csv") training_data, attributes= trnclsfr.convertTableToInstances(rtc3) classifier = RandomForest() classifier.setNumIterations(200) classifier.setNumFeatures(5) classifier.setMaxDepth(1) classifier.buildClassifier(training_data) SerializationHelper.write(filePath + "/" + date + " classifier.model", classifier) with open(filePath+"/"+date+' classifier_fileheaders.json', 'w') as f: # f.write(str(columns)) json.dump(columns, f) self.src=clij2.push(imp2) return if Source.label[:5]=="label": [int(s) for s in Source.label.split() if s.isdigit()] s=int(s) print "s " +str(s) #xy position filter pixels roi1 = self.labelPreviewImp.getRoi() selectedPixels=roi1.getContainedPoints() roiPixelLoc= map(lambda i : i.y * width +i.x, selectedPixels) pixels2=filter(lambda i: self.nucLoc[i] in roiPixelLoc, xrange(len(self.nucLoc))) #on first run set self.top and self.bottom to top and bottom of the stack try: print self.top except: self.top=1 try: print self.bottom except: self.bottom=self.labelPreviewImp.getNSlices() #on first run set self.top and self.bottom to top and bottom of the stack try: print self.maxSize except: self.maxSize=10000 try: print self.minSize except: self.minSize=0 choice= self.gd.getChoices().get(0) if choice.getSelectedItem()== "Slice": top=self.current bottom=self.current else: top=self.top bottom=self.bottom #z position filter pixels pixels2=filter(lambda i :max(self.bbze[i],bottom)-min(top,self.bbz[i]) <= (self.bbze[i]-self.bbz[i])+ (bottom-top),pixels2) pixels2=filter(lambda i :self.vol[i] <= self.maxSize, pixels2) pixels2=filter(lambda i :self.vol[i] >= self.minSize, pixels2) self.nucleiLabels=map(lambda i : i,pixels2) self.labelDict[s] =self.nucleiLabels+self.labelDict[s] self.labelDict[s] =sorted(list(set(self.labelDict[s]))) for key in self.labelDict: if key != s : self.labelDict[key]=filter(lambda x: x not in self.labelDict[s], self.labelDict[key]) #print labelDict for key in self.labelDict: for value in self.labelDict[int(key)]: self.labelValues[value]=int(key) self.labelValues[0]=0 self.labelDict[0]=[0] for key in self.labelDict.keys()[1:]: self.labelDict[key]=filter(lambda x: x != 0, self.labelDict[key]) self.renderPreview(0) #except: # print "No roi" return # Do an action depending on the button clicked def adjustmentValueChanged(self, event): """ event: an AdjustmentEvent with data on the state of the scroll bar. """ #return if the user is still dragging if event.getValueIsAdjusting(): return Source = event.getSource() try: self.slider1=self.gd.getSliders().get(0) except: "huh?" try: self.slider2=self.gd.getSliders().get(1) except: print "huh?2" self.top = self.slider1.getValue() self.bottom = self.slider2.getValue() #make top and bottom sliders autocorrect for each other if Source== self.slider1: if self.top > self.bottom: self.slider2.setValue(int(self.top)) self.bottom=self.top print "A" if Source== self.slider2: if self.bottom < self.top: self.slider1.setValue(int(self.bottom)) self.top = self.bottom print "B" try: self.slider3=self.gd.getSliders().get(2) except: print "huh?" try: self.slider4=self.gd.getSliders().get(3) except: print "huh?2" self.minSize = self.slider3.getValue() self.maxSize = self.slider4.getValue() #make top and bottom sliders autocorrect for each other if Source== self.slider3: if self.maxSize < self.minSize: self.slider4.setValue(int(self.minSize)+1) self.minSize=self.maxSize print "A" if Source== self.slider4: if self.minSize > self.maxSize: self.slider3.setValue(int(self.maxSize)-1) self.maxSize = self.minSize print "B" # *******************************functions************************************************ def concatStacks(masterStack, impToAdd): #takes an IMP and adds it to a stack, returning the concatenated stack impToAddStack=impToAdd.getImageStack() for i in xrange(1, impToAdd.getNSlices()+1): try: masterStack.addSlice(impToAddStack.getProcessor(i)) except: print "FAILED To addto stack for: "+ impToAdd.getTitle() +" " + str(i) return masterStack def errorDialog(message): gd = GenericDialogPlus("Error") gd.addMessage(message) gd.showDialog() return def previewDisplaySettings(image, title, zoom, cal): """Apply wanted settings for previews""" ImageConverter.setDoScaling(0) ImageConverter(image).convertToGray16() image.show() IJ.run("glasbey_on_dark") IJ.setMinAndMax(image, 0, 255) image.setTitle(title) image.setCalibration(cal) IJ.run("Set... ", "zoom="+str(zoom)) def extractFrame(imp, nFrame): """extract a frame from the image, returning a new 16 bit imagePlus labelled with the channel name""" stack = imp.getImageStack() fr=ImageStack(imp.width, imp.height) for i in range(1, imp.getNSlices() + 1): for nChannel in range(1, imp.getNChannels()+1): index = imp.getStackIndex(nChannel, i, nFrame) fr.addSlice(str(i), stack.getProcessor(index)) imp3 = ImagePlus("Frame " + str(nFrame), fr).duplicate() imp3.setDimensions(imp.getNChannels(), imp.getNSlices(), 1) comp = CompositeImage(imp3, CompositeImage.COMPOSITE) #comp.show() return comp def dialog(imp2, labelColorBarImp): gd = GenericDialogPlus("ROI labeller") categories=11 #placeholder slider variables so the class can be initiated slider1=0 slider2=0 slider3=0 slider4=0 test=previewLabelerAndListeners(imp2, slider1,slider2,slider3,slider4, gd) gd.addMessage("Assign selection to:") gd.addMessage(" ") for i in range(1,categories): gd.addButton("label "+str(i), test) #imp7.close() - causes an error as image needed for the dialog gd.addMessage("Automatic labelling:") gd.addMessage(" ") gd.addButton("Use pretrained classifier", test) gd.addButton("Use existing classified image", test) gd.addButton("Train classifier from this image", test) gd.addMessage(" ") gd.addMessage("Z labelling controls:") gd.addMessage(" ") gd.addButton("Set top",test) gd.addButton("Whole stack",test) gd.addButton("Set bottom", test) gd.addMessage(" ") gd.addMessage(" Label colour key:") gd.addImage(labelColorBarImp) gd.addSlider("Top",1, imp2.getStackSize(), 1) gd.addSlider("Bottom",1, imp2.getStackSize(),imp2.getStackSize()) gd.addSlider("Minimum ROI size", 0, 9999, 0, 1) gd.addSlider("Maximum ROI size", 1, 10000, 10000, 1) slider1=gd.getSliders().get(0) slider2=gd.getSliders().get(1) test.slider1=slider1 test.slider2=slider2 slider3=gd.getSliders().get(2) slider4=gd.getSliders().get(3) test.slider3=slider3 test.slider4=slider4 slider1.addAdjustmentListener(test) slider2.addAdjustmentListener(test) slider3.addAdjustmentListener(test) slider4.addAdjustmentListener(test) gd.addChoice("Apply labeling to:", ["(Sub)stack", "Slice"], "(Sub)stack") gd.setLayout(GridLayout(0,2)) gd.setModal(False) buttons=gd.getButtons() gd.showDialog() while ((not gd.wasCanceled()) and not (gd.wasOKed())): Thread.sleep(50) return test def selectionDialog(categories,labelColorBarImp): gd = GenericDialogPlus("ROI labeller -image picking") imps = WM.getImageTitles() nonimages=WM.getNonImageTitles() gd.addChoice("Image to quantify", imps, imps[0]) try: gd.addChoice("FRETENTATOR results table", nonimages, nonimages[0]) fail=0 except: gd.addMessage("No results table open") fail=1 gd.addImage(labelColorBarImp) for i in range(categories): gd.addStringField("Label "+str(i) +" name:", "Label "+str(i)) gd.addChoice("Quantify selected image or add labels to open results table?", ["Image", "Results table"], "Results table") #quantImp= IJ.getImage(gd.getNextChoice()) gd.setModal(False) gd.showDialog() while ((not gd.wasCanceled()) and not (gd.wasOKed())): Thread.sleep(50) names=dict() for i in range(categories): names[i]=str(gd.getNextString()) imageName=gd.getNextChoice() if fail==0: resultsName=gd.getNextChoice() imageOrTable=gd.getNextChoice() else: imageOrTable="Image" resultsName=0 return names, imageName, resultsName, imageOrTable def createLabelColorBar(): imp7 = ImagePlus("labelColorBar", ShortProcessor(180, 20)) ip7 = imp7.getProcessor() pix=ip7.getPixels() n_pixels = len(pix) # catch width w = imp7.getWidth() # create a ramp gradient from left to right for i in range(len(pix)): pix[i] = int((i % w)/18)+1 # adjust min and max ip7.setMinAndMax(0, 255) font = Font("SansSerif", Font.PLAIN, 12) overlay = Overlay() for i in range(len(pix)): roi = TextRoi(i*18+2, 2, str(i+1), font) roi.setStrokeColor(Color.black) overlay.add(roi) imp7.setOverlay(overlay) imp7.show() IJ.run("glasbey_on_dark") imp7=imp7.flatten() return imp7 def importROIImage(roiimp, src): rt=ResultsTable() roigtx2=clij2.push(roiimp) clij2.statisticsOfBackgroundAndLabelledPixels(roigtx2,src, rt) labeledrois = rt.getColumn('MEAN_INTENSITY') labelDict={0:[],1:[], 2:[], 3:[], 4:[], 5:[], 6:[], 7:[], 8:[], 9:[], 10:[], 11:[]} for i,v in enumerate(labeledrois): labelDict[v].append(i) return labelDict def fileSelectDialog(): """Select roi map""" imps = WM.getImageTitles() gdFS = GenericDialogPlus("Select ROI map") gdFS.addChoice("ROI map to import", imps, imps[0]) gdFS.showDialog() imageName=gdFS.getNextChoice() if gdFS.wasCanceled(): imageName= 0 return imageName # *****************************body of code starts**************************************** if __name__ == "__main__": clij2 = CLIJ2.getInstance() clij2.clear() imp1=IJ.getImage() height=imp1.getHeight() width=imp1.getWidth() depth=imp1.getStackSize() frames=imp1.getNFrames() if frames > 1: imp2 = extractFrame(imp1, 1) else: imp2=imp1 stats=StackStatistics(imp2) labelColorBarImp= createLabelColorBar() categories=11 #print dir(WM) test = dialog(imp2, labelColorBarImp) names, imageName, resultsName, imageOrTable = selectionDialog(categories,labelColorBarImp) test.maxZPreviewImp.close() test.maxYPreviewImp.close() listOfNames =["Untracked"]*65535 names[11]='Untracked' invertedDict=dict() labelDict=test.labelDict for key in labelDict.keys(): for value in labelDict[key]: if value not in test.errors: listOfNames[value]=names[key] invertedDict[value]= key else: listOfNames[value]="Untracked" invertedDict[value]= 11 print 'not in labeldict' if imageOrTable == "Results table": rt = ResultsTable.getResultsTable(resultsName).clone() else: measureImp=WM.getImage(imageName) src2=clij2.push(measureImp) rt = ResultsTable() clij2.statisticsOfBackgroundAndLabelledPixels(src2, test.src, rt) src2.close() resultsName="Results table" try: labels = rt.getColumn(rt.getColumnIndex('TrackID')) frame = rt.getColumn(rt.getColumnIndex('Frame (Time)')) except: try: labels = rt.getColumn(rt.getColumnIndex('Label')) except: labels = rt.getColumn(rt.getColumnIndex('IDENTIFIER')) for i in range(len(labels)): try: rt.setValue("Label name", i,listOfNames[int(labels[i])]) rt.setValue("Label value", i,invertedDict[int(labels[i])]) except: print i, 'eye eye' rt.show(resultsName+ " with labels") clij2.clear() labelColorBarImp.close() imp1Stats=imp1.getStatistics() print imp1Stats.max tracked=[11]*int(65535) for i,v in enumerate(test.labelValues): if value not in test.errors: tracked[i]= v else: tracked[i]= 11 print test.labelValues fp= ShortProcessor(len(tracked), 1,tracked , None) labelerImp= ImagePlus("labeler", fp) src2=clij2.push(labelerImp) conLabeledStack=ImageStack(imp1.width, imp1.height) if frames>1: for nFrame in range(1,frames+1): imp3=extractFrame(imp1, nFrame) src=clij2.push(imp3) dst=clij2.create(src) clij2.replaceIntensities(src, src2, dst) LabeledImp=clij2.pull(dst) conLabeledStack = concatStacks( conLabeledStack, LabeledImp) concatLabeledImp= ImagePlus("Labeled "+imageName, conLabeledStack) ImageConverter.setDoScaling(0) ImageConverter(concatLabeledImp).convertToGray16() IJ.setMinAndMax(concatLabeledImp, 0, 255) concatLabeledImp.setCalibration(imp1.getCalibration()) concatLabeledImp.setDimensions(1, imp1.getNSlices(), imp1.getNFrames()) concatLabeledImp = CompositeImage(concatLabeledImp, CompositeImage.COMPOSITE) concatLabeledImp.show() IJ.run("glasbey_on_dark") labelerImp.close() labelColorBarImp.close()