#!/opt/anaconda3/bin/python3 import streamlit as st import glob import os import read_lif import numpy as np import cv2 import matplotlib.pyplot as plt import math import plotly.graph_objects as go import SessionState from plotly.offline import plot from skimage.measure import label, regionprops from skimage import data from skimage import color from skimage.morphology import extrema from skimage import exposure from scipy import spatial def bytescale(data, in_min, in_max): data = np.clip(data, in_min, in_max) data = (data - float(in_min)) / float(in_max - in_min) return np.array(data * 255, dtype=np.uint8) def pixel_list_by_radius(radius): pixel_list = [] max_pix = math.ceil(radius) for x in range(max_pix+1): for y in range(max_pix+1): if math.sqrt(x*x+y*y) < radius: pixel_list.append((x,y)) if x > 0: pixel_list.append((-x,y)) if y > 0: pixel_list.append((x,-y)) if x >0 and y > 0: pixel_list.append((-x,-y)) return(pixel_list) @st.cache def load_zero_frame(filename,sel_series,sel_channel): reader = read_lif.Reader(filename) series = reader.getSeries() image = series[sel_series].getFrame2D(T=0, channel = sel_channel, dtype=np.uint16) image_b = bytescale(image,np.min(image),np.max(image)) return image_b @st.cache def load_final_frame(filename,sel_series,sel_channel): reader = read_lif.Reader(filename) series = reader.getSeries() num_timepoints = series[sel_series].getNbFrames() image = series[sel_series].getFrame2D(T=num_timepoints-1, channel = sel_channel, dtype=np.uint16) image_b = bytescale(image,np.min(image),np.max(image)) return image_b @st.cache def load_time_series(filename,sel_series,sel_channel): reader = read_lif.Reader(filename) series = reader.getSeries() num_timepoints = series[sel_series].getNbFrames() shape = series[sel_series].get2DShape() data = np.zeros((shape[0],shape[1],num_timepoints)) for tp in range(num_timepoints): im = series[sel_series].getFrame2D(T=tp, channel=sel_channel,dtype=np.uint16) data[:,:,tp] = im return data @st.cache def get_peaks_positions(blur, percent): h_maxima = extrema.h_maxima(blur, percent) label_h_maxima = label(h_maxima) return [(x.centroid[1], x.centroid[0]) for x in regionprops(label_h_maxima)] st.sidebar.title("Input selection") filename = st.sidebar.selectbox( 'Input File', glob.glob("*.lif")) reader = read_lif.Reader(filename) series = reader.getSeriesHeaders() sel_series_i = st.sidebar.selectbox( 'Series', list(range(len(series))),format_func=lambda x:series[x].getName()) channels = series[sel_series_i].getChannels() sel_channel_i = st.sidebar.selectbox( 'Channel', list(range(len(channels))),format_func=lambda x:channels[x].getAttribute('ChannelTag')) #Load timepoint 0 of desired image image_b = load_zero_frame(filename,sel_series_i,sel_channel_i) image_b_final = load_final_frame(filename,sel_series_i,sel_channel_i) # Blur image st.sidebar.title("Peak picking parameters") blurring = st.sidebar.slider("Blurring", min_value=1, max_value=21, value=5) blur = cv2.GaussianBlur(image_b,(blurring,blurring),0) blur_final = cv2.GaussianBlur(image_b_final,(blurring,blurring),0) #st.image(blur) # Threshold image based on percentage percent = st.sidebar.slider("H-height", min_value=1., max_value=100., value=10.,step=0.1,format="%.3f") percent_final = st.sidebar.slider("Exclude based on final H-height", min_value=1., max_value=100., value=10.,step=0.1,format="%.3f") coor_list_start = get_peaks_positions(blur, percent) coor_list_final = get_peaks_positions(blur_final, percent_final) #st.write(len(coor_list_start)) #st.write(len(coor_list_final)) final_tree = spatial.KDTree(coor_list_final) coor_list = [ x for x in coor_list_start if final_tree.query(x,distance_upper_bound=3)[1] == len(coor_list_final)] #st.write(len(coor_list)) skip = st.button('skip') step1 = st.button('1 step') step2 = st.button('2 step') step3 = st.button('3 step') step4 = st.button('4 step') step5 = st.button('5 step') ss = SessionState.get(position=1, count0=0, count1=0, count2=0, count3=0, count4=0, count5=0) #ss1 = SessionState.get(count=1) widget = st.empty() #ss1 = SessionState.get(count1=1) if skip: ss.position = ss.position + 1 ss.count0 = ss.count0 + 1 if step1: ss.position = ss.position + 1 ss.count1 = ss.count1 + 1 if step2: ss.position = ss.position + 1 ss.count2 = ss.count2 + 1 if step3: ss.position = ss.position + 1 ss.count3 = ss.count3 + 1 if step4: ss.position = ss.position + 1 ss.count4 = ss.count4 + 1 if step5: ss.position = ss.position + 1 ss.count5 = ss.count5 + 1 st.write('skip=',ss.count0,'1 step=',ss.count1, '2 step=',ss.count2, '3 step=',ss.count3, '4 step=',ss.count4, '5 step=',ss.count5) max=len(coor_list) ss.position = widget.slider('Peak', 0, max, ss.position) #ss1.count1 = st.write("", ss1.count1) #selected_keypoint_i = st.slider("Select peak", min_value=0, max_value=len(coor_list), value=0,step=1) selected_keypoint_i=ss.position st.sidebar.title("Trace calculation parameters") extraction_radius = st.sidebar.slider("Extraction radius", min_value=0.0, max_value=20.0, value=3.0,step=0.1) time_series_images = load_time_series(filename,sel_series_i,sel_channel_i) coord = (int(round(coor_list[selected_keypoint_i][0])),int(round(coor_list[selected_keypoint_i][1]))) #st.write(coord) pixel_list = pixel_list_by_radius(extraction_radius) trace = [] for tp in range(time_series_images.shape[2]): summation = 0 for pix in pixel_list: ex_pix = (coord[0] + pix[0], coord[1] + pix[1]) summation += time_series_images[ex_pix[1],ex_pix[0],tp] trace.append(summation) st.line_chart(trace) min_v = np.min(time_series_images) max_v = np.max(time_series_images) plot_points = np.linspace(0,time_series_images.shape[2]-1,num = 10,dtype=int) image_to = np.zeros((30,31*10)) for i, timepoint in enumerate(plot_points): image = bytescale(time_series_images[coord[1]-15:coord[1]+15,coord[0]-15:coord[0]+15,timepoint],min_v,max_v) image_to[0:30,31*i:31 * (i+1) -1] = image #plt.imshow(image_to) #st.pyplot() plt.clf() st.write(coor_list[selected_keypoint_i]) # Show image with picked peaks show_image=st.button('SHOW IMAGE') if show_image: # Create figure fig = go.Figure() # Constants img_width = blur.shape[0] img_height = blur.shape[1] scale_factor = 1.0 # Add invisible scatter trace. # This trace is added to help the autoresize logic work. fig.add_trace( go.Scatter( x=[0, img_width * scale_factor], y=[0, img_height * scale_factor], mode="markers", marker_opacity=0 ) ) # Configure axes fig.update_xaxes( visible=False, range=[0, img_width * scale_factor] ) fig.update_yaxes( visible=False, range=[0, img_height * scale_factor], # the scaleanchor attribute ensures that the aspect ratio stays constant scaleanchor="x" ) fig.add_trace( go.Heatmap( z=blur, showscale=False, colorscale='Greys', reversescale=True ) ) # Add image #fig.update_layout( # images=[go.layout.Image( # x=0, # sizex=img_width * scale_factor, # y=img_height * scale_factor, # sizey=img_height * scale_factor, # xref="x", # yref="y", # opacity=1.0, # layer="below", # sizing="stretch", # source=blur)] #) # Configure other layout fig.update_layout( width=img_width * scale_factor, height=img_height * scale_factor, margin={"l": 0, "r": 0, "t": 0, "b": 0}, ) shapes = [ go.layout.Shape( type="rect", xref="x", yref="y", x0=point[0]-2, y0=point[1]-2, x1=point[0]+2, y1=point[1]+2, line=dict( color="yellow", width=3, ) ) for point in coor_list] shapes[selected_keypoint_i]["line"]["color"] = "red" fig.update_layout( shapes=shapes) # Disable the autosize on double click because it adds unwanted margins around the image # More detail: https://plot.ly/python/configuration-options/ st.plotly_chart(fig) #fig.show(config={'doubleClick': 'reset'})