app.py

import numpy as np
import gradio as gr
import spaces
import cv2
from cellpose import models
from matplotlib.colors import hsv_to_rgb
import matplotlib.pyplot as plt
import os, io, base64
from PIL import Image 
from cellpose.io import imread, imsave
import glob 

from huggingface_hub import hf_hub_download

img = np.zeros((96, 128), dtype = np.uint8)
fp0 = Image.fromarray(img)
#fp0 = "0.png"
#imsave(fp0, img)

# data  retrieval
def download_weights():    
    return hf_hub_download(repo_id="mouseland/cellpose-sam", filename="cpsam")
    
    #os.system("wget -q https://huggingface.co/mouseland/cellpose-sam/resolve/main/cpsam")

def download_weights_old():
    import os, requests
    
    fname = ['cpsam']
    
    url = ["https://osf.io/d7c8e/download"]
    
    for j in range(len(url)):
      if not os.path.isfile(fname[j]):
        ntries = 0
        while ntries<10:
            try:
              r = requests.get(url[j])
            except:
                print("!!! Failed to download data !!!")
                ntries += 1 
                print(ntries)
            
      if r.status_code != requests.codes.ok:
        print("!!! Failed to download data !!!")
      else:
        with open(fname[j], "wb") as fid:
          fid.write(r.content)

try:
    fpath = download_weights()
    model = models.CellposeModel(gpu=True, pretrained_model = fpath)
except Exception as e:
    print(f"Error loading model: {e}")
    exit(1)



            
def plot_flows(y):
    Y = (np.clip(normalize99(y[0][0]),0,1) - 0.5) * 2
    X = (np.clip(normalize99(y[1][0]),0,1) - 0.5) * 2
    H = (np.arctan2(Y, X) + np.pi) / (2*np.pi)
    S = normalize99(y[0][0]**2 + y[1][0]**2)
    HSV = np.concatenate((H[:,:,np.newaxis], S[:,:,np.newaxis], S[:,:,np.newaxis]), axis=-1)
    HSV = np.clip(HSV, 0.0, 1.0)
    flow = (hsv_to_rgb(HSV) * 255).astype(np.uint8)
    return flow

def plot_outlines(img, masks):
    img = normalize99(img)
    img = np.clip(img, 0, 1)
    outpix = []
    contours, hierarchy = cv2.findContours(masks.astype(np.int32), mode=cv2.RETR_FLOODFILL, method=cv2.CHAIN_APPROX_SIMPLE)
    for c in range(len(contours)):
        pix = contours[c].astype(int).squeeze()
        if len(pix)>4:
            peri = cv2.arcLength(contours[c], True)
            approx = cv2.approxPolyDP(contours[c], 0.001, True)[:,0,:]
            outpix.append(approx)
    
    figsize = (6,6)
    if img.shape[0]>img.shape[1]:
        figsize = (6*img.shape[1]/img.shape[0], 6)
    else:
        figsize = (6, 6*img.shape[0]/img.shape[1])
    fig = plt.figure(figsize=figsize, facecolor='k')
    ax = fig.add_axes([0.0,0.0,1,1])
    ax.set_xlim([0,img.shape[1]])
    ax.set_ylim([0,img.shape[0]])
    ax.imshow(img[::-1], origin='upper', aspect = 'auto')
    if outpix is not None:
        for o in outpix:
            ax.plot(o[:,0], img.shape[0]-o[:,1], color=[1,0,0], lw=1)
    ax.axis('off')
    
    #bytes_image = io.BytesIO()
    #plt.savefig(bytes_image, format='png', facecolor=fig.get_facecolor(), edgecolor='none')
    #bytes_image.seek(0)
    #img_arr = np.frombuffer(bytes_image.getvalue(), dtype=np.uint8)
    #bytes_image.close()
    #img = cv2.imdecode(img_arr, 1)
    #img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
    #del bytes_image
    #fig.clf()
    #plt.close(fig)

    buf = io.BytesIO()
    fig.savefig(buf, bbox_inches='tight')
    buf.seek(0)
    pil_img = Image.open(buf)

    plt.close(fig)

    return pil_img

def plot_overlay(img, masks):
    if img.ndim>2:
        img_gray = img.astype(np.float32).mean(axis=-1)
    else:
        img_gray = img.astype(np.float32)
        
    img = normalize99(img_gray)
    #img = np.clip(img, 0, 1)
    HSV = np.zeros((img.shape[0], img.shape[1], 3), np.float32)
    HSV[:,:,2] = np.clip(img*1.5, 0, 1.0)
    for n in range(int(masks.max())):
        ipix = (masks==n+1).nonzero()
        HSV[ipix[0],ipix[1],0] = np.random.rand()
        HSV[ipix[0],ipix[1],1] = 1.0
    RGB = (hsv_to_rgb(HSV) * 255).astype(np.uint8)
    return RGB

def normalize99(img):
    X = img.copy()
    X = (X - np.percentile(X, 1)) / (1e-10 + np.percentile(X, 99) - np.percentile(X, 1))
    return X

def image_resize(img, resize=400):
    ny,nx = img.shape[:2]
    if np.array(img.shape).max() > resize:
        if ny>nx:
            nx = int(nx/ny * resize)
            ny = resize
        else:
            ny = int(ny/nx * resize)
            nx = resize
        shape = (nx,ny)
        img = cv2.resize(img, shape)
    img = img.astype(np.uint8)
    return img

    
@spaces.GPU(duration=10)
def run_model_gpu(img, max_iter, flow_threshold, cellprob_threshold):
    masks, flows, _ = model.eval(img, niter = max_iter, flow_threshold = flow_threshold, cellprob_threshold = cellprob_threshold)
    return masks, flows

@spaces.GPU(duration=60)
def run_model_gpu60(img, max_iter, flow_threshold, cellprob_threshold):
    masks, flows, _ = model.eval(img, niter = max_iter, flow_threshold = flow_threshold, cellprob_threshold = cellprob_threshold)
    return masks, flows

@spaces.GPU(duration=240)
def run_model_gpu240(img, max_iter, flow_threshold, cellprob_threshold):
    masks, flows, _ = model.eval(img, niter = max_iter, flow_threshold = flow_threshold, cellprob_threshold = cellprob_threshold)
    return masks, flows

import datetime
from zipfile import ZipFile
def cellpose_segment(filepath, resize = 1000,max_iter = 250, flow_threshold= 0.4, cellprob_threshold = 0):

    zip_path = os.path.splitext(filepath[-1])[0]+"_masks.zip"
    #zip_path = 'masks.zip'
    with ZipFile(zip_path, 'w') as myzip:
        for j in range((len(filepath))):
            now = datetime.datetime.now()
            formatted_now = now.strftime("%Y-%m-%d %H:%M:%S")            
            
            img_input = imread(filepath[j])
            #img_input = np.array(img_pil)
            img = image_resize(img_input, resize = resize)
            
            maxsize = np.max(img.shape)
            if maxsize<=1000:
                masks, flows = run_model_gpu(img, max_iter, flow_threshold, cellprob_threshold)
            elif maxsize < 5000:
                masks, flows = run_model_gpu60(img, max_iter, flow_threshold, cellprob_threshold)
            elif maxsize < 20000:
                masks, flows = run_model_gpu240(img, max_iter, flow_threshold, cellprob_threshold)
            else:
                raise ValueError("Image size must be less than 20,000")

            print(formatted_now, j, masks.max(), os.path.split(filepath[j])[-1])
            
            target_size = (img_input.shape[1], img_input.shape[0])
            if (target_size[0]!=img.shape[1] or target_size[1]!=img.shape[0]):
                # scale it back to keep the orignal size
                masks_rsz = cv2.resize(masks.astype('uint16'), target_size, interpolation=cv2.INTER_NEAREST).astype('uint16')
            else:
                masks_rsz = masks.copy()
                
            fname_masks = os.path.splitext(filepath[j])[0]+"_masks.tif"
            imsave(fname_masks, masks_rsz)
    
            myzip.write(fname_masks, arcname = os.path.split(fname_masks)[-1])
            
    
    #masks, flows, _ = model.eval(img, channels=[0,0])
    flows = flows[0]
    # masks = np.zeros(img.shape[:2])
    # flows = np.zeros_like(img)

    outpix = plot_outlines(img, masks)
    #overlay = plot_overlay(img, masks)
    
        
    
    #crand = .2 + .8 * np.random.rand(np.max(masks.flatten()).astype('int')+1,).astype('float32')
    #crand[0] = 0

    #overlay = Image.fromarray(overlay)
    flows = Image.fromarray(flows)

    Ly, Lx = img.shape[:2]
    outpix = outpix.resize((Lx, Ly), resample  = Image.BICUBIC)
    #overlay = overlay.resize((Lx, Ly), resample  = Image.BICUBIC)
    flows = flows.resize((Lx, Ly), resample  = Image.BICUBIC)

    fname_out  = os.path.splitext(filepath[-1])[0]+"_outlines.png"
    outpix.save(fname_out) #"outlines.png")
    
    #fname_flows  = os.path.splitext(filepath[-1])[0]+"_flows.png"
    #flows.save(fname_flows) #"outlines.png")

    if len(filepath)>1:
        b1 = gr.DownloadButton(visible=True, value = zip_path)
    else:
        b1 = gr.DownloadButton(visible=True, value = fname_masks)
    b2 = gr.DownloadButton(visible=True, value = fname_out) #"outlines.png")
    
    return outpix, flows, b1, b2

def download_function(): 
    b1 = gr.DownloadButton("Download masks as TIFF", visible=False)
    b2 = gr.DownloadButton("Download outline image as PNG", visible=False)
    return b1, b2

def tif_view(filepath):
    fpath, fext = os.path.splitext(filepath)
    if fext in ['tiff', 'tif']:
        img = imread(filepath[-1])
        if img.ndim==2:
            img = np.tile(img[:,:,np.newxis], [1,1,3])
        elif img.ndim==3:
            imin = np.argmin(img.shape)
            if imin<2:
                img = np.tranpose(img, [2, imin])
        else:
            raise ValueError("TIF cannot have more than three dimensions")

        Ly, Lx, nchan = img.shape
        imgi = np.zeros((Ly, Lx, 3))
        nn = np.minimum(3, img.shape[-1])
        imgi[:,:,:nn] = img[:,:,:nn]
        
        #filepath = fpath+'.png'
        imsave(filepath, imgi)
    return filepath

def norm_path(filepath):
    img = imread(filepath)
    img = normalize99(img)
    img = np.clip(img, 0, 1)
    fpath, fext = os.path.splitext(filepath)
    filepath = fpath +'.png'
    pil_image = Image.fromarray((255. * img).astype(np.uint8))
    pil_image.save(filepath)
    #imsave(filepath, pil_image)
    return filepath 
    
def update_image(filepath): 
    for f in filepath:
        f = tif_view(f)
    filepath_show = norm_path(filepath[-1])
    return filepath_show, filepath, fp0, fp0

def update_button(filepath):
    filepath = tif_view(filepath)
    filepath_show = norm_path(filepath)
    return filepath_show, [filepath], fp0, fp0
    
with gr.Blocks(title = "Hello", 
               css=".gradio-container {background:purple;}") as demo:

    #filepath = ""
    with gr.Row():
        with gr.Column(scale=2):
            gr.HTML("""<div style="font-family:'Times New Roman', 'Serif'; font-size:20pt; font-weight:bold; text-align:center; color:white;">Cellpose-SAM for cellular 
            segmentation <a style="color:#cfe7fe; font-size:14pt;" href="https://www.biorxiv.org/content/10.1101/2025.04.28.651001v1" target="_blank">[paper]</a> 
            <a style="color:white; font-size:14pt;" href="https://github.com/MouseLand/cellpose" target="_blank">[github]</a>
            <a style="color:white; font-size:14pt;" href="https://www.youtube.com/watch?v=KIdYXgQemcI" target="_blank">[talk]</a>                        
            </div>""")
            gr.HTML("""<h4 style="color:white;">You may need to login/refresh for 5 minutes of free GPU compute per day (enough to process hundreds of images). </h4>""")
            
            input_image = gr.Image(label = "Input", type = "filepath")

            with gr.Row():
                with gr.Column(scale=1):                    
                    with gr.Row():
                        resize = gr.Number(label = 'max resize', value = 1000)
                        max_iter = gr.Number(label = 'max iterations', value = 250)
                        flow_threshold = gr.Number(label = 'flow threshold', value = 0.4)
                        cellprob_threshold = gr.Number(label = 'cellprob threshold', value = 0)
                        
                    up_btn = gr.UploadButton("Multi-file upload (png, jpg, tif etc)", visible=True, file_count = "multiple")                        
                    
                    #gr.HTML("""<h4 style="color:white;"> Note2: Only the first image of a tif will display the segmentations, but you can download segmentations for all planes. </h4>""")
                    
                with gr.Column(scale=1):
                    send_btn = gr.Button("Run Cellpose-SAM")
                    down_btn = gr.DownloadButton("Download masks (TIF)", visible=False)            
                    down_btn2 = gr.DownloadButton("Download outlines (PNG)", visible=False)  
                    
        with gr.Column(scale=2):     
            outlines = gr.Image(label = "Outlines", type = "pil", format = 'png', value = fp0) #, width = "50vw", height = "20vw")
            #img_overlay = gr.Image(label = "Overlay", type = "pil", format = 'png') #, width = "50vw", height = "20vw")
            flows = gr.Image(label = "Cellpose flows", type = "pil", format = 'png', value = fp0) #, width = "50vw", height = "20vw")

            
    
    sample_list = glob.glob("samples/*.png")
    #sample_list = []
    #for j in range(23):
    #    sample_list.append("samples/img%0.2d.png"%j)
        
    gr.Examples(sample_list, fn = update_button, inputs=input_image, outputs = [input_image, up_btn, outlines, flows], examples_per_page=50, label = "Click on an example to try it")
    input_image.upload(update_button, input_image, [input_image, up_btn, outlines, flows])
    up_btn.upload(update_image, up_btn, [input_image, up_btn, outlines, flows])
    
    send_btn.click(cellpose_segment, [up_btn, resize, max_iter, flow_threshold, cellprob_threshold], [outlines, flows, down_btn, down_btn2])

    #down_btn.click(download_function, None, [down_btn, down_btn2])
        
    gr.HTML("""<h4 style="color:white;"> Notes:<br> 
                    <li>you can load and process 2D, multi-channel tifs.
                    <li>the smallest dimension of a tif --> channels
                    <li>you can upload multiple files and download a zip of the segmentations
                    <li>install Cellpose-SAM locally for full functionality.
                    </h4>""")
    
                    
demo.launch()