cellpose-demo-pyotritonclient-jupyterlite.py

import micropip
await micropip.install('pyotritonclient')

import io
from PIL import Image
import matplotlib.pyplot as plt
import numpy as np
from js import fetch
from pyotritonclient import get_config, execute_model

async def fetch_image(url, name=None, grayscale=False, size=None):
    response = await fetch(url)
    bytes = await response.arrayBuffer()
    bytes = bytes.to_py()
    buffer = io.BytesIO(bytes)
    buffer.name = name or url.split('?')[0].split('/')[1]
    image = Image.open(buffer)
    if grayscale:
        image = image.convert('L')
    if size:
        image = image.resize(size=size)
    image = np.array(image)
    return image

# obtain the model config
config = await get_config('https://ai.imjoy.io/triton', 'cellpose-cyto-python')

image = await fetch_image('https://static.imjoy.io/img/img02.png')
image = image.astype('float32')
diameter = {'diameter': 30}

# run inference
results = await execute_model([image.transpose(2, 0, 1), diameter], config=config, decode_bytes=True)
mask = results['mask']

# display the output
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(8, 4))
ax1.imshow(image.astype('uint8'))
ax1.set_title('input image')
ax2.imshow(mask[0])
ax2.set_title('predicted mask')
plt.show()

hpa-single-cell-classification-pyotritonclient-jupyterlite.py

import micropip
await micropip.install('pyotritonclient')

from js import fetch
import io
from PIL import Image
import matplotlib.pyplot as plt
import numpy as np
from pyotritonclient import get_config, execute_model

LABELS = {
  0: 'Nucleoplasm',
  1: 'Nuclear membrane',
  2: 'Nucleoli',
  3: 'Nucleoli fibrillar center',
  4: 'Nuclear speckles',
  5: 'Nuclear bodies',
  6: 'Endoplasmic reticulum',
  7: 'Golgi apparatus',
  8: 'Intermediate filaments',
  9: 'Actin filaments',
  10: 'Microtubules',
  11: 'Mitotic spindle',
  12: 'Centrosome',
  13: 'Plasma membrane',
  14: 'Mitochondria',
  15: 'Aggresome',
  16: 'Cytosol',
  17: 'Vesicles and punctate cytosolic patterns',
  18: 'Negative',
}

COLORS =  ["red", "green", "blue", "yellow"]

async def fetch_image(url, name=None, size=None):
    response = await fetch(url)
    bytes = await response.arrayBuffer()
    bytes = bytes.to_py()
    buffer = io.BytesIO(bytes)
    buffer.name = name or url.split('?')[0].split('/')[1]
    image = Image.open(buffer)
    if size:
        image = image.resize(size=size).convert('L')
    image = np.array(image)
    return image

async def fetch_hpa_image(image_id):
    crops = []
    for color in COLORS:
        image = await fetch_image(f'https://images.proteinatlas.org/{image_id}_{color}.jpg', size=(340, 340))
        crops.append(image)
    image = np.stack(crops, axis=0)
    # assert image.shape == (4, 128, 128)
    return image

image = await fetch_hpa_image('115/672_E2_1')
# crop the image to a single cell
image = image[:, 60:188, 120:248]

# make sure the image size is 128x128
assert image.shape == (4, 128, 128), "please make sure the image size is 128x128"
# display the image
plt.imshow(image.transpose(1,2,0)[:,:,:3])
plt.show()

config = await get_config('https://ai.imjoy.io/triton', 'bestfitting-inceptionv3-single-cell')
results = await execute_model([image.astype('float32')/255], config=config)
classes = results['classes']
pred = [(LABELS[i], prob) for i, prob in enumerate(classes.tolist()) if prob>0.5]

print(f'Prediction: {pred}')

stardist-triton-jupyterlite.py

import micropip
await micropip.install('pyotritonclient')

import io
from PIL import Image
import matplotlib.pyplot as plt
import numpy as np
from js import fetch
from pyotritonclient import get_config, execute_model

async def fetch_image(url, name=None, grayscale=False, size=None):
    response = await fetch(url)
    bytes = await response.arrayBuffer()
    bytes = bytes.to_py()
    buffer = io.BytesIO(bytes)
    buffer.name = name or url.split('?')[0].split('/')[1]
    image = Image.open(buffer).convert('L')
    if grayscale:
        image = image.convert('L')
    if size:
        image = image.resize(size=size)
    image = np.array(image)
    return image

# obtain the model config
config = await get_config('https://ai.imjoy.io/triton', 'stardist')

image = await fetch_image('https://raw.githubusercontent.com/stardist/stardist/3451a4f9e7b6dcef91b09635cc8fa78939fb0d29/stardist/data/images/img2d.tif', grayscale=True)
image = image.astype('uint16')
diameter = {'diameter': 30}

# run inference
results = await execute_model([image, diameter], config=config, decode_bytes=True)
mask = results['mask']

# display the output
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(8, 4))
ax1.imshow(image)
ax1.set_title('input image')
ax2.imshow(mask)
ax2.set_title('predicted mask')
plt.show()