cellpose_finetuning_dataset_creation

# import numpy as np
# import json
# import io
# from PIL import Image
# import requests
# from tqdm import tqdm

# with open('N=5.json', 'r') as f:
#     data = json.load(f)

# LB_API_KEY = 'eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJ1c2VySWQiOiJjbGhteTloeGgwMG0zMDc0Z2JuN2QwazhxIiwib3JnYW5pemF0aW9uSWQiOiJjbGhteTloeDQwMG0yMDc0ZzRwdXFhN296IiwiYXBpS2V5SWQiOiJjbGlteGFqZmgwMWRlMDcwNDl6NjczenFuIiwic2VjcmV0IjoiNTFjMmFiZGQ3NzAwOTBiNzQyMWI3NjhjZWI3YWM1MjQiLCJpYXQiOjE2ODYyMTU3MDMsImV4cCI6MjMxNzM2NzcwM30.c_Cxgs4oyc15Q8jzE3WOlB1XCjKQ5yHVCzUYY_SCQz0'
# headers = {"Authorization": f"Bearer " + LB_API_KEY}

# for i in tqdm(range(4, len(data))):
#     img = requests.get(data[i]['data_row.row_data'])
#     image = Image.open(io.BytesIO(img.content))
#     image = np.array(image) 
#     final_mask = np.zeros((1200, 1920, 4), dtype=np.uint8)

#     num_masks = len(data[i]['projects.clim82h1j07uk07y3b1dsdiuq.labels'][0]['annotations']['objects'])

#     for n in tqdm(range(num_masks)):
#         # print(data[i]['projects.clim82h1j07uk07y3b1dsdiuq.labels'][0]['annotations']['objects'][n]['mask']['url'])
#         mask = requests.get(data[i]['projects.clim82h1j07uk07y3b1dsdiuq.labels'][0]['annotations']['objects'][n]['mask']['url'], headers=headers)
#         temp_mask = Image.open(io.BytesIO(mask.content))
#         temp_mask = np.array(temp_mask)
#         ones_mask = np.where(temp_mask != 0, 1, temp_mask)
#         final_mask += (n+1) * ones_mask     # this line effectively creates instances of each individual object

#     filepath_seg = 'img_' + str(i) + '_seg.npy'
#     np.save(filepath_seg, final_mask)


import numpy as np
arr = np.load("/Users/ayyeung/Downloads/human_in_the_loop/test/img_1213_seg.npy", allow_pickle=True)
dic = arr.item()
print(dic['masks'])