PMD Compression

Simple PMD compression, the hello world of masknmf

/home/runner/work/masknmf-toolbox/masknmf-toolbox/masknmf/utils/_cuda.py:10: UserWarning: You've explicitly selected to perform computations on the cpu, performance will be significantly slower
  warn(
[26-03-22 03:45:35]: Starting compression
[26-03-22 03:45:35]: sampled from the following regions: [0]
[26-03-22 03:45:35]: We are initializing on a total of 2000 frames
[26-03-22 03:45:35]: Loading data to estimate complete spatial basis
[26-03-22 03:45:35]: skipping the pruning step for frame cutoff
[26-03-22 03:45:35]: Finding spatiotemporal roughness thresholds

  0%|          | 0/250 [00:00<?, ?it/s]
  2%|▏         | 5/250 [00:00<00:05, 46.91it/s]
  4%|▍         | 10/250 [00:00<00:05, 47.76it/s]
  6%|▌         | 15/250 [00:00<00:04, 47.85it/s]
  8%|▊         | 20/250 [00:00<00:04, 48.06it/s]
 10%|█         | 25/250 [00:00<00:04, 48.10it/s]
 12%|█▏        | 30/250 [00:00<00:04, 47.78it/s]
 14%|█▍        | 35/250 [00:00<00:04, 48.25it/s]
 16%|█▌        | 40/250 [00:00<00:04, 48.38it/s]
 18%|█▊        | 45/250 [00:00<00:04, 48.51it/s]
 20%|██        | 50/250 [00:01<00:04, 48.42it/s]
 22%|██▏       | 55/250 [00:01<00:04, 48.35it/s]
 24%|██▍       | 60/250 [00:01<00:03, 48.28it/s]
 26%|██▌       | 65/250 [00:01<00:03, 48.34it/s]
 28%|██▊       | 71/250 [00:01<00:03, 48.82it/s]
 30%|███       | 76/250 [00:01<00:03, 48.91it/s]
 32%|███▏      | 81/250 [00:01<00:03, 48.69it/s]
 34%|███▍      | 86/250 [00:01<00:03, 48.85it/s]
 37%|███▋      | 92/250 [00:01<00:03, 49.35it/s]
 39%|███▉      | 97/250 [00:01<00:03, 49.33it/s]
 41%|████      | 102/250 [00:02<00:03, 48.86it/s]
 43%|████▎     | 107/250 [00:02<00:02, 48.76it/s]
 45%|████▍     | 112/250 [00:02<00:02, 49.04it/s]
 47%|████▋     | 117/250 [00:02<00:02, 48.77it/s]
 49%|████▉     | 122/250 [00:02<00:02, 49.13it/s]
 51%|█████     | 127/250 [00:02<00:02, 49.09it/s]
 53%|█████▎    | 132/250 [00:02<00:02, 49.00it/s]
 55%|█████▍    | 137/250 [00:02<00:02, 49.26it/s]
 57%|█████▋    | 142/250 [00:02<00:02, 49.25it/s]
 59%|█████▉    | 147/250 [00:03<00:02, 48.81it/s]
 61%|██████    | 152/250 [00:03<00:02, 48.91it/s]
 63%|██████▎   | 157/250 [00:03<00:01, 48.78it/s]
 65%|██████▌   | 163/250 [00:03<00:01, 49.46it/s]
 67%|██████▋   | 168/250 [00:03<00:01, 49.45it/s]
 69%|██████▉   | 173/250 [00:03<00:01, 49.54it/s]
 71%|███████   | 178/250 [00:03<00:01, 49.42it/s]
 73%|███████▎  | 183/250 [00:03<00:01, 49.08it/s]
 75%|███████▌  | 188/250 [00:03<00:01, 49.10it/s]
 77%|███████▋  | 193/250 [00:03<00:01, 48.80it/s]
 79%|███████▉  | 198/250 [00:04<00:01, 48.46it/s]
 81%|████████  | 203/250 [00:04<00:00, 48.47it/s]
 83%|████████▎ | 208/250 [00:04<00:00, 48.30it/s]
 85%|████████▌ | 213/250 [00:04<00:00, 48.19it/s]
 87%|████████▋ | 218/250 [00:04<00:00, 48.44it/s]
 89%|████████▉ | 223/250 [00:04<00:00, 48.63it/s]
 91%|█████████ | 228/250 [00:04<00:00, 48.47it/s]
 93%|█████████▎| 233/250 [00:04<00:00, 48.68it/s]
 96%|█████████▌| 239/250 [00:04<00:00, 49.23it/s]
 98%|█████████▊| 244/250 [00:05<00:00, 49.17it/s]
100%|█████████▉| 249/250 [00:05<00:00, 48.87it/s]
100%|██████████| 250/250 [00:05<00:00, 48.76it/s]
[26-03-22 03:45:40]: Running Blockwise Decompositions

  0%|          | 0/3 [00:00<?, ?it/s]
 67%|██████▋   | 2/3 [00:00<00:00, 12.71it/s]
100%|██████████| 3/3 [00:00<00:00, 13.15it/s]
[26-03-22 03:45:40]: Constructed U matrix. Rank of U is 182
[26-03-22 03:45:40]: PMD Objected constructed
/opt/hostedtoolcache/Python/3.12.13/x64/lib/python3.12/site-packages/fastplotlib/graphics/features/_base.py:18: UserWarning: casting float64 array to float32
  warn(f"casting {array.dtype} array to float32")

# test_example = true

import masknmf
import torch
import fastplotlib as fpl
from urllib.request import urlretrieve
import tifffile


urlretrieve(
    "https://github.com/flatironinstitute/CaImAn/raw/refs/heads/main/example_movies/demoMovie.tif",
    "./demo.tif"
)

# always lazy load raw data by memmaping or other methods
data = tifffile.imread("./demo.tif")

block_sizes = [32, 32]
max_components = 20

# it's recommended to use masknmf on a machine with a GPU
# device = "cuda" if torch.cuda.is_available() else "cpu"
device = "cpu"
# number of frames used to estimate the spatial basis in PMD
num_frames_for_spatial_fit = data.shape[0]

# perform PMD
pmd_result = masknmf.compression.pmd_decomposition(
    data,
    block_sizes,
    num_frames_for_spatial_fit,
    max_components=max_components,
    device=device,
    frame_batch_size=1024
)

# get the residual
pmd_residual = masknmf.PMDResidualArray(data, pmd_result)

# view the movies, note that all these array are LAZY evaluated, allowing you to view extremely large datasets!
iw = fpl.ImageWidget(
    data=[data, pmd_result, pmd_residual],
    names=["raw", "pmd", "residual"],
    figure_kwargs={"size": (1000, 340), "shape": (1, 3)},
    cmap="gnuplot2",
)

iw.show()

# use the time slider or set the frame index programmatically
iw.current_index = {"t": 1610}

# manually set vmin-vmax to emphasize noise in raw video
# you can also adjust the vmin-vmax using the histogram tool
# reset the vmin-vmax by clicking the buttons under "ImageWidget Controls"
for image in iw.managed_graphics:
    image.vmax = 3_200

# remove toolbar to reduce clutter
for subplot in iw.figure:
    subplot.toolbar = False


# ignore the remaining lines these are just for docs generation
figure = iw.figure
if __name__ == "__main__":
    print(__doc__)
    fpl.loop.run()