Example Usage

Here is a simple example of how to use PyMS3D to compute the Morse-Smale complex of a 3D scalar field:

import pyms3d_core as pyms3d

# Identify the dataset and its dimensions
ds = "../testing/Hydrogen_128x128x128.raw" # f32 raw file in fortran order
dims = (128, 128, 128)

# Select the device to be used for processing
# NOTE: This must be called before any other pyms3d function
dev = pyms3d.GPU # Options: pyms3d.GPU (0) or pyms3d.CPU (1)
pyms3d.select_device(dev)

# Compute & save the MS complex
msc = pyms3d.compute_bin(ds, dims)
msc.save("Hydrogen_128x128x128_bin.msc") # extension .msc is recommended, but not mandatory

# Alternatively, compute from a numpy array
# import numpy as np
# data = np.fromfile(ds, dtype=np.float32).reshape(dims)
# msc = pyms3d.compute_arr(data)

# Simplify the MS complex using persistence simplification
msc.simplify_pers(0.1)
msc.save("Hydrogen_128x128x128_bin_simplified.msc")

# Load the MS complex from a file
msc2 = pyms3d.MSComplex()
msc2.load("Hydrogen_128x128x128_bin.msc")

Note that the package can take input either as a numpy array or as a raw binary file (with specified dimensions) in fortran order.

In either case, the datatype should be float32. See the dataset downloader script for an example of how datasets of other formats can be converted to f32.