PyMCubes

0.1.6 verified Fri May 01 auth: no python

A Python library for extracting isosurfaces from volumetric data using the marching cubes algorithm. Version 0.1.6 is the latest, with irregular releases.

pip install pymcubes

Common errors

error ModuleNotFoundError: No module named 'mcubes' ↓

cause The Python module is 'mcubes', but users try to import 'pymcubes'.

fix

Run 'pip install pymcubes' and then import using 'import mcubes'.

error ImportError: cannot import name 'marching_cubes' from 'mcubes' ↓

cause The function is named 'marching_cubes' (singular), but users often try 'marching_cubes' (plural). Check spelling.

fix

Use 'from mcubes import marching_cubes' (singular).

Warnings

gotcha The correct import is 'from mcubes import ...' not 'from pymcubes import ...'. The package installs as pymcubes but the module is mcubes. ↓

fix Use 'import mcubes' or 'from mcubes import ...' after pip install pymcubes.

gotcha Marching cubes output vertices in a (N,3) array, not (3,N). Ensure your downstream code expects the correct shape. ↓

fix Check the shape: vertices.shape should be (N,3).

Imports

marching_cubes
wrong
```
from mcubes import marching_cubes
```
correct
```
from skimage.measure import marching_cubes
```
pymcubes is a different library; the marching_cubes function is in scikit-image.
marching_cubes_func
wrong
```
from pymcubes import marching_cubes_func
```
correct
```
from mcubes import marching_cubes_func
```
The actual export is from the 'mcubes' module, not 'pymcubes'.

Quickstart

Extract an isosurface from volumetric data and save as OBJ.

import numpy as np
from mcubes import marching_cubes, export_obj

# Create volumetric data: a sphere
X, Y, Z = np.mgrid[-2:2:50j, -2:2:50j, -2:2:50j]
volume = np.sqrt(X**2 + Y**2 + Z**2) - 1.0

# Extract isosurface at level 0
vertices, triangles = marching_cubes(volume, 0)

# Export to OBJ file
export_obj(vertices, triangles, 'sphere.obj')
print('Isosurface extracted and saved to sphere.obj')