neuroglancer.LocalVolume(
    data,
    dimensions=None,
    volume_type=None,
    voxel_offset=None,
    encoding="npz",
    max_voxels_per_chunk_log2=None,
    mesh_options=None,
    downsampling="3d",
    chunk_layout=None,
    max_downsampling=64,
    max_downsampled_size=128,
    max_downsampling_scales=inf,)

Initializes a LocalVolume.

@param data: Source data.

@param downsampling: ‘3d’ to use isotropic downsampling, ‘2d’ to

downsample separately in XY, XZ, and YZ, None to use no downsampling.

@param max_downsampling: Maximum amount by which on-the-fly downsampling

may reduce the volume of a chunk. For example, 4x4x4 downsampling reduces the volume by 64.

@param volume_type: either ‘image’ or ‘segmentation’. If not specified,

guessed from the data type.

@param voxel_size: Sequence [x, y, z] of floats. Specifies the voxel

size.

@param mesh_options: A dict with the following keys specifying options

for mesh simplification for ‘segmentation’ volumes:

  • max_quadrics_error: float. Edge collapses with a larger associated quadrics error than this amount are prohibited. Set this to a negative number to disable mesh simplification, and just use the original mesh produced by the marching cubes algorithm. Defaults to 1e6. The effect of this value depends on the voxel_size.

  • max_normal_angle_deviation: float. Edge collapses that change a triangle normal by more than this angle are prohibited. The angle is specified in degrees. Defaults to 90.

  • lock_boundary_vertices: bool. Retain all vertices along mesh surface boundaries, which can only occur at the boundary of the volume. Defaults to true.