A package to work with meshes, designed around use cases for analyzing neuronal morphology.
documentation https://meshparty.readthedocs.io/
From pypi:
pip install meshparty
We try to keep the pypi version up to date. To install the git version do:
git clone https://github.com/sdorkenw/MeshParty.git cd MeshParty pip install . --upgrade
to make optional features of ray tracing and interaction with the PyChunkedGraph work properly you need to install the optional dependencies
conda install pyembree pip install annotationframeworkclient
from mesh_party import trimesh_io meshmeta = trimesh_io.MeshMeta() mesh = meshmeta.mesh(path_to_mesh) # mesh gets cached local_vertices = mesh.get_local_view(n_points, pc_align=True, method="kdtree")
Meshes can be downloaded in parallel using
trimesh_io.download_meshes(seg_ids, target_dir, cv_path)
where cv_path points to the cloudvolume bucket. For downloading proofread meshes one needs to specify the mesh_endpoint of the chunkedgraph server:
trimesh_io.download_meshes(seg_ids, target_dir, cv_path, mesh_endpoint="https://...")
The mesh needs to be watertight In order to compute reliable information. To test whether a mesh is watertight, run
mesh.is_watertight
To make a mesh watertight do
mesh.fix_mesh()
Since trimesh_io.Mesh() inherits from trimesh.Trimesh all trimesh functionality is available to mesh, e.g.:
mesh.volume mesh.area mesh.center_mass