pardesline

3D vision and medical imaging in one Python package

73 algorithms · 43 datasets loadable in one line · Open3D / SimpleITK / PyVista · an MCP server for AI agents

pip install "pardesline[all]"

GitHub · PyPI · Showcase · MCP server · Apache-2.0

import pardesline as pl

ds = pl.load("msd_task02_heart")                 # 30 cardiac MRIs, nothing downloaded yet
s = ds.filter(split="Tr")[0]                     # roles: mri, mask - licence attached
atrium = s.mask.to_mesh(label=1)                 # exact left-atrium surface, LPS millimetres
atrium.decimate(target_reduction=0.5).save("atrium.stl")

ct = pl.load("tcia://Pancreas-CT")[0].ct         # a DICOM series from The Cancer Imaging Archive
moved = ct.align(pl.load("other.nii.gz"))        # rigid registration, transform in moved.transform
ct.to_sitk(), atrium.to_open3d(), atrium.to_pyvista()   # the libraries you already use

Any dataset in one line

MSD, AMOS, TotalSegmentator, LUNA16, IXI, TCIA collections, OpenNeuro, Hugging Face repos, KITTI... 43 manifests, lazy access by HTTP Range, licence and citation attached. Your own folder through pl.scan.

One typed object

Point clouds, meshes, volumes, label maps that know their frame, units, labels, licence and provenance. Conversions are exact or say they are not.

73 algorithms, local or cloud

Registration, reconstruction, segmentation, SDF, LiDAR, SimpleITK volumes; validated parameters; the same names on your machine and on the hosted GPU.

For AI agents

pardesline mcp gives Claude, Gemini, Cursor or Codex the datasets, the algorithms and twenty expert skills as MCP tools. No account.