mudm-tools 0.6.3

Processing pipelines, tiling engines, and format converters for muDM spatial data.

mudm-tools

Processing pipelines, tiling engines, and format converters for muDM spatial data. Includes optional Rust acceleration via prebuilt wheels.

📖 Documentation: https://novagenresearch.github.io/mudm-tools/

Install

pip install mudm-tools

On supported platforms (Linux x86_64, macOS x86_64/arm64, Windows x86_64), this automatically includes the Rust-accelerated tiling engine. On other platforms, a pure-Python fallback is used.

import mudm_tools
print(mudm_tools.RUST_AVAILABLE)  # True with prebuilt wheel

Features

• 2D tiling: Rust-accelerated quadtree pipeline (StreamingTileGenerator2D) + legacy Python pipeline
• 3D tiling: Octree-based mesh tiling with meshopt/Draco compression (StreamingTileGenerator)
• Format converters: Xenium spatial transcriptomics, OBJ meshes, GeoJSON
• GeoParquet/Arrow: Import and export via Apache Arrow
• glTF/GLB export: 3D mesh export with optional Draco compression
• Neuroglancer: Precomputed format for browser-based 3D visualization
• Multiple output formats: PBF (MVT), 3D Tiles (GLB), tiled Parquet (ZSTD), Neuroglancer precomputed

Quick start

from mudm import MuDM  # core data model (separate package)
from mudm_tools._rs import StreamingTileGenerator2D  # Rust-accelerated
from mudm_tools.tiling2d import generate_pbf

gen = StreamingTileGenerator2D(min_zoom=0, max_zoom=7, buffer=64/4096)
gen.add_geojson(open("data.json").read(), bounds)
generate_pbf(gen, "tiles/", bounds)

License

MIT