vixar 1.0.2

Proprietary geoscientific 3D/2D visualization engine, distributed as a Python library.

Vixar

Proprietary geoscientific 3D/2D visualization engine for Python.

Vixar renders point clouds, boreholes, block models, surfaces, and volumes in WebGL2 — inside a Jupyter notebook, a local browser, or a self-contained HTML file. No Node.js required; the engine ships pre-built inside the wheel.

import vixar as vx

viewer = vx.Viewer(theme="dark")
viewer.add_point_cloud("survey.las", color_by="elevation", cmap="spectral")
viewer.serve()           # local web server — opens http://localhost:8050
# viewer.show()          # inline Jupyter / VS Code widget
# viewer.to_html("scene.html")  # standalone HTML file

Install

pip install vixar                # core
pip install "vixar[las]"         # + LAS / LAZ point-cloud reading
pip install "vixar[all]"         # + LAS, LAZ, and mesh formats (meshio)

Features

• Full geology geometry suite — point clouds (inline + streamed), boreholes, block models, ore bodies, surfaces (OBJ / GOCAD TS), volumes, isosurfaces.
• Scales to 20+ GB via spatial tiling, octree frustum-culling, LOD, and an LRU tile cache with optional IndexedDB persistence.
• LAS / LAZ support — reads LAS 1.x–1.4 and LASzip-compressed LAZ files transparently with the vixar[las] extra.
• Interactive tools — slicer planes, split-pane cross-section, distance and angle measurement, 3-D annotations, slicer animation.
• Rust/WASM acceleration for LAS parsing and Marching Cubes (≈2–5× faster).
• Export — high-res PNG screenshot, WebM video via record_video(), standalone HTML via to_html().
• Accessible — ARIA-labelled controls, keyboard shortcuts, high-contrast mode (H key or viewer.set_high_contrast(True)), graceful WebGL2 fallback.
• Full UTM support — float64 in Python, camera-relative float32 in the GPU; no precision jitter at any scale.

Quick examples

Tiled streaming (large datasets)

from vixar.tiler import tile_point_cloud

# Pre-tile once (can be done offline)
tile_point_cloud("large_survey.las", output_dir="./tiles/", tile_size=100.0)

viewer = vx.Viewer()
viewer.add_tiled_point_cloud("./tiles/meta.json", name="Survey")
viewer.serve()

Block model + isosurface

import numpy as np

viewer = vx.Viewer()
viewer.add_block_model(
    df,                        # DataFrame with x, y, z, au_ppm columns
    x_col="x", y_col="y", z_col="z",
    value_col="au_ppm", cmap="hot",
    block_size=(10, 10, 5),
)
viewer.add_isosurface(grid, threshold=0.5, name="0.5 g/t shell")
viewer.serve()

CLI tiling

vixar tile survey.las --output ./tiles/ --tile-size 100 --color-by intensity

Keyboard shortcuts

R reset · T/F/S/I axis views · P split-pane · [/] nudge slicer · M measure · Space play/pause · H high-contrast · Ctrl+S screenshot · F3 dev overlay · Esc cancel

Documentation

Full API reference, architecture notes, and format support: omkar3344.github.io/Vixar

Development

pnpm install
pnpm build:viewer          # compiles viewer.js → python/src/vixar/static/
cd python && pip install -e ".[dev]"
pytest
pnpm test                  # JS unit tests
pnpm test:e2e              # Playwright E2E (4 browsers)