pangolin-fast 0.1.0

Minimal pybind11 bindings for Pangolin on Windows (users install Pangolin deps themselves).

pangolin_fast

pangolin_fast is a lightweight Python wrapper around Pangolin for fast, simple realtime 3D visualization.

It is designed as a pragmatic alternative to heavier viewers when you want:

• High-FPS rendering with a dedicated C++ render thread
• Minimal Python API for points, lines, meshes, poses, and UI controls
• Low-overhead updates for dynamic scenes (e.g. SLAM, tracking, robotics, point-cloud streams)

---

What This Repo Contains

• pangolin_fast/ - Python package entrypoint
• src/ - C++ viewer implementation + pybind11 bindings
• demo.py - runnable usage example
• INSTALL_PANGOLIN_WINDOWS.md - Pangolin/vcpkg/Windows setup guide

If you need help installing Pangolin itself on Windows, follow:

• INSTALL_PANGOLIN_WINDOWS.md

---

Quick Start (Library)

# From this repo root:
python -m pip install -e .
python demo.py

Public PyPI package:

• https://pypi.org/project/pangolin-windows-easy/0.1.0/

pip install pangolin-windows-easy==0.1.0

Note: the package still requires Pangolin and its Windows toolchain/runtime dependencies to be installed and configured first. Follow:

• INSTALL_PANGOLIN_WINDOWS.md

If your Pangolin dependencies are in vcpkg and not on PATH, set:

$env:VCPKG_ROOT = "$env:USERPROFILE\vcpkg"

pangolin_fast will try to load DLLs from:

• VCPKG_ROOT\installed\x64-windows\bin
• VCPKG_INSTALLATION_ROOT\installed\x64-windows\bin
• %USERPROFILE%\vcpkg\installed\x64-windows\bin

---

Minimal Usage

import time
import numpy as np
from pangolin_fast import Viewer

v = Viewer(width=1280, height=720, title="pangolin_fast")
v.add_bool("show_grid", True)
v.add_float("point_size", 2.0, 1.0, 8.0)
v.add_grid("grid", half=10.0, step=1.0)

xyz = np.random.uniform(-2, 2, size=(50000, 3)).astype(np.float32)
rgb = np.clip((xyz + 2.0) / 4.0, 0.0, 1.0)
v.add_points("cloud", xyz, rgb=rgb, dynamic=True)

v.start()
try:
    while v.is_running():
        ui = v.get_ui()
        v.set_visible("grid", bool(ui.get("show_grid", True)))
        v.set_point_size("cloud", float(ui.get("point_size", 2.0)))
        v.update_points("cloud", xyz=xyz)
        time.sleep(1.0 / 30.0)
finally:
    v.stop()

---

API Overview

Viewer methods currently exposed:

• Lifecycle: start, stop, is_running
• Camera: set_camera, set_projection
• Geometry:
  • add_points, update_points, set_point_size
  • add_lines
  • add_axis, add_grid
  • add_mesh
  • set_pose
  • set_visible, remove
• UI vars: add_bool, add_float, get_ui

Input conventions:

• Coordinates are NumPy arrays of shape (N, 3) and float32-compatible.
• Colors accept float [0,1] or uint8 [0,255], shape (N, 3).
• Mesh faces are integer arrays of shape (M, 3).
• Poses are 4x4 transform matrices.

---

Current Scope

• Target platform: Windows (native, Pangolin + vcpkg workflow)
• Key focus: fast visualization and simple control loops from Python
• on_key is currently a stub in this build

---

Installation Notes

This README is intentionally focused on using the Python library.

For Pangolin installation/toolchain setup, see:

• INSTALL_PANGOLIN_WINDOWS.md