Skip to content

CLAUDE.md

Latest commit

 

History

History
107 lines (73 loc) · 3.19 KB

CLAUDE.md

File metadata and controls

107 lines (73 loc) · 3.19 KB

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

PotreeConverter v2.0 is a C++ application that generates octree LOD structures for streaming and real-time rendering of massive point clouds. It converts LAS/LAZ files into a format viewable in web browsers with Potree or PotreeDesktop. Version 2.0 is 10-50x faster than v1.7 and produces only 3 files instead of thousands.

Build Commands

Native Build (Linux/macOS/Windows)

mkdir build
cd build
cmake ../
make        # Linux/macOS
# Or open Visual Studio 2019 Project ./Converter/Converter.sln on Windows

Docker Build (Recommended)

# Build locally
docker build -t potreeconverter .

# Or pull pre-built image
docker pull ghcr.io/in-for-disaster-analytics/potreeconverter:latest

Usage Commands

Docker Usage

# Basic conversion
docker run -it --user $(id -u):$(id -g) \
  -v /path/to/input:/data \
  potreeconverter -i /data/input.laz -o /data/output

# With sampling strategy
docker run -it --user $(id -u):$(id -g) \
  -v /path/to/input:/data \
  potreeconverter -i /data/input.laz -o /data/output -m poisson

TACC/Apptainer Usage

module load tacc-apptainer
apptainer exec docker://ghcr.io/in-for-disaster-analytics/potreeconverter:latest \
  PotreeConverter -i /path/to/input.laz -o /path/to/output -m poisson

Native Binary Usage

PotreeConverter <input> -o <outputDir> -m <sampling_method>
# Sampling methods: poisson (default), random

Architecture

Core Components

  • Main Application: Converter/src/main.cpp - Entry point and CLI handling
  • Chunker: chunker_countsort_laszip.cpp/.h - Point cloud chunking with LAZ compression
  • Indexer: indexer.cpp/.h - Octree index generation and spatial organization
  • LasLoader: modules/LasLoader/ - LAS/LAZ file parsing and loading
  • Utilities: converter_utils.h, logger.cpp/.h, structures.h

Dependencies

  • LASlip: libs/laszip/ - LAZ compression/decompression
  • Brotli: libs/brotli/ - General compression library
  • JSON: libs/json/ - Single-header JSON library
  • TBB: Intel Threading Building Blocks (system dependency)
  • Arguments: libs/arguments/ - Command-line argument parsing

Output Structure

The converter produces 3 files:

  1. metadata.json - Point cloud metadata and structure
  2. Octree hierarchy file
  3. Point data file

Sampling Strategies

  • Poisson-disk sampling (default): Even distribution, better quality
  • Random sampling: Faster but less uniform distribution

Testing

The project includes test utilities:

  • tools/testing.mjs - Node.js testing script for Windows
  • testing/testing.mjs - Cross-platform testing utilities
  • Test data located in scripts/test_data/

Python Integration

The scripts/potree_scene_generator.py provides Python utilities for scene generation and metadata handling. Uses only Python standard library (no external dependencies).

TACC Configuration

The app.json file configures the application for TACC systems with Tapis integration. The run.sh script handles TACC execution workflow with proper input/output directory handling.