Harmonic Measure Numerics

Thin glue layer for running numerical experiments related to the harmonic measure paper.

Overview

This package provides:

• Configuration-driven experiments: Run computations from YAML/JSON config files
• Parallel parameter sweeps: Multi-process scaling studies
• Result aggregation: JSON output for further analysis

The heavy lifting is done by the underlying packages:

• harmonic-measure - Core harmonic measure computation
• panel-bie - Boundary integral equation solver
• superellipse - Lamé curve geometry

Installation

pip install harmonic-measure-numerics

# Or install from source
pip install -e .

Usage

Single experiment

# Run convergence test
hm-run configs/paper_default.yaml -o results/

# Run scaling sweep
hm-run configs/scaling_sweep.yaml -o results/

Parallel parameter sweep

# Sweep over multiple n values
hm-sweep --ns 8 12 16 24 32 48 64 --workers 4 -o sweep.json

Configuration format

# convergence experiment
type: convergence
n: 64
a: 1.0
c: 2.0
quality: mid  # fast, mid, high
tol: 1.0e-10

# scaling experiment
type: scaling
ns: [8, 12, 16, 24, 32, 48, 64]
a: 1.0
c: 2.0

Project Structure

.
├── configs/           # Experiment configurations
│   ├── paper_default.yaml
│   └── scaling_sweep.yaml
├── src/hm_numerics/
│   ├── runner.py     # Config-driven experiment runner
│   └── sweep.py      # Parallel parameter sweep
└── results/          # Output directory (gitignored)

Related Repositories

• harmonic-measure-paper - Paper LaTeX source
• harmonic-measure - Core computation package

License

MIT License