How formingworlds/proteus works
Project type and purpose
• PROTEUS is an open-source, modular Python framework for simulating coupled evolution of rocky-planet and exoplanet atmospheres and interiors.
• It’s invoked as a CLI tool (proteus), ingests user configurations (TOML), orchestrates multiple physics modules, calls external binaries (e.g., VULCAN, JANUS, PHOENIX), and produces outputs and plots.
File structure
• Top-level:
– src/proteus/cli.py & src/proteus/proteus.py: CLI entrypoints
– src/proteus/config: configuration parsing and validation
– src/proteus/[atmos_chem, atmos_clim, escape, interior, orbit, outgas, observe, star]: domain modules with wrappers to external tools
– src/proteus/grid: grid-management utilities
– src/proteus/inference: Bayesian/global optimization and analysis
– src/proteus/plot: plotting routines for coupled model outputs
– src/proteus/utils: shared helpers (I/O, logging, termination, etc.)
• External-tool wrappers: each domain subpackage has a wrapper.py to launch compiled executables.
• Tests cover each module, integration tests under tests/.
• CI, Dockerfile, docs, examples, and cluster guides show automated build, test, and deployment.
README and docs
• Describes a modular pipeline: configuration → coupling → physics modules → outputs & plots.
• Points to detailed docs on model, installation, usage, cluster deployment.
Elements
• Orange for data stores
Detailed capture of essential architecture
• Don’t simply list modules—illustrate how they connect via the Coupler and wrappers.
• Sequence of calls and data artifacts (e.g., “atmosphere.nc → interior.nc → escape.csv → plots”).