Quick Start

This example generates a zero-temperature nucleonic equation of state (EOS) from a set of bulk saturation parameters.

Edit input/config.yaml to choose the model regime, the bulk properties (or a named parametrization) and the output grid:

model_regime: non-relativistic     # or "relativistic"
bulk_properties:
  parametrization: ''              # or SLy4, SLy4_bar, RATP, SKI4, NL3
  BE:    -15.97                    # MeV
  n0:     0.1595                   # fm^-3
  K0:     230.0                    # MeV
  Q0:    -225.0                    # MeV
  Z0:    -443.0                    # MeV
  Esym:   32.01                    # MeV
  Lsym:   46.0                     # MeV
  Ksym:  -120.0                    # MeV
  Qsym:   350.0                    # MeV
  Zsym:  -690.0                    # MeV
  meff_norm:      0.89
  meff_splitting: 0.0              # MeV
  b0:     6.9
  bsym:   0.0
grid:
  nB_grid: linear                  # or "log"
  nB_min:  0.05                    # fm^-3
  nB_max:  0.8                     # fm^-3
  nB_points: 25
  delta_grid: linear               # or "log"
  delta_min: 0.0
  delta_max: 1.0
  delta_points: 10

To use a tabulated parametrization instead of the explicit values, set bulk_properties.parametrization to one of SLy4, SLy4_bar, RATP, SKI4 or NL3; the remaining bulk_properties fields are then ignored. See Configuration File for the full field reference.

Run from src/:
```
cd src
python3 validate_config.py && python3 interface.py
```
validate_config.py checks input/config.yaml against the OpenAPI schema and rewrites it with any missing fields filled in from their defaults. interface.py then computes the EOS.
Results are written to the output/ directory:
- output/eos.csv — the tabulated EOS, one row per grid point and no header row. See Output for the column layout.
- output/status.yaml — an integer status code and a message (200 on success, 400 for a bad config, 500 for a runtime error).

Note

Running locally requires the libnseos.so shared library. Build it once with bash install.sh from the repository root (needs gfortran and make) before the first run. See Installation for details.

Running with Docker

The bundled run_docker.sh builds the image and runs the same two-step pipeline, mounting the local input/ and output/ directories:

./run_docker.sh

Equivalently:

docker build -t metamodel:latest .
docker run --rm -it \
    -v "$(pwd)/input:/opt/input" \
    -v "$(pwd)/output:/opt/output" \
    metamodel:latest \
    bash -c "python3 validate_config.py && python3 interface.py"

Config Validation

To validate (and normalise) a config without running the model, run from src/:

python3 validate_config.py --config_file_path ../input/

Missing fields are filled with their default values and the file is written back in place. On failure the script writes output/status.yaml with code 400 and a message naming the offending field.