Standalone Usage¶

You can run ANUGA simulations outside of the Hydrata web platform using the run_anuga Python package. This is useful for:

Local development and testing
Running on your own HPC infrastructure
Automating batch simulations
CI/CD pipelines

Installation¶

Prerequisites¶

Python 3.10+
GDAL system libraries (gdal-bin, libgdal-dev)
ANUGA (for running simulations)

Install run_anuga¶

# Clone the repository
git clone https://github.com/Hydrata/run_anuga.git
cd run_anuga

# Install in editable mode
pip install -e .

# With ANUGA support (required to actually run simulations)
pip install -e ".[anuga]"

# With development tools (pytest, ruff)
pip install -e ".[dev]"

Running a simulation¶

Using the CLI¶

run-anuga --package_dir /path/to/package/

With Hydrata server authentication (for progress reporting):

run-anuga user@example.com password --package_dir /path/to/package/

Using the Python API¶

from run_anuga.run import run_sim

run_sim("/path/to/package")

Using MPI (parallel execution)¶

For large domains, run across multiple CPU cores:

mpirun -np 4 python -m run_anuga.run --package_dir /path/to/package/

Running the example¶

The repository includes an example package:

cd run_anuga
run-anuga --package_dir examples/australian_floodplain/

Note

The example includes a 200x200m DEM clipped from 1m elevation data sourced from the Geoscience Australia ELVIS portal (CC BY 4.0). See examples/australian_floodplain/LICENSE_DATA.md for attribution.

Inspecting outputs¶

After a run completes, outputs are in the outputs_* directory:

ls outputs_1_1_1/
# run_1_1_1.sww
# run_1_1_1_depth_max.tif
# run_1_1_1_velocity_max.tif
# run_1_1_1_depthIntegratedVelocity_max.tif
# run_1_1_1_stage_max.tif
# run_anuga_1.log

Open the GeoTIFF files in QGIS or any GIS software to visualise results.

Validating a package¶

Before running, you can validate a scenario.json against the schema:

from run_anuga.schema import validate_scenario
import json

with open("scenario.json") as f:
    config = json.load(f)

validate_scenario(config)
print("Valid!")

Parsing inputs without running¶

To inspect a package without executing a simulation:

from run_anuga.run_utils import setup_input_data

input_data = setup_input_data("/path/to/package")

print(f"Project: {input_data['scenario_config']['project']}")
print(f"Duration: {input_data['scenario_config']['duration']}s")
print(f"EPSG: {input_data['scenario_config']['epsg']}")
print(f"Mesh file: {input_data['mesh_filepath']}")