SWAN Workflow Example

Note

This is a basic example demonstrating SWAN model execution with rompy-oceanum. Advanced features and detailed configurations will be added in future releases.

Overview

This example shows how to run a SWAN wave model using rompy-oceanum with the Oceanum Prax backend. The workflow demonstrates:

Basic SWAN model configuration
Submitting to the Prax pipeline system
Monitoring execution progress
Retrieving results

Prerequisites

Before starting this example, ensure you have:

# Install required packages
pip install rompy-oceanum oceanum

# Authenticate with Oceanum
oceanum auth login

# Set environment variables
export PRAX_ORG="your-organization"
export PRAX_PROJECT="your-project"

Basic Configuration

Create a simple SWAN model configuration:

Listing 1 config/swan_model.yml

# Basic SWAN Configuration Example
model:
  type: swan

grid:
  # Basic grid configuration
  # Details depend on your specific rompy setup

time:
  start: "2024-01-01T00:00:00"
  end: "2024-01-01T12:00:00"

outputs:
  - type: field
    variables: ["hs", "tm01"]

Execution with CLI

The simplest way to run the model is using the oceanum CLI:

# Submit model to Prax pipeline
oceanum rompy run config/swan_model.yml swan --pipeline-name my-swan-run

# Monitor progress
oceanum rompy status <run-id>

# View logs
oceanum rompy logs <run-id>

# Download results when complete
oceanum rompy sync <run-id> ./outputs

Programmatic Execution

You can also execute models programmatically:

import rompy
from rompy_oceanum.backends import PraxPipelineBackend

# Load your model configuration
config = rompy.load_config("config/swan_model.yml")

# Create model run
model_run = rompy.ModelRun(config=config)

# Execute using Prax backend
result = model_run.pipeline(
    backend="prax",
    pipeline_name="my-swan-run",
    org="your-org",
    project="your-project"
)

print(f"Submitted with run ID: {result.run_id}")

Monitoring Progress

Check the status of your model run:

# Get current status
oceanum rompy status <run-id>

# Follow logs in real-time
oceanum rompy logs <run-id> --follow

Result Management

Once your model completes, download the results:

# Download all outputs
oceanum rompy sync <run-id> ./outputs

# List available outputs before downloading
oceanum rompy list <run-id>

Next Steps

This basic example gets you started with rompy-oceanum. For more advanced usage:

Explore different model configurations in the rompy documentation
Customize resource requirements for your models
Set up automated workflows for operational forecasting