US Fire Modelling Automation

Python utilities for preparing and running the Missoula Fire Sciences Laboratory FireBehaviorModels command-line applications on Windows.

Supported applications

• FlamMap
• MTT
• TOM
• Farsite
• Randig
• FSPro

The main module is src/flame_components/flammap_cli.py. It downloads the current vendor package when needed, creates landscapes and FlamMap-family input files, and invokes vendor executables.

Requirements

• Windows
• Python 3.10 or later
• Conda is recommended
• Python packages used by the module: rasterio, requests, and psutil

Create an environment from the supplied Conda specification:

conda create --name firemodelling --file conda-spec-file-windows.txt
conda activate firemodelling

Install the repository as an editable package before using its examples:

python -m pip install --editable .

Supporting data

download_apps() downloads FireBehaviorModels.zip and extracts it to supporting_data/FB/. The directory contains vendor executables in bin/ and vendor sample data in sampledata/. It is ignored by Git; an optional prior package belongs in supporting_data/FB_old/.

For an editable checkout, supporting_data/ at the repository root is the default. For a normal installed package, the default is %LOCALAPPDATA%\flame_components\supporting_data. Set FLAME_COMPONENTS_DATA_DIR before importing flame_components to use another writable location.

Basic workflow

1. Create an LCP landscape with gen_lcp() or gen_lcp_gdal().
2. Create a model input file.
3. Run the selected executable with run_app().

FlamMap, MTT, TOM, and Farsite use gen_flammap_input_file() and gen_command_file():

from flame_components import flammap_cli as fm

input_path = fm.gen_flammap_input_file(
    out_dir='outputs',
    out_name='flammap_run',
    app_select='FlamMap',
    fuel_moisture_data=(1, '0 4 6 9 60 90'),
)

command_path = 'outputs/flammap_run_command.txt'
fm.gen_command_file(
    out_path=command_path,
    command_list=[['landscape.tif', input_path, 'outputs/flammap_run']],
)

stdout, stderr = fm.run_app('FlamMap', command_path)

Randig and FSPro

Randig and FSPro are integrated as direct-argument applications. Their executables do not consume FlamMap command files.

# Randig: landscape, Randig input file, output base
fm.run_app('Randig', [
    'landscape.tif',
    'randig.input',
    'outputs/randig_run',
])

# FSPro: landscape, FSPro input file, output base, ignition shapefile, barrier path or 0
fm.run_app('FSPro', [
    'landscape.tif',
    'fspro.input',
    'outputs/fspro_run',
    'ignition.shp',
    '0',
])

app_test('Randig') and app_test('FSPro') run the vendor sample datasets. These are long-running model executions and are not suitable for automated CI.

Use gen_randig_input_file() and gen_fspro_input_file() to create vendor-format input files. Randig and FSPro model runs remain manual workstation checks because the vendor executables can take several minutes or longer.

Tests and examples

• tests/test_flammap_cli.py contains automated coverage for downloader, executable discovery, headers, and launcher argument construction.
• examples/ contains user-facing Farsite, MTT, Randig, and FSPro drivers. They run vendor applications and should be launched manually as modules, for example python -m examples.farsite_example.
• Vendor smoke tests can take several minutes or longer and must remain outside pytest and GitHub Actions.

Run the automated test module with the interpreter configured for this project:

python -m pytest tests\test_flammap_cli.py -v

Notes

• Generated model outputs should be stored outside committed fixtures.
• Ignition and barrier shapefiles require their companion .dbf, .prj, and .shx files and must match the LCP projection.
• gen_flammap_input_file() supports FlamMap, MTT, TOM, and Farsite. Use gen_randig_input_file() or gen_fspro_input_file() for those separate vendor schemas.