app-ml4floods

ML4Floods is an end-to-end flood delineation workflow for optical satellite data. This repository contains:

• a CWL workflow in cwl-workflow/app-ml4floods.cwl
• a Python command-line tool in command-line-tools/ml4floods-cli
• generated user documentation in docs/ml4floods.md

Documentation

Published documentation: https://eoap.github.io/app-ml4floods/

Local documentation sources:

• docs/index.md
• docs/ml4floods.md

Pull The CWL Package

You can pull the packaged CWL artifact from GHCR with oras:

oras pull ghcr.io/eoap/app-ml4floods/app-ml4floods:latest-dev

Tag mapping:

• latest-dev for the develop branch
• latest for the main branch
• x.y.z for a released tag version

Run The CLI

From the repository root:

task run

Or directly from the CLI package:

cd command-line-tools/ml4floods-cli
hatch run ml4floods-cli \
  --product-uri "https://planetarycomputer.microsoft.com/api/stac/v1/collections/sentinel-2-l2a/items/S2B_MSIL2A_20241031T105109_R051_T31SBD_20241031T133016" \
  --water-threshold 0.7 \
  --brightness-threshold 3500

Run The CWL Workflow

From the repository root:

task run-cwl

This generates a temporary CWL inputs file with the product_uri object expected by the workflow and removes it after execution.

If you pulled the CWL package from GHCR, you can run the extracted workflow with cwltool using the same product_uri object shape described in the generated documentation.

Development Tasks

task validate
task test
task lint
task check
task generate-doc

Notes

• The CLI now expects --product-uri, not --input-item.
• The direct CLI interface takes a plain URL string.
• The CWL workflow takes a structured URI input and maps its value field to --product-uri.

License

app-ml4floods is distributed under the terms of the MIT license.