Skip to content

Solutions Integration and Interoperability

Every PEOPLE-ECCO Solution is packaged as an Earth Observation Application Package (EOAP): a CWL (Common Workflow Language) workflow that runs the algorithm inside a Docker container. Because the packaging is standardized, any Solution can be executed locally with a CWL runner – outside of the PEOPLE-ECCO Platform – as long as the requirements below are met.

This page walks through running a Solution locally, using SAV (Submerged Aquatic Vegetation, from ecco-aquatic-solutions) as a worked example.

Requirements

  • Docker: to pull and run the Solution's published container image.
  • A CWL runner: e.g. cwltool/cwl-runner, installed via: 
    python3 -m venv venv
source venv/bin/activate
pip install cwlref-runner
  • CDSE credentials: Copernicus Data Space Ecosystem OIDC client ID and secret, registered against your own CDSE account. The Solution uses these to authenticate against the openEO federation and pull Copernicus data.
  • The Solution's EOAP package: downloaded from a GitHub release.

Step-by-step: running SAV locally

  1. Download and extract the EOAP package

Each Solution is published as a .tar.gz release asset on GitHub. For SAV:

curl -LO https://github.com/PEOPLE-ECCO/ecco-aquatic-solutions/releases/download/v1.0.0/dhi-sav-v1.0.0.tar.gz
mkdir dhi-sav-v1.0.0 && tar -xzf dhi-sav-v1.0.0.tar.gz -C dhi-sav-v1.0.0
cd dhi-sav-v1.0.0

The package contains the CWL workflow (run.cwl), example parameter files (sav_run_parameters_malaysia.json, sav_run_parameters_PEI.json), and an example.env template. Other Solution releases can be found under their repository's /releases page, e.g. ecco-aquatic-solutions/releases.

  1. Set your CDSE credentials as environment variables

Copy example.env to .env and fill in your own CDSE OIDC client credentials, or export them directly:

export OPENEO_AUTH_CLIENT_ID="<your CDSE client id>"
export OPENEO_AUTH_CLIENT_SECRET="<your CDSE client secret>"
  1. Choose or edit a parameters file

The package ships ready-to-use examples, e.g. sav_run_parameters_malaysia.json:

{
    "rangestart": "2023-01-01",
    "rangeend": "2023-01-23",
    "spatial_extent": {
    "west": 118.54074,
    "south": 4.31173,
    "east": 118.58351,
    "north": 4.34025,
    "crs": "EPSG:4326"
    },
    "maxcloudcover": 40,
    "area_name": "Malaysia"
}

rangestart/rangeend define the temporal extent, spatial_extent the area of interest (a bbox or GeoJSON polygon), and maxcloudcover a cloud-cover threshold. Note that the SAV model is only trained for two regions, Malaysia and PEI (Prince Edward Island) – pick sav_run_parameters_malaysia.json or sav_run_parameters_PEI.json as a starting point and keep any custom extent close to one of these two areas.

  1. Run the CWL workflow
cwl-runner run.cwl \
    --cdse_client_id=${OPENEO_AUTH_CLIENT_ID} \
    --cdse_client_secret=${OPENEO_AUTH_CLIENT_SECRET} \
    --parameters sav_run_parameters_malaysia.json \
    --run_name sav_run

cwltool pulls the published ghcr.io/people-ecco/dhi-sav:latest image, starts the container, which authenticates against CDSE, pulls the required Sentinel-2 data via openEO, and runs the SAV algorithm for the given extent and time range.

  1. Inspect the results

Once the run finishes, results are written to sav_run/output/, including:

  • collection.json – a STAC ItemCollection describing the outputs, with run metadata (solution, solutionVersion, cdseVersion, inputParameters, executionDateTime).
  • Result rasters (e.g. *_aggregated_SAV_probability.tif, *_aggregated_classification.tif) and a GeoJSON file.
  • openeo_logs.txt – logs from the underlying openEO job.

These files are self-contained and can be inspected locally (e.g. in QGIS) without any further platform involvement.

Integration with ESA APEx

The OGC Best Practice for Earth Observation Application Packages (APEx, OGC BP 20-089) standardizes CWL/Docker/STAC-based packaging for portability across processing platforms. PEOPLE-ECCO Solutions share the same building blocks: CWL workflows, Docker containers, STAC-described outputs.

APEx supports the integration of algorithms into their central Algorithm Catalogue. Full OGC APEx compatibility is a current work task of the PEOPLE-ECCO and APEx teams. Once the Solutions are availble via the Catalogue, this section will be updated.