OGC Application Package Examples

A collection of example OGC Application Packages for the MAAP. Each example is a small, self-contained processing unit bundled with everything needed to register and run it as an OGC process.

What is an OGC Application Package?

An OGC Application Package is a standardized way to describe a data-processing application so it can be deployed and executed portably across platforms. The application's logic is containerized, and a Common Workflow Language (CWL) document declares its inputs, outputs, container image, and resource requirements. A platform (like MAAP) can then ingest that description and run the process on demand, wiring the declared inputs to command-line arguments and collecting the declared outputs.

The ogc-app-pack-generator GitHub Action

Pushing a change under an example directory triggers a per-example GitHub workflow (e.g. .github/workflows/stac-clip.yml), which calls the shared .github/workflows/generate-app-pack.yml. That reusable workflow runs the MAAP-Project/ogc-app-pack-generator action, which:

1. Reads the example's algorithm_config.yml and Containerfile.
2. Builds and publishes the container image to the repository's GHCR.
3. Generates the CWL workflow, writes it to cwl_workflows/ by default and commits it back to the working branch.
4. Registers the process with the MAAP OGC processes endpoint.

Examples

Example	What it does	Runtime	Highlights
Write String to File	Writes an input string to a text file.	Python script	Simplest example: string in, file out
Estimate Pi	Estimates π by numerically integrating 4/(1+x²) over [0, 1] (midpoint rule).	Compiled Fortran	A non-Python, compiled-binary process
Color to Greyscale	Converts an input image to greyscale using GDAL.	Python notebook (papermill)	File input; Jupyter notebook
STAC Raster Subset	Clips a raster asset from a staged STAC Catalog to a bbox and emits a new STAC Catalog.	Python notebook (papermill)	STAC I/O; Jupyter notebook

Running the CWL workflows locally

The generated CWL workflows can be run on your own machine with the reference CWL runner, cwltool, plus a container engine.

Prerequisites

• A container engine: Docker or Podman.

• cwltool:

  pip install cwltool
  # or: conda install -c conda-forge cwltool

Run an example

Each example ships a sample job file (input.yml). From the repository root, run the workflow with that job file:

cwltool color_to_greyscale/cwl_workflows/process_color-to-greyscale_main.cwl \
        color_to_greyscale/input.yml

The process writes its results into a Directory output (globbed as ./output*), which cwltool stages into your current working directory when the run completes.

Other examples follow the same pattern — point cwltool at the example's cwl_workflows/process_<name>_main.cwl and its input.yml:

cwltool write_string_to_file/cwl_workflows/process_write-string-to-file_main.cwl write_string_to_file/input.yml
cwltool estimate_pi/cwl_workflows/process_estimate-pi_main.cwl estimate_pi/input.yml
cwltool stac_clip/cwl_workflows/process_stac-clip_main.cwl stac_clip/input.yml

Examples that need extra setup to run locally

stac_clip follows the OGC stage-in / stage-out convention: on the platform, MAAP stages the input STAC Item into a local catalog directory before the container runs. To run this example locally, emulate this behavior by creating a directory named input and download the STAC item to this directory.

The following examples run locally with no extra setup: write_string_to_file, estimate_pi, color_to_greyscale.