imod_coupler couples hydrological kernels. It currently focuses on groundwater and supports coupling between MetaSWAP, MODFLOW 6, and Ribasim.
It is a command-line application that can be run via:
imodc /path/to/imod_coupler.tomlFor usage information, run:
imodc --helpDeltares colleagues can find the issue tracker on Jira.
To develop imod_coupler locally:
-
Download and install pixi.
-
Clone the repository and navigate to the project root.
-
Create the development environment:
pixi install -e dev
-
Install the test dependencies:
pixi run -e dev install-test-dependencies
This command:
- Downloads the kernel dependencies (MetaSWAP, MODFLOW 6 & Ribasim) and the regression imod_coupler
- Downloads the MetaSWAP lookup table
- Generates a
.envfile in the project root with environment variables pointing to the downloaded binaries in the.pixifolder
-
Run the tests:
pixi run -e dev tests
-
Lint the codebase:
pixi run -e dev lint
Tip: When developing with Visual Studio Code, open the project via
open-vscode.batto launch a new window with the correct environment variables set.
The user acceptance tests currently use the LHM model. More models may be added in the future. These tests can only be run locally on Windows.
-
Pull the data from the Minio/DVC remote:
pixi run -e user-acceptance fetch_lhm
This unpacks the LHM model data and the MetaSWAP database required for the tests.
-
Run the user acceptance tests:
pixi run -e user-acceptance user_acceptance_test
The tests take approximately 60 minutes to complete.
Various versions of test data are tracked using DVC, which allows different data versions to exist across branches. The storage bucket is read-only for the public. To push new or updated data, contact one of the project maintainers.
When debugging unit tests in Visual Studio Code using the Test Explorer, you may encounter issues because MODFLOW 6 and MetaSWAP can behave unpredictably when initialized and finalized multiple times in the same process.
This does not occur when running (not debugging) unit tests, because a conditional check determines whether to call subprocess.run() or stay within the main thread. See the run_coupler_function fixture for details.
If you encounter errors after pulling the latest changes, your pip dependencies may be outdated. Update them by running:
pixi run update-git-dependencies