taskmaestro-resinsight

taskmaestro tasks and example workflows for ResInsight.

Installation

pip install taskmaestro-resinsight

rips must match the ResInsight build it talks to. To use the API from a local ResInsight checkout instead of the released one on PyPI:

pip install -e /path/to/resinsight/GrpcInterface/Python

Development setup

git clone https://github.com/OPM/taskmaestro-resinsight ~/src/taskmaestro-resinsight
cd ~/src/taskmaestro-resinsight

# install this repo and its test/lint tools (including Ruff 0.16+)
pip install -e ".[dev]"

Discovering tasks and workflows

Installing the package registers its tasks and completions workflow through the taskmaestro.tasks and taskmaestro.workflows entry-point groups. No source-tree scan, PYTHONPATH change, or workflow-directory symlink is needed:

from taskmaestro import (
    get_registered_workflow,
    registered_task_names,
    registered_workflow_names,
)

print(
    sorted(name for name in registered_task_names() if name.startswith("resinsight."))
)
# ['resinsight.add_perforation', 'resinsight.connect', ...]

print(registered_workflow_names())
# {'resinsight.completions'}

workflow = get_registered_workflow("resinsight.completions")
print(workflow.to_mermaid())

The example in workflows/resinsight_completions/workflow.yaml also uses entry-point identifiers such as resinsight.connect in its task: and depends_on: fields. It can be loaded normally after installation:

from taskmaestro import load_workflow_from_yaml

loaded = load_workflow_from_yaml(
    "workflows/resinsight_completions/workflow.yaml",
    "workflows/resinsight_completions/input.yaml",
)

Applications can use registered_workflows() as their workflow catalogue and avoid the old convention of scanning ~/.taskmaestro/workflows.

Running tests

RESINSIGHT_EXECUTABLE=/path/to/ResInsight pytest tests/ -v