Glacium

Glacium is a lightweight command line tool to manage small simulation workflows. Projects are created inside the runs/ directory of the current working directory and consist of a global configuration, a set of jobs and rendered templates. The focus lies on easily defining new recipes and executing jobs in dependency order. Programmatic control is available through a high level pipeline API; see docs/high_level_api/intro.rst for an overview.

Installation

Install the package with pip (Python 3.12 or newer is required):

pip install .

Warning: make sure the old pyfpdf package is not installed alongside fpdf2. The two libraries conflict and can lead to runtime errors. If you see a warning about PyFPDF, run:

pip uninstall --yes pyfpdf

This exposes a glacium command via the console script entry point.

The DejaVuSans.ttf font used for PDF reports ships with the package.

Usage

Below is a quick tour through the most important CLI commands. Each command provides --help for additional options.

Create a project

# create a new project from the default recipe
glacium new MyWing

The multishot recipe runs ten solver cycles by default. Override the count with --multishots:

glacium new MyWing --multishots 5

The command prints the generated project UID. All projects live below ./runs/<UID> in the current working directory. glacium new and glacium init parse case.yaml and write global_config.yaml automatically. When running multishot jobs the template files for each shot are generated automatically. After editing case.yaml you can run glacium update to regenerate the configuration. Set CASE_MULTISHOT in case.yaml to a list of icing times for each shot.

Case sweep

glacium case-sweep --param CASE_AOA=0,4 --param CASE_VELOCITY=50,100

Use --multishots to change the number of solver cycles per project (defaults to 10):

glacium case-sweep --param CASE_AOA=0,4 --multishots 20

One project is created for each parameter combination and global_config.yaml is generated from the project's case.yaml. The command prints the generated UIDs.

Pipeline

Run a grid convergence study and spawn follow-up projects::

glacium pipeline --level 1 --level 2 --multishot "[10,300,300]"

The call executes the grid-convergence pipeline layout which creates one project per grid level using the grid_dep recipe, selects the mesh with the lowest drag and then generates and runs a single-shot project and optional MULTISHOT case with the chosen grid. Multishot projects use the multishot recipe. Use --layout to select another workflow and --pdf to merge all report PDFs into a single summary file.

List projects

glacium projects

Select a project

# select by number from `glacium projects`
glacium select 1

The selected UID is stored in ~/.glacium_current and used by other commands.

Run jobs

# run all pending jobs in the current project
glacium run

You can run specific jobs by name as well:

glacium run XFOIL_REFINE XFOIL_POLAR

Show job status

glacium list

The table now includes an index column so you can refer to jobs by number.

Manage individual jobs

# reset a job to PENDING
glacium job reset XFOIL_POLAR
glacium job reset 1  # via index

You can list all available job types with numbers:

glacium job --list

Select a job of the current project by its index:

glacium job select 1

Jobs can also be added or removed via their index:

glacium job add 1
glacium job remove 1

Sync projects with recipes

# refresh the job list of the current project
glacium sync

Update configuration

# rebuild global_config.yaml from case.yaml
glacium update

Display project info

glacium info

Print the case.yaml parameters and a few values from global_config.yaml for the current project.

Remove projects

# delete the selected project
glacium remove

Use --all to remove every project under runs in the current working directory.

External executables

Paths to third party programs can be configured in runs/<UID>/_cfg/global_config.yaml inside the current working directory. Important keys include POINTWISE_BIN, FENSAP_BIN and the newly added FLUENT2FENSAP_EXE pointing to fluent2fensap.exe on Windows.

Logging

Set GLACIUM_LOG_LEVEL to control the verbosity of the CLI. For example::

export GLACIUM_LOG_LEVEL=DEBUG

Development

All tests can be run with:

pytest

To enable automatic version management install the plugin once:

poetry self add "poetry-dynamic-versioning[plugin]"

poetry install will pull setuptools_scm as specified in pyproject.toml. Versions are taken from Git tags, e.g.:

git tag v1.2.0 -m "release"