FEBio workflows¶

porous_media prepares FEBio input templates and processes exported results. Install the solver, FEBio Studio, and any required research plugins separately. The package does not execute simulations or decode .xplt databases directly.

Render input templates¶

Use create_feb_files to render one file per simulation context:

from pathlib import Path

from porous_media.febio.febio_manipulation import create_feb_files

work_dir = Path("results") / "febio"
work_dir.mkdir(parents=True, exist_ok=True)
template = work_dir / "model.feb.template"
template.write_text(
    '<febio_spec version="4.0">\n'
    "  <!-- Example parameter: {{ boundary_pressure }} -->\n"
    "</febio_spec>\n",
    encoding="utf-8",
)
create_feb_files(
    feb_template=template,
    feb_data={
        "low": {"boundary_pressure": 133},
        "high": {"boundary_pressure": 266},
    },
    output_dir=work_dir / "inputs",
)

This demonstrates template rendering and creates model_low.feb and model_high.feb. The minimal template is not a complete simulation model. Replace it with a validated model containing the geometry, materials, boundary conditions, units, and plugins required for your study.

Template variables use Jinja2 syntax, for example {{ boundary_pressure }}. The mapping keys identify simulations and become part of output filenames. The filename prefix is the template name before its first dot. Simulation identifiers must be single portable filename components, without directory separators, traversal, or Windows device names. This prevents a context key from writing outside the selected output directory. Generated output files cannot be symbolic links. Existing files are preserved by default; pass overwrite=True to replace regular output files explicitly.

Missing template variables now raise an error. Context values are XML-escaped, and an immutable Jinja2 sandbox blocks private Python attributes. Templates and context objects must still be trusted and reviewed: sandboxing does not impose CPU or memory limits or prevent side effects in supplied Python objects. Avoid passing callables or objects with privileged methods. The helper does not validate the resulting FEBio model; validate it with the solver before use.

When migrating existing templates, supply every required variable and remove assumptions that unescaped strings can insert XML. Use Jinja2's safe filter only for reviewed XML fragments. Parameter contexts are no longer printed to output.

Run the solver and export results¶

1. Open and validate the rendered model using your FEBio installation.
2. Run the model using the solver version and plugins required by that model.
3. Export the desired time steps and fields from FEBio Studio as legacy VTK files, keeping one simulation in each directory.
4. Convert the exported files using vtks_to_xdmf and inspect their time range and fields before analysis.

Use the official FEBio resources for solver installation, current commands, model specifications, and export support. Solver binaries and plugin compatibility depend on the installation.

The converter expects the final token of each legacy VTK file's second line to contain the numerical time. Keep filenames ordered by time and preserve a fixed mesh. See VTK and XDMF data for the precise input assumptions and visualization for image and video export.

See the FEBio API reference and research archive for related material.