tud_lbm.io

Input/Output operations for LBM simulations (public API).

This module consolidates all I/O functionality: configuration readers, output writers, and visualization tools.

Main Classes

SimulationIO

Manage simulation output to disk. Provides convenient interface for saving trajectories in multiple formats.

Example:

io = SimulationIO(base_dir="results", config=config)
final_state, trajectory = run(setup, state, nt=1000)
io.save(final_state, trajectory)
output_writers

Registry of output format handlers (NumPy, VTK).

Example:

writers = output_writers.available()  # ['numpy', 'vtk']
numpy_writer = output_writers['numpy']()

Submodules

readers

Configuration adapters for loading settings from files.

  • DictAdapter : Load from Python dicts

  • TomlAdapter : Load from TOML files

Example:

from tud_lbm.io.readers import TomlAdapter
config = TomlAdapter().load("config.toml")
plotting

Visualization operators and figure builders.

  • FigureBuilder : Assemble multi-panel figures from simulation data

  • Animator : Encode saved timestep frames into mp4/gif

  • PlotOperator : Base class for custom visualization panels

Example:

from tud_lbm.io.plotting import FigureBuilder
builder = FigureBuilder(config, "results/run_001/")
builder.build_all()
callbacks

Host-callback-based streaming I/O for use inside a lax.scan body.

  • make_save_callback : Build a do_save(state, t) callback that writes snapshots to disk via jax.debug.callback without breaking the JIT trace. Used internally by run() when io_handler is given.

Submodules