Autosave¶

Contract¶

This page records the table, field, and stage autosave policies that schedule persistent output during a stage.

Governing equations¶

No physical equation is introduced; autosave is transport and persistence policy.

Symbols and SI units¶

Sampling periods are in seconds or in steps; quantities carry the units defined by their observables.

Assumptions and validity¶

Exactly one of the time or step cadence must be set; formats and layouts are validated immediately.

Python API¶

# %% Relax stage with table and field autosave
import fullmag as fm

nm = 1.0e-9

study = fm.study("autosave_api_example")
study.engine("fdm")
study.device("cpu", precision="double")
study.mode("strict")

study.objects.mesh.defaults(cell_size=(2 * nm, 2 * nm, 5 * nm))
film = study.geometry(fm.Box(100 * nm, 20 * nm, 5 * nm), name="film")
film.Ms = 800.0e3
film.Aex = 13.0e-12
film.alpha = 1.0
film.m = fm.init.UniformMagnetization((1.0, 0.1, 0.0))
study.exchange()

study.stages.add_relax(stage_id="relax", algorithm="projected_gradient_bb", max_steps=1000, tolT=1e-8).autosave(
    fm.StageAutosave(
        table=fm.TableAutosave(
            every_steps=10,
            quantities=["step", "mx", "my", "mz", "e_ex", "e_total", "max_torque_T"],
        ),
        fields=[fm.FieldAutosave("m", every_steps=100)],
    )
)

ProblemIR¶

StageAutosave.to_ir() emits kind, target, layout, format, table, and fields; a time policy lower to sample_period_s/every_seconds and "auto" lower to a Sinc-based sample_period_policy.

Round-trip and failure semantics¶

Requested intent is the value authored by Python and preserved in ProblemIR; resolved execution is the planner or realization result. Validation errors identify the violated domain rule, and unsupported combinations are rejected explicitly rather than silently substituted.

Ambiguous (both/neither) cadence, unsupported formats, duplicate field quantities, and txt with field outputs fail immediately.

Discrete realization¶

Cadence is consumed by the runtime sampler; persistence formats map to JSON/Zarr/HDF5 writers.

Implementation mapping¶

Anchor: packages/fullmag-py/src/fullmag/model/study.py (TableAutosave, FieldAutosave, StageAutosave).

Validation¶

Ownership and cadence tests exercise the policy normalizations.

Limitations¶

Autosave is authoring policy; actual artifact production requires an executed lane.

Scientific bibliography¶

No physical model is introduced.

Control Room crosswalk¶

Status: Table/field autosave and result inspection are partial; unsupported output formats remain not implemented.

not implemented: frontend support for output parameters not rendered by the autosave/result inspectors. See Control Room capability register for the support matrix and not implemented policy. Frontend source owner: apps/control-room/src/modules/inspector/panels/stages/AutosaveStageInspector.tsx (AutosaveStageInspector).

Source-code index¶