pyforestry.sweden.simulation.mortality package

Submodules

pyforestry.sweden.simulation.mortality.context module

Unified mortality context models for Sweden simulation orchestration.

pyforestry.sweden.simulation.mortality.engine module

Simulation-layer mortality orchestration for Swedish mortality modules.

This module is pyforestry’s own composition layer and carries no publication of its own: it selects among, and applies, mortality models that each carry their own provenance. The scientific sources are those of the composed modules – pyforestry.sweden.mortality.elfving_2013, pyforestry.sweden.mortality.fridman_stahl_2001, pyforestry.sweden.mortality.siipilehto_2020, pyforestry.sweden.mortality.soderberg_1986, pyforestry.sweden.mortality.bengtsson and pyforestry.sweden.mortality.naslund_1986 – reachable through MortalityEngine.components.

class pyforestry.sweden.simulation.mortality.engine.MortalityEngine(config: MortalityConfig = <factory>, rng: KeyedRNG | None = None)

Bases: object

Simulation-layer mortality orchestration for Sweden.

run is a single linear pipeline: resolve the probability source (retained override -> configured tree model), optionally calibrate (Bengtsson then Söderberg), apply adjustment factors, and realize the result (deterministic or stochastic).

property component_id: str

Stable identifier for this composition.

property components: tuple[Describable, ...]

The mortality modules this engine can compose, each with its own citation.

Returns the modules’ descriptors, not their ids. components means Sequence[Describable] everywhere else in this package, and every consumer reads component_id and source off each entry – so returning bare strings did not merely differ in type, it dropped the citations. pyforestry.projection._provenance reported a run through this engine as citing one thing, itself, and none of the seven papers that produced the numbers.

run(context: MortalityContext) → MortalityResult

Predict and realize mortality for the stand described by context.

property source: SourceReference

Provenance for the engine itself, which is a pyforestry composition.

The engine has no publication of its own. The scientific content belongs to the composed mortality modules; see components.

class pyforestry.sweden.simulation.mortality.engine.MortalityModel(*args, **kwargs)

Bases: Protocol

Structural interface shared by the interchangeable single-tree strategies.

The three tree-model facades (Fridman & Ståhl 2001, Elfving 2013, Siipilehto et al. 2020) all expose this signature, so the engine treats the configured model as one strategy instead of branching on its concrete type.

predict_probabilities(*, context: MortalityContext, period_years: float, default_stems_per_tree: float) → tuple[list[float], dict[str, Any]]

Return per-tree 5-year mortality probabilities and model diagnostics.

pyforestry.sweden.simulation.mortality.normalize module

Normalization helpers for unified mortality context.

pyforestry.sweden.simulation.mortality.normalize.normalize_mortality_context(context: MortalityContext) → MortalityContext

Return normalized context with copied mutable fields and non-null history.

Module contents

Sweden simulation-layer mortality context and orchestration engine.