Struct Classification 

pub struct Classification {
    pub ode_states: Vec<String>,
    pub observed_unknowns: Vec<String>,
    pub inlined_unknowns: Vec<String>,
    pub algebraic_unknowns: Vec<String>,
    pub brownian_parameters: Vec<String>,
    pub discrete_parameters: Vec<String>,
    pub sampled_parameters: Vec<String>,
    pub constant_parameters: Vec<String>,
    pub system_kind: SystemKind,
    pub observed_definitions: BTreeMap<String, Expr>,
}

One model’s complete §6.3.1 classification, computed once. Every name list is sorted.

The free functions below (ode_states, brownian_parameters, …) are thin wrappers over this; build it directly when a caller needs several sets, so the equation walk happens once.

Fields

ode_states: Vec<String>

Unknowns appearing under D(·, t) on some equation LHS.

observed_unknowns: Vec<String>

Unknowns defined by a bare-variable OR indexed-variable LHS (esm-spec §6.3.1) — the semantic “defined by an equation” set.

inlined_unknowns: Vec<String>

The NARROWER set alongside Self::observed_unknowns: the observed unknowns whose defining LHS is a BARE variable, so they are eliminable by inlining (esm-spec §6.3.1, “Note that eliminable and inlineable are not the same thing”). An ARRAYED definition (y[i] ~ f(i)) is observed too, but it materializes into a buffer its consumers index rather than being substituted away, so it is NOT here. Spelled inlined_unknowns in the Python oracle; it does not narrow the §6.3.1 partition, it sits beside it.

algebraic_unknowns: Vec<String>

Unknowns constrained only implicitly.

brownian_parameters: Vec<String>

Parameters whose update is wiener.

discrete_parameters: Vec<String>

Parameters carrying any OTHER update.

sampled_parameters: Vec<String>

Parameters with a distribution and no update.

constant_parameters: Vec<String>

Parameters with neither.

system_kind: SystemKind

The DERIVED system kind.

observed_definitions: BTreeMap<String, Expr>

Each observed unknown’s defining RHS, keyed by the unknown’s name — the 1.0.0 home of what used to sit in variables[v].expression.

Implementations

impl Classification

pub fn of(model: &Model) -> Classification

Classify one model node. Classification is per MODEL NODE: a subsystem’s equations classify the subsystem’s variables, never the parent’s.

pub fn from_parts( variables: &IndexMap<String, ModelVariable>, equations: &[Equation], ) -> Classification

Classify from the raw parts, for callers holding a variables map and an equation list rather than a whole Model (the JSON-level cadence pass, the flatten/compile pipelines).

pub fn from_json(model: &Value) -> Result<Classification, Error>

Classify a model held as raw JSON — the form the cadence-partition pass and the CLI work in.

Reads only variables and equations, and IGNORES every other key, so a model carrying the document registries merged down onto it (as crate::cadence::model_with_loaders does) classifies the same as the bare one.

A VARIABLE that does not deserialize is an error: mis-reading the declared type set is exactly the failure this module exists to prevent, and it would silently classify as “no variables”. An individual EQUATION that does not deserialize is SKIPPED, because the passes that call this run over partially-lowered intermediate JSON whose well-formedness is the schema layer’s business, not classification’s.

pub fn is_ode_state(&self, name: &str) -> bool

Membership test for Classification::ode_states.

pub fn is_observed(&self, name: &str) -> bool

Membership test for Classification::observed_unknowns.

pub fn is_inlined(&self, name: &str) -> bool

Membership test for Classification::inlined_unknowns.

pub fn is_brownian(&self, name: &str) -> bool

Membership test for Classification::brownian_parameters.

pub fn is_discrete_parameter(&self, name: &str) -> bool

Membership test for Classification::discrete_parameters.

pub fn unknowns(&self) -> Vec<String>

Every unknown of the model, sorted — the union of the three unknown sets.

pub fn parameters(&self) -> Vec<String>

Every parameter of the model, sorted — the union of the four parameter sets.

pub fn assert_partitions(&self, model: &Model) -> Result<(), String>

Assert the two §6.3.1 partition laws against the model this was derived from: the three unknown sets partition the unknowns, and the four parameter sets partition the parameters — both disjointly and totally.

Returns the first violation as an Err, so a caller can surface it as a diagnostic rather than panicking.

Trait Implementations

impl Clone for Classification

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Classification

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for Classification

fn default() -> Self

Returns the “default value” for a type.