Struct CausalGraph 

pub struct CausalGraph {
    pub n_nodes: usize,
    pub parents: Vec<Vec<usize>>,
    pub children: Vec<Vec<usize>>,
    pub names: Vec<String>,
}

A directed acyclic graph (DAG) representing causal relationships between variables.

Nodes are identified by usize indices. Edges represent direct causal effects from parent to child. The graph must be acyclic for causal semantics to be valid.

Fields

n_nodes: usize

Number of nodes in the graph.

parents: Vec<Vec<usize>>

Adjacency list: parents[v] gives all parent nodes of v.

children: Vec<Vec<usize>>

Adjacency list: children[v] gives all children of v.

names: Vec<String>

Optional variable names.

Implementations

impl CausalGraph

pub fn new(n: usize) -> Self

Create a new empty causal graph with n nodes.

Arguments

• n — number of nodes

pub fn set_names(&mut self, names: &[&str])

Set variable names.

Arguments

• names — slice of names, length must equal n_nodes

pub fn add_edge(&mut self, from: usize, to: usize)

Add a directed edge from from (parent/cause) to to (child/effect).

Panics

Panics if adding this edge would create a cycle.

pub fn creates_cycle(&self, from: usize, to: usize) -> bool

Check whether adding edge from→to would create a cycle.

pub fn topological_sort(&self) -> Option<Vec<usize>>

Return nodes in topological order (Kahn’s algorithm).

Returns None if the graph has a cycle (should not happen if edges are added via add_edge).

pub fn ancestors(&self, v: usize) -> HashSet<usize>

Return all ancestors of node v (nodes from which v is reachable).

pub fn descendants(&self, v: usize) -> HashSet<usize>

Return all descendants of node v.

pub fn d_separated(&self, x: &[usize], y: &[usize], z: &[usize]) -> bool

Test d-separation: are node sets x and y d-separated by conditioning set z?

Uses the Bayes Ball algorithm. Returns true if x ⊥ y | z.

pub fn markov_blanket(&self, v: usize) -> HashSet<usize>

Return the Markov blanket of node v: parents, children, and co-parents.

pub fn is_acyclic(&self) -> bool

Check if the graph is acyclic.

Trait Implementations

impl Clone for CausalGraph

fn clone(&self) -> CausalGraph

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CausalGraph

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

Formats the value using the given formatter.