Module graph 

DAG Traversal Algorithms for Trajectory Structure

Trajectories are DAGs (Directed Acyclic Graphs) where:

• Each node is an episode (unit of experience)
• Edges represent parent→child relationships
• Multiple children = regenerations or branching

This module provides high-performance algorithms for:

• Path finding (root to leaf traversal)
• Branch detection (identifying decision points)
• Primary path selection (choosing the “best” linear path through the DAG)

Structs

BranchInfo

Information about a branch point in the DAG.

Edge

Edge in the trajectory DAG.

Episode

An episode in the trajectory DAG.

PathResult

Result of finding a path through the DAG.

TrajectoryGraph

Trajectory DAG structure optimized for traversal operations.

Enums

EdgeType

Type of edge in the DAG.

PathSelectionPolicy

Policy for selecting which child to follow at branch points.

TraversalOrder

Traversal order for DAG walking.

Type Aliases

NodeId

Unique identifier for a node in the trajectory DAG.