Struct GraphIndex 

pub struct GraphIndex<'a> { /* private fields */ }

A borrowed, per-traversal accelerator for SchemaGraph::lookup.

SchemaGraph::lookup is a linear scan over defs. When a single traversal (value validation, rendering, ref resolution) resolves many SchemaType::Refs against a wide graph, those repeated scans dominate. GraphIndex borrows a graph once and, for graphs above a small size threshold, builds a one-shot hash index so each subsequent lookup is O(1). Small graphs keep the plain linear scan, because building and hashing an index is a net loss there (and would otherwise regress narrow inputs such as a typical agent method parameter record).

It only borrows the graph and never mutates it, so SchemaGraph stays a pure data/wire model (no embedded lazy cache that would travel through serde / desert / protobuf / poem derives).

Implementations

impl<'a> GraphIndex<'a>

pub fn new(graph: &'a SchemaGraph) -> GraphIndex<'a>

Borrow graph and prepare a lookup accelerator for it.

pub fn defs_len(&self) -> usize

Number of named definitions in the indexed graph.

pub fn lookup(&self, id: &TypeId) -> Option<&'a SchemaTypeDef>

Look up a named definition by its TypeId, with the same semantics as SchemaGraph::lookup.