Struct CodeGraph 

pub struct CodeGraph {
    pub snapshot_id: SnapshotId,
    /* private fields */
}

The canonical dependency graph for a codebase snapshot.

Fields

snapshot_id: SnapshotId

Snapshot identifier for this graph as discussed before.

Implementations

impl CodeGraph

pub fn new(snapshot_id: SnapshotId) -> Self

Creates an empty CodeGraph for the given snapshot.

pub fn graph(&self) -> &DiGraph<NodeData, EdgeData>

Immutable access to the underlying petgraph graph for SCC analysis, serialization, and deploy algorithms.

pub fn add_node(&mut self, node: NodeData) -> NodeIndex

Adds a raw node to the graph. Node additions do not affect edge normalization, so this is the only direct mutation needed by deploy/test code that injects synthetic nodes post-build.

The caller remains responsible for registering the node in the index maps (file_to_index, symbol_to_index, external_index) when it is a File/Symbol/External node. This method does not sync those maps; a bare Data node needs no registration.

pub fn add_file_node(&mut self, node: FileNode) -> NodeIndex

Add a file node and register it in the file index.

pub fn add_symbol_node(&mut self, node: SymbolNode) -> NodeIndex

Add a symbol node and register it in the symbol index.

pub fn file_node_index(&self, file_id: FileId) -> Option<NodeIndex>

pub fn symbol_node_index(&self, symbol_id: SymbolId) -> Option<NodeIndex>

pub fn file_node(&self, file_id: FileId) -> Option<&FileNode>

pub fn symbol_node(&self, symbol_id: SymbolId) -> Option<&SymbolNode>

pub fn file_count(&self) -> usize

pub fn symbol_count(&self) -> usize

pub fn external_count(&self) -> usize

pub fn node_count(&self) -> usize

pub fn edge_count(&self) -> usize

pub fn get_or_create_external_node( &mut self, raw_path: String, language: LangId, ) -> NodeIndex

Resolves or creates the External node for a raw unresolved path, keeping external_index consistent so repeated loads reuse a single node.

pub fn add_edge_normalized( &mut self, source: NodeIndex, target: NodeIndex, kind: EdgeKind, confidence: f32, )

Adds an edge, normalizing duplicate (src, dst, kind) triples by max-merging confidence so injected edges obey the same invariant as builder-constructed ones. The O(1) edge_index makes duplicate detection independent of out-degree.

pub fn add_edge_normalized_with_flow( &mut self, source: NodeIndex, target: NodeIndex, kind: EdgeKind, confidence: f32, flow_kind: Option<FlowKind>, )

Adds a Flow edge with a specific flow kind, normalizing duplicates by max-merging confidence (same (src, dst, Flow) triple). When a duplicate edge exists, the first non-None flow_kind is preserved. The lookup is O(1) via the dedup index.

pub fn files(&self) -> impl Iterator<Item = (FileId, &FileNode)> + '_

pub fn symbols(&self) -> impl Iterator<Item = (SymbolId, &SymbolNode)> + '_

Files in path order.

Callers that need a stable sequence, such as a shard export or a partition, must not rely on hash map order.

pub fn edges_of_kind( &self, kind: EdgeKind, ) -> impl Iterator<Item = (NodeIndex, NodeIndex)> + '_

pub fn reference_edges( &self, ) -> impl Iterator<Item = (SymbolId, SymbolId, f32)> + '_

Reference edges as (source symbol, target symbol, confidence).

Skips edges whose endpoints are not symbol nodes. Use for navigation and confidence-ranked completion.

Trait Implementations

impl Clone for CodeGraph

fn clone(&self) -> CodeGraph

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CodeGraph

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

Formats the value using the given formatter.