Classes

NanGraph

Directed multi-edge in-memory graph.

Method (TS)	Method (Python)	Description
new NanGraph()	NanGraph()	Constructor
g.addNode(node)	g.add_node(dict)	Add node (replaces if exists)
g.addEdge(edge)	g.add_edge(from_, to, rel_type)	Add directed typed edge (auto-creates nodes)
g.getNode(id)	g.get_node(id)	Get node (or undefined/None)
g.hasNode(id)	g.has_node(id)	Does node exist?
g.getRelated(id, relType?, direction?)	g.get_related(id, rel_type?, direction?)	Outgoing/incoming/both neighbors with filter
g.nodeCount	g.node_count	Node count
g.edgeCount	g.edge_count	Edge count
g.edges	g.edges	Edge list
g.nodes	g.nodes	Node map

g.addNode({ id: 'api', type: 'aws.apigateway', label: 'API GW', attrs: { monthly_cost: 20 } });
g.addEdge({ from: 'api', to: 'lambda', relType: 'routes_to' });

g.add_node({"id": "api", "type": "aws.apigateway", "label": "API GW", "attrs": {"monthly_cost": 20}})
g.add_edge("api", "lambda", "routes_to")

All construction methods return the graph itself → chainable.

Analysis functions

TS	Python	Formula
traverseBFS(g, start, opts?)	traverse_bfs(g, start, rel_types?, max_depth?, direction?)	Reachable nodes (BFS), excl. start
traverseDFS(g, start, opts?)	traverse_dfs(g, start, rel_types?, max_depth?, direction?)	Reachable nodes (DFS), excl. start
blastRadius(g, start, opts?)	blast_radius(g, start, rel_types?, max_depth?)	downstream affected nodes
costChain(g, start, opts?)	cost_chain(g, start, rel_types?, max_depth?)	Σ monthly_cost (start + downstream)
teamCost(g, owner)	team_cost(g, owner)	Σ monthly_cost by attrs.owner
fanIn(g, id)	fan_in(g, id)	incoming edge count
singlePointsOfFailure(g, threshold?)	single_points_of_failure(g, threshold?)	nodes with fan-in >= threshold

TS options:

interface TraversalOptions {
  relTypes?: string[];      // only these relations
  maxDepth?: number;        // depth limit
  direction?: 'outgoing' | 'incoming' | 'both';
}

Importers

Build a NanGraph from data.

TS	Python	Input
fromObject(obj)	from_object(dict)	JS object / dict (BYaML or dependency-graph)
fromYaml(str)	from_yaml(str)	YAML string
fromJson(str)	from_json(str)	JSON string

Supported formats:

• BYaML v0.3: { components: [...], relationships: [{from,to,type}] }
• dependency-graph.yaml: { nodes: {id: props}, edges: [{from,to,relType}] }

Types

interface GraphNode {
  id: string;
  label?: string;
  type?: string;
  attrs?: Record<string, unknown>;
}
interface GraphEdge {
  from: string;
  to: string;
  relType: string;
  label?: string;
}

Direction semantics

Value	Behavior
outgoing	this → others
incoming	others → this
both	both

Notes

• Zero deps in TS core (YAML import uses lightweight yaml). Python uses PyYAML.
• Works in Node.js, Cloudflare Workers, Lambda (TS) and Python 3.10+ (Py).
• Same API across TS and Python — column by column parity.