neograph::graph::GraphNode Class Reference

Abstract base class for all graph nodes. More...

#include <node.h>

Inheritance diagram for neograph::graph::GraphNode:

Public Member Functions
virtual std::string	get_name () const =0
	Get the node's unique name within the graph.
virtual asio::awaitable< NodeOutput >	run (NodeInput in)=0
	Unified dispatch entry for sync work, async work, and streaming.

Detailed Description

Abstract base class for all graph nodes.

Nodes are the building blocks of the graph. Each node reads from the graph state and produces channel writes (and optionally Command or Send directives). Override run(NodeInput) for every node shape.

Thread safety

Node instances are owned by the GraphEngine and shared across all concurrent run() invocations on that engine — including runs with different thread_ids. Implementations MUST therefore be either stateless (the recommended default — derive everything from the GraphState argument) or fully self-synchronized. Storing per-run scratch data in a member variable will silently corrupt parallel runs.

Per-execution state belongs in the channels: read inputs from state.get(...) and emit outputs as ChannelWrites.

Definition at line 96 of file node.h.

Member Function Documentation

get_name()

virtual std::string neograph::graph::GraphNode::get_name ( ) const

pure virtual

Get the node's unique name within the graph.

Returns

Node name string.

Implemented in neograph::graph::SyncGraphNode, neograph::a2a::A2ACallerNode, neograph::a2a::CopyNinjaNode, neograph::graph::LLMCallNode, neograph::graph::ToolDispatchNode, neograph::graph::IntentClassifierNode, and neograph::graph::SubgraphNode.

run()

virtual asio::awaitable< NodeOutput > neograph::graph::GraphNode::run ( NodeInput  in )

pure virtual

Unified dispatch entry for sync work, async work, and streaming.

New nodes override THIS method. Read in.state for channel inputs, read in.ctx for available per-run metadata (cancel token, thread ID, current step, …), check in.stream_cb for an optional LLM_TOKEN sink, and return a NodeOutput (alias for NodeResult) populated with channel writes plus optional Command / Send directives.

The engine (NodeExecutor::execute_node_with_retry_async) dispatches only through this method. Sync-vs-async is no longer a public virtual cross-product: perform immediate work before co_return, or co_await asynchronous dependencies.

class MyNode : public GraphNode {
    asio::awaitable<NodeOutput> run(NodeInput in) override {
        auto request = make_provider_request(*provider_, model_, in.state.get_provider_messages());
        auto outcome = co_await provider_->invoke_async(std::move(request));
        record_usage(in.ctx, outcome);
        outcome_or_throw(outcome);
        NodeOutput out;
        out.writes.push_back(provider_messages_write(outcome));
        co_return out;
    }
    std::string get_name() const override { return "my_node"; }
};

See also

NodeInput and NodeOutput above.

Lifetime note: in is taken by value so the coroutine frame owns its own copy of the struct. The state / ctx references inside still point to objects on the engine's suspended-frame stack (which outlive the run by construction), but the struct itself is no longer a reference parameter. The coroutine-reference-parameter UAF fixed in the v0.2.0 RunConfig path does not apply here.

Implemented in neograph::a2a::CopyNinjaNode, neograph::a2a::A2ACallerNode, neograph::graph::SyncGraphNode, neograph::graph::LLMCallNode, neograph::graph::ToolDispatchNode, neograph::graph::IntentClassifierNode, and neograph::graph::SubgraphNode.

---

The documentation for this class was generated from the following file:

• include/neograph/graph/node.h