hap.models.graph

Attributes

AgentGraph
AgentNode

Classes

HAPGraph	HAP graph with agent orchestration capabilities.
HAPNode	HAP node that can contain an agent.

Module Contents

class hap.models.graph.HAPGraph(/, **data: Any)

Bases: pydantic.BaseModel

HAP graph with agent orchestration capabilities.

Parameters:

data (Any)

add_agent_node(node_id: str, agent: haive.agents.base.agent.Agent, next_nodes: List[str] = None) → HAPNode

Add an agent as a node to the graph.

Parameters:

• node_id (str)

• agent (haive.agents.base.agent.Agent)

• next_nodes (List[str])

Return type:

HAPNode

add_entrypoint_node(node_id: str, entrypoint: str, next_nodes: List[str] = None) → HAPNode

Add a node by entrypoint string.

Parameters:

• node_id (str)

• entrypoint (str)

• next_nodes (List[str])

Return type:

HAPNode

async execute(initial_context: Dict[str, Any]) → hap.models.context.HAPContext

Execute the entire graph.

Parameters:

initial_context (Dict[str, Any])

Return type:

hap.models.context.HAPContext

topological_order() → List[str]

Get topological ordering of nodes.

Return type:

List[str]

entry_node: str = None

nodes: Dict[str, HAPNode] = None

class hap.models.graph.HAPNode(/, **data: Any)

Bases: pydantic.BaseModel

HAP node that can contain an agent.

Parameters:

data (Any)

async execute(context: HAPContext) → HAPContext

Execute this node’s agent.

Parameters:

context (HAPContext)

Return type:

HAPContext

load_agent() → haive.agents.base.agent.Agent

Load agent from entrypoint if not already loaded.

Return type:

haive.agents.base.agent.Agent

agent_entrypoint: str = None

agent_instance: haive.agents.base.agent.Agent | None = None

id: str = None

next_nodes: List[str] = None

hap.models.graph.AgentGraph

hap.models.graph.AgentNode