haive.agents.base.types

Core type system for the Haive agent framework.

Defines type variables, constraints, and base protocols for type-safe agent design.

Classes

Agent	Complete agent protocol combining all capabilities.
AgentInput	Default input schema for agents.
AgentOutput	Default output schema for agents.
AgentState	Default state schema for agents.
EngineProvider	Protocol for objects that provide engines.
GraphProvider	Protocol for objects that provide graphs.
GraphSegment	Represents a segment of a graph that can be composed.
HookContext	Context passed to hooks.
HookPoint	Standard hook points in agent lifecycle.
Invokable	Protocol for objects that can be invoked.
NodeConnection	Represents a connection between nodes in a graph.
StateProvider	Protocol for objects that provide state schemas.

Module Contents

class haive.agents.base.types.Agent

Bases: GraphProvider[TState], StateProvider[TState], Invokable[TInput, TOutput], EngineProvider[TEngine], Protocol[TEngine, TInput, TOutput, TState]

Complete agent protocol combining all capabilities.

class haive.agents.base.types.AgentInput(/, **data)

Bases: pydantic.BaseModel

Default input schema for agents.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:

data (Any)

class haive.agents.base.types.AgentOutput(/, **data)

Bases: pydantic.BaseModel

Default output schema for agents.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:

data (Any)

class haive.agents.base.types.AgentState(/, **data)

Bases: pydantic.BaseModel

Default state schema for agents.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:

data (Any)

class haive.agents.base.types.EngineProvider

Bases: Protocol[TEngine]

Protocol for objects that provide engines.

property engine: TEngine

Get the primary engine.

Return type:

TEngine

property engines: dict[str, haive.core.engine.base.Engine]

Get all engines.

Return type:

dict[str, haive.core.engine.base.Engine]

class haive.agents.base.types.GraphProvider

Bases: Protocol[TState]

Protocol for objects that provide graphs.

build_graph()

Build and return the graph.

Return type:

Any

class haive.agents.base.types.GraphSegment(/, **data)

Bases: pydantic.BaseModel, Generic[TState]

Represents a segment of a graph that can be composed.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:

data (Any)

class haive.agents.base.types.HookContext(/, **data)

Bases: pydantic.BaseModel, Generic[TState]

Context passed to hooks.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:

data (Any)

class haive.agents.base.types.HookPoint

Bases: str, enum.Enum

Standard hook points in agent lifecycle.

Initialize self. See help(type(self)) for accurate signature.

class haive.agents.base.types.Invokable

Bases: Protocol[TInput, TOutput]

Protocol for objects that can be invoked.

async ainvoke(input_data, config=None)

Async invoke with input data.

Parameters:

• input_data (TInput)

• config (dict[str, Any] | None)

Return type:

TOutput

invoke(input_data, config=None)

Invoke with input data.

Parameters:

• input_data (TInput)

• config (dict[str, Any] | None)

Return type:

TOutput

class haive.agents.base.types.NodeConnection(/, **data)

Bases: pydantic.BaseModel, Generic[TState]

Represents a connection between nodes in a graph.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:

data (Any)

class haive.agents.base.types.StateProvider

Bases: Protocol[TState]

Protocol for objects that provide state schemas.

property state_schema: type[TState]

Get the state schema type.

Return type:

type[TState]