agents.sequential.config¶

Configuration for SequentialAgent that connects components in a linear workflow.

from typing import Any This module defines the configuration class for SequentialAgent, which automates the process of connecting multiple engine components in a sequence.

Classes¶

`SequentialAgentConfig`	Configuration for a SequentialAgent that connects components linearly.
`StepConfig`	Configuration for a single step in a sequential workflow.

Functions¶

`build_agent`(config)	Build a SequentialAgent from configuration.
`from_aug_llms`(aug_llms[, name, id, state_schema])	Create a SequentialAgentConfig from a list of AugLLMConfig instances.
`from_components`(components[, name, id, state_schema])	Create a SequentialAgentConfig from a list of components.
`from_steps`(steps[, name, id, entry_point, state_schema])	Create a SequentialAgentConfig from a list of steps.
`get_step_by_name`(config, name)	Get a step configuration by name.
`setup_components`(config)	Setup components for a configuration.
`validate_steps`(steps)	Validate that steps list is not empty.

Module Contents¶

class agents.sequential.config.SequentialAgentConfig¶

Bases: haive.core.engine.agent.config.AgentConfig

Configuration for a SequentialAgent that connects components linearly.

This agent automates the process of connecting engine components in a sequence, handling the data flow between them through the state schema.

Components can be any engine types, particularly AugLLMConfig instances for chaining language model steps.

build_agent()¶

Build and return a SequentialAgent instance.

Return type:: Any

classmethod from_aug_llms(aug_llms, name=None, id=None, state_schema=None, **kwargs)¶

Create a SequentialAgentConfig from a list of AugLLMConfig instances.

Convenience method for the common case of chaining LLM steps.

Parameters:

aug_llms (list[haive.core.engine.aug_llm.AugLLMConfig]) – List of AugLLMConfig instances
name (str | None) – Optional agent name
id (str | None) – Optional unique identifier
state_schema (type[haive.core.schema.state_schema.StateSchema] | None) – Optional state schema
**kwargs – Additional configuration parameters

Returns:

SequentialAgentConfig instance

Return type:

SequentialAgentConfig

classmethod from_components(components, name=None, id=None, state_schema=None, step_names=None, **kwargs)¶

Create a SequentialAgentConfig from a list of components.

This automatically creates step configurations for each component.

Parameters:

components (list[Any]) – List of components to use as steps (AugLLMConfig, Engine, etc.)
name (str | None) – Optional agent name
id (str | None) – Optional unique identifier
state_schema (type[haive.core.schema.state_schema.StateSchema] | None) – Optional state schema
step_names (list[str] | None) – Optional list of step names (must match length of components)
**kwargs – Additional configuration parameters

Returns:

SequentialAgentConfig instance

Return type:

SequentialAgentConfig

classmethod from_steps(steps, name=None, id=None, entry_point=None, state_schema=None, **kwargs)¶

Create a SequentialAgentConfig from a list of steps.

Parameters:

steps (list[StepConfig]) – List of step configurations
name (str | None) – Optional agent name
id (str | None) – Optional unique identifier
entry_point (str | None) – Optional entry point step name
state_schema (type[haive.core.schema.state_schema.StateSchema] | None) – Optional state schema
**kwargs – Additional configuration parameters

Returns:

SequentialAgentConfig instance

Return type:

SequentialAgentConfig

get_step_by_name(name)¶

Get a step configuration by name.

Parameters:: name (str)
Return type:: StepConfig | None

setup_components()¶

Collect all step components into the components list for schema derivation.

Return type:: SequentialAgentConfig

classmethod validate_steps(v)¶

Ensure we have at least one step.

Return type:: Any

class agents.sequential.config.StepConfig(/, **data)¶

Bases: pydantic.BaseModel

Configuration for a single step in a sequential workflow.

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)

model_config¶: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

agents.sequential.config.build_agent(config)¶

Build a SequentialAgent from configuration.

Parameters:: config (SequentialAgentConfig)
Return type:: haive.agents.sequential.agent.SequentialAgent

agents.sequential.config.from_aug_llms(aug_llms, name=None, id=None, state_schema=None, **kwargs)¶

Create a SequentialAgentConfig from a list of AugLLMConfig instances.

Parameters:

aug_llms (list[haive.core.engine.aug_llm.AugLLMConfig])
name (str | None)
id (str | None)
state_schema (type[haive.core.schema.state_schema.StateSchema] | None)

Return type:

SequentialAgentConfig

agents.sequential.config.from_components(components, name=None, id=None, state_schema=None, **kwargs)¶

Create a SequentialAgentConfig from a list of components.

Parameters:

components (list[Any])
name (str | None)
id (str | None)
state_schema (type[haive.core.schema.state_schema.StateSchema] | None)

Return type:

SequentialAgentConfig

agents.sequential.config.from_steps(steps, name=None, id=None, entry_point=None, state_schema=None, **kwargs)¶

Create a SequentialAgentConfig from a list of steps.

Parameters:

steps (list[StepConfig])
name (str | None)
id (str | None)
entry_point (str | None)
state_schema (type[haive.core.schema.state_schema.StateSchema] | None)

Return type:

SequentialAgentConfig

agents.sequential.config.get_step_by_name(config, name)¶

Get a step configuration by name.

Parameters:

config (SequentialAgentConfig)
name (str)

Return type:

StepConfig | None

agents.sequential.config.setup_components(config)¶

Setup components for a configuration.

Parameters:: config (SequentialAgentConfig)
Return type:: SequentialAgentConfig

agents.sequential.config.validate_steps(steps)¶

Validate that steps list is not empty.

Parameters:: steps (list[StepConfig])
Return type:: bool