agents.planning.rewoo_tree_agent¶

ReWOO Tree-based Planning Agent with Parallelizable Execution.

This agent implements the ReWOO (Reasoning without Observation) methodology with tree-based planning for parallelizable execution. It features:

Hierarchical tree planning with recursive decomposition
Parallelizable node execution with proper dependencies
Tool aliasing and forced tool choice
Structured output models with field validators
Plan-and-execute pattern with dynamic recompilation
LLM Compiler inspired parallelization

Reference: - ReWOO: https://langchain-ai.github.io/langgraph/tutorials/rewoo/rewoo/#solver - LLM Compiler: https://langchain-ai.github.io/langgraph/tutorials/llm-compiler/LLMCompiler/

Classes¶

`PlanNode`	A node in the planning tree representing a task.
`PlanTree`	A tree structure representing the complete execution plan.
`ReWOOTreeAgent`	ReWOO Tree-based Planning Agent with Parallelizable Execution.
`ReWOOTreeAgentState`	Enhanced state for ReWOO tree agent.
`ReWOOTreeExecutorOutput`	Structured output for the ReWOO tree executor.
`ReWOOTreePlannerOutput`	Structured output for the ReWOO tree planner.
`TaskPriority`	Priority levels for tasks.
`TaskStatus`	Status of tasks in the planning tree.
`TaskType`	Types of tasks in the planning tree.
`ToolAlias`	Tool alias configuration for forced tool choice.

Module Contents¶

class agents.planning.rewoo_tree_agent.PlanNode(/, **data)¶

Bases: pydantic.BaseModel

A node in the planning tree representing a task.

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)

add_child(child_id)¶

Add a child node.

Parameters:: child_id (str)
Return type:: None

add_dependency(dependency_id)¶

Add a dependency.

Parameters:: dependency_id (str)
Return type:: None

can_execute(completed_nodes)¶

Check if this node can be executed given completed nodes.

Parameters:: completed_nodes (set[str])
Return type:: bool

mark_completed(result=None)¶

Mark the task as completed.

Parameters:: result (Any)
Return type:: None

mark_failed(error)¶

Mark the task as failed.

Parameters:: error (str)
Return type:: None

mark_started()¶

Mark the task as started.

Return type:: None

validate_dependencies()¶

Validate dependency relationships.

Return type:: PlanNode

classmethod validate_id(v)¶

Validate node ID format.

Parameters:: v (str)
Return type:: str

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

class agents.planning.rewoo_tree_agent.PlanTree(/, **data)¶

Bases: pydantic.BaseModel

A tree structure representing the complete execution plan.

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)

add_node(node)¶

Add a node to the tree.

Parameters:: node (PlanNode)
Return type:: None

get_completion_percentage()¶

Get the completion percentage of the plan.

Return type:: float

get_node(node_id)¶

Get a node by ID.

Parameters:: node_id (str)
Return type:: PlanNode | None

get_parallelizable_nodes()¶

Get nodes organized by parallelizable execution levels.

Return type:: list[list[PlanNode]]

get_ready_nodes()¶

Get nodes that are ready for execution.

Return type:: list[PlanNode]

has_failures()¶

Check if there are any failed nodes.

Return type:: bool

is_complete()¶

Check if the entire plan is complete.

Return type:: bool

mark_node_completed(node_id, result=None)¶

Mark a node as completed.

Parameters:

node_id (str)
result (Any)

Return type:

None

mark_node_failed(node_id, error)¶

Mark a node as failed.

Parameters:

node_id (str)
error (str)

Return type:

None

classmethod validate_id(v)¶

Validate plan tree ID format.

Parameters:: v (str)
Return type:: str

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

class agents.planning.rewoo_tree_agent.ReWOOTreeAgent(**kwargs)¶

Bases: haive.agents.base.agent.Agent

ReWOO Tree-based Planning Agent with Parallelizable Execution.

This agent implements the ReWOO methodology with enhancements: - Hierarchical tree planning with recursive decomposition - Parallelizable execution with dependency management - Tool aliasing for forced tool choice - Structured output models with validation - LLM Compiler inspired optimizations

Initialize ReWOO Tree Agent.

add_tool_alias(alias, actual_tool, force_choice=True, **params)¶

Add a tool alias for forced tool choice.

Parameters:

alias (str)
actual_tool (str)
force_choice (bool)

Return type:

None

build_graph()¶

Build the execution graph for ReWOO tree agent.

Return type:: haive.core.graph.state_graph.base_graph2.BaseGraph

class agents.planning.rewoo_tree_agent.ReWOOTreeAgentState(/, **data)¶

Bases: haive.core.schema.prebuilt.messages_state.MessagesState

Enhanced state for ReWOO tree agent.

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 agents.planning.rewoo_tree_agent.ReWOOTreeExecutorOutput(/, **data)¶

Bases: pydantic.BaseModel

Structured output for the ReWOO tree executor.

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].

class agents.planning.rewoo_tree_agent.ReWOOTreePlannerOutput(/, **data)¶

Bases: pydantic.BaseModel

Structured output for the ReWOO tree planner.

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)

validate_consistency()¶

Validate consistency between plan components.

Return type:: ReWOOTreePlannerOutput

classmethod validate_plan_id(v)¶

Validate plan ID format.

Parameters:: v (str)
Return type:: str

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

class agents.planning.rewoo_tree_agent.TaskPriority¶

Bases: str, enum.Enum

Priority levels for tasks.

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

class agents.planning.rewoo_tree_agent.TaskStatus¶

Bases: str, enum.Enum

Status of tasks in the planning tree.

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

class agents.planning.rewoo_tree_agent.TaskType¶

Bases: str, enum.Enum

Types of tasks in the planning tree.

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

class agents.planning.rewoo_tree_agent.ToolAlias(/, **data)¶

Bases: pydantic.BaseModel

Tool alias configuration for forced tool choice.

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)

classmethod validate_alias(v)¶

Validate alias format.

Parameters:: v (str)
Return type:: str

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