agents.planning_v2.base.tree_models¶

Tree-based planning models using the enhanced tree_leaf structure.

This module provides planning models that leverage the generic tree/leaf structure from haive-core for more flexible and type-safe planning.

Classes¶

PlanContent

Content for plan nodes (both tasks and sub-plans).

PlanResult

Result of executing a plan node.

PlanStatus

Status for plan nodes.

TaskPlan

A concrete plan implementation using the tree structure.

Functions¶

create_phased_plan(objective, phases)

Create a plan with phases (sub-plans).

create_simple_plan(objective, tasks)

Create a simple linear plan from a list of task names.

Module Contents¶

class agents.planning_v2.base.tree_models.PlanContent(/, **data)¶

Bases: pydantic.BaseModel

Content for plan nodes (both tasks and sub-plans).

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_v2.base.tree_models.PlanResult(/, **data)¶

Bases: pydantic.BaseModel

Result of executing a plan node.

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_v2.base.tree_models.PlanStatus¶

Bases: str, enum.Enum

Status for plan nodes.

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

class agents.planning_v2.base.tree_models.TaskPlan¶

Bases: PlanTree

A concrete plan implementation using the tree structure.

This class extends the generic Tree to provide planning-specific functionality while maintaining type safety.

Examples

# Create a plan plan = TaskPlan(content=PlanContent(

objective=”Deploy new feature”, priority=4

))

# Add simple tasks plan.add_task(“Write tests”, priority=5) plan.add_task(“Code review”, priority=3)

# Add sub-plan deploy_plan = plan.add_subplan(“Deploy to production”) deploy_plan.add_task(“Deploy to staging”) deploy_plan.add_task(“Run smoke tests”) deploy_plan.add_task(“Deploy to prod”)

# Check status print(f”Total tasks: {plan.total_nodes}”) print(f”Progress: {plan.progress_percentage}%”)

add_parallel_tasks(tasks)¶

Add multiple tasks that can execute in parallel.

Parameters:

tasks (list[tuple[str, int]]) – List of (objective, priority) tuples

Returns:

List of created task nodes

Return type:

list[PlanLeaf]

add_subplan(objective, description='', priority=1)¶

Add a sub-plan that can contain its own tasks.

Parameters:

  • objective (str)

  • description (str)

  • priority (int)

Return type:

TaskPlan

add_task(objective, description='', priority=1)¶

Add a simple task to the plan.

Parameters:

  • objective (str)

  • description (str)

  • priority (int)

Return type:

PlanLeaf

get_blocked_tasks()¶

Get all tasks that are blocked.

Return type:

list[Union[PlanLeaf, TaskPlan]]

get_current_task()¶

Get the current task to execute (first pending or in-progress).

Return type:

Union[PlanLeaf, TaskPlan] | None

get_tasks_by_priority(min_priority=1)¶

Get all tasks with priority >= min_priority.

Parameters:

min_priority (int)

Return type:

list[Union[PlanLeaf, TaskPlan]]

mark_current_completed(output='Done')¶

Mark the current active task as completed.

Parameters:

output (str)

Return type:

bool

mark_current_failed(error)¶

Mark the current active task as failed.

Parameters:

error (str)

Return type:

bool

to_markdown(indent=0)¶

Convert plan to markdown representation.

Parameters:

indent (int)

Return type:

str

agents.planning_v2.base.tree_models.create_phased_plan(objective, phases)¶

Create a plan with phases (sub-plans).

Parameters:

  • objective (str) – Main plan objective

  • phases (dict[str, list[str]]) – Dict of phase_name -> list of tasks

Return type:

TaskPlan

Examples

plan = create_phased_plan(

“Launch Product”, {

“Development”: [“Code”, “Test”, “Review”], “Deployment”: [“Stage”, “Verify”, “Prod”], “Marketing”: [“Announce”, “Demo”, “Feedback”]

}

)

agents.planning_v2.base.tree_models.create_simple_plan(objective, tasks)¶

Create a simple linear plan from a list of task names.

Parameters:
Return type:

TaskPlan