agents.research.storm.state¶

State management for STORM research workflow.

This module provides Pydantic models for managing state throughout the STORM (Synthesis of Topic Outline through Retrieval and Multi-perspective questioning) research process, including topic definition, article generation, and research coordination.

Classes:: TopicState: Simple state container for research topic ArticleState: State container for final article content ResearchState: Complete research workflow state management

Example

Basic research state management:

from haive.agents.research.storm.state import ResearchState

state = ResearchState(
    topic=TopicState(topic="AI Safety"),
    outline=outline_instance,
    editors=editor_list,
    interview_results=interview_list,
    sections=section_list
)

draft = state.draft  # Get compiled article draft

Classes¶

`ArticleState`	State container for final article content.
`ResearchState`	Complete research workflow state management.
`TopicState`	Simple state container for research topic.

Module Contents¶

class agents.research.storm.state.ArticleState(/, **data)¶

Bases: pydantic.BaseModel

State container for final article content.

Parameters:: data (Any)

article¶: The complete final article text.

Example

>>> state = ArticleState(article="This is the final article...")
>>> print(len(state.article))
25

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.

class agents.research.storm.state.ResearchState(/, **data)¶

Bases: TopicState, ArticleState

Complete research workflow state management.

This class manages the entire STORM research process state, including topic definition, outline generation, editor perspectives, interview results, and final section compilation.

Parameters:: data (Any)

topic¶: The research topic state container.

outline¶: Generated outline for the research article.

editors¶: List of editor perspectives for multi-angle research.

interview_results¶: Results from perspective-based interviews.

sections¶: Final compiled sections for the article.

Properties:: draft: Compiled draft article from all sections.

Example

Complete research workflow:

from haive.agents.research.storm.state import ResearchState

state = ResearchState(
    topic=TopicState(topic="Climate Change"),
    outline=generated_outline,
    editors=editor_perspectives,
    interview_results=interview_data,
    sections=compiled_sections
)

# Get the complete draft
article_draft = state.draft
print(f"Draft length: {len(article_draft)} characters")

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.

property draft: str¶

Compile all sections into a single draft article.

Returns:: Complete article draft with all sections joined by double newlines.
Return type:: str

Example

>>> draft_text = research_state.draft
>>> print(draft_text[:100])
# Introduction

Climate change refers to…

class agents.research.storm.state.TopicState(/, **data)¶

Bases: pydantic.BaseModel

Simple state container for research topic.

Parameters:: data (Any)

topic¶: The research topic as a string.

Example

>>> state = TopicState(topic="Machine Learning Ethics")
>>> print(state.topic)
Machine Learning Ethics

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.