agents.patterns.react_with_structured_output¶

ReactAgent with Structured Output Pattern.

This pattern demonstrates ReactAgent → SimpleAgentV3 sequential execution using the generalized hook system for structured output workflows.

Pattern: ReactAgent (reasoning/tools) → SimpleAgent (structured output) Use Cases: - Analysis with structured results - Research with formatted reports - Problem-solving with structured solutions - Tool-based workflows with typed outputs

Classes¶

`AnalysisResult`	Structured analysis result.
`ProblemSolution`	Structured problem solution.
`ReactWithStructuredOutput`	ReactAgent → SimpleAgentV3 with structured output pattern.
`ResearchReport`	Structured research report.

Functions¶

`calculator`(expression)	Calculate mathematical expressions.
`create_react_analysis_workflow`([name, tools])	Create a ReactAgent analysis workflow with structured output.
`create_react_problem_solving_workflow`([name, tools])	Create a ReactAgent problem-solving workflow with structured output.
`create_react_research_workflow`([name, tools])	Create a ReactAgent research workflow with structured output.
`data_analyzer`(data)	Analyze data and provide insights.
`example_analysis_workflow`()	Example: ReactAgent analysis with structured output.
`example_problem_solving_workflow`()	Example: ReactAgent problem-solving with structured solution.
`example_research_workflow`()	Example: ReactAgent research with structured report.
`main`()	Run all workflow examples.
`web_search`(query)	Search the web for information.

Module Contents¶

class agents.patterns.react_with_structured_output.AnalysisResult(/, **data)¶

Bases: pydantic.BaseModel

Structured analysis result.

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.patterns.react_with_structured_output.ProblemSolution(/, **data)¶

Bases: pydantic.BaseModel

Structured problem solution.

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.patterns.react_with_structured_output.ReactWithStructuredOutput(**data)¶

Bases: haive.agents.multi.agent.MultiAgent

ReactAgent → SimpleAgentV3 with structured output pattern.

This pattern uses ReactAgent for reasoning and tool usage, then SimpleAgentV3 for converting the results to structured output.

Initialize the pattern with agents.

classmethod create_analysis_pattern(name='analysis_workflow', tools=None, reasoning_config=None, structuring_config=None)¶

Create a ReactAgent → StructuredOutput pattern for analysis tasks.

Parameters:

name (str) – Name for the workflow
tools (list | None) – Tools for the ReactAgent
reasoning_config (haive.core.engine.aug_llm.AugLLMConfig | None) – Configuration for reasoning agent
structuring_config (haive.core.engine.aug_llm.AugLLMConfig | None) – Configuration for structuring agent

Returns:

Configured ReactWithStructuredOutput instance

Return type:

ReactWithStructuredOutput

classmethod create_problem_solving_pattern(name='problem_solving_workflow', tools=None, reasoning_config=None, structuring_config=None)¶

Create a ReactAgent → StructuredOutput pattern for problem-solving tasks.

Parameters:

name (str)
tools (list | None)
reasoning_config (haive.core.engine.aug_llm.AugLLMConfig | None)
structuring_config (haive.core.engine.aug_llm.AugLLMConfig | None)

Return type:

ReactWithStructuredOutput

classmethod create_research_pattern(name='research_workflow', tools=None, reasoning_config=None, structuring_config=None)¶

Create a ReactAgent → StructuredOutput pattern for research tasks.

Parameters:

name (str)
tools (list | None)
reasoning_config (haive.core.engine.aug_llm.AugLLMConfig | None)
structuring_config (haive.core.engine.aug_llm.AugLLMConfig | None)

Return type:

ReactWithStructuredOutput

class agents.patterns.react_with_structured_output.ResearchReport(/, **data)¶

Bases: pydantic.BaseModel

Structured research report.

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)

agents.patterns.react_with_structured_output.calculator(expression)¶

Calculate mathematical expressions.

Parameters:: expression (str)
Return type:: str

agents.patterns.react_with_structured_output.create_react_analysis_workflow(name='react_analysis', tools=None)¶

Create a ReactAgent analysis workflow with structured output.

Parameters:

name (str) – Workflow name
tools (list | None) – Tools for analysis (web search, calculators, etc.)

Returns:

Configured analysis workflow

Return type:

ReactWithStructuredOutput

agents.patterns.react_with_structured_output.create_react_problem_solving_workflow(name='react_problem_solver', tools=None)¶

Create a ReactAgent problem-solving workflow with structured output.

Parameters:

name (str) – Workflow name
tools (list | None) – Tools for problem-solving (analysis tools, simulators, etc.)

Returns:

Configured problem-solving workflow

Return type:

ReactWithStructuredOutput

agents.patterns.react_with_structured_output.create_react_research_workflow(name='react_research', tools=None)¶

Create a ReactAgent research workflow with structured output.

Parameters:

name (str) – Workflow name
tools (list | None) – Tools for research (web search, document analysis, etc.)

Returns:

Configured research workflow

Return type:

ReactWithStructuredOutput

agents.patterns.react_with_structured_output.data_analyzer(data)¶

Analyze data and provide insights.

Parameters:: data (str)
Return type:: str

async agents.patterns.react_with_structured_output.example_analysis_workflow()¶: Example: ReactAgent analysis with structured output.

async agents.patterns.react_with_structured_output.example_problem_solving_workflow()¶: Example: ReactAgent problem-solving with structured solution.

async agents.patterns.react_with_structured_output.example_research_workflow()¶: Example: ReactAgent research with structured report.

async agents.patterns.react_with_structured_output.main()¶: Run all workflow examples.

agents.patterns.react_with_structured_output.web_search(query)¶

Search the web for information.

Parameters:: query (str)
Return type:: str