_structured - Chirp

ai._structured

module

Structured output — LLM JSON responses to typed models.

Generates JSON schema from dataclass fields or Pydantic models and parses LLM responses back into typed instances.

Structured output — LLM JSON responses to typed models.

Generates JSON schema from dataclass fields or Pydantic models and parses LLM responses back into typed instances.

ai._structured

Name	Type	Default	Description
`type`		—
`qualified_name`		—
`element_type`		—
`description`		—
`source_file`		—
`line_number`		—
`is_autodoc`		—
`autodoc_element`		—
`_autodoc_template`		—
`_autodoc_url_path`		—
`_autodoc_page_type`		—
`title`		—
`doc_content_hash`		—

Symbols on this page

function is_structured_type

Return True if cls is a supported structured-output target.

function schema_for_type

Generate a JSON schema for a dataclass or Pydantic model.

function dataclass_to_schema

Generate a JSON schema from a frozen dataclass.

function pydantic_to_schema

Generate a JSON schema from a PydanticBaseModelsubclass.

function parse_structured

Parse an LLM text response into a dataclass or Pydantic model.

function structured_repair_prompt

Build a follow-up user message asking the model to repair bad JSON.

function _parse_pydantic_model

Parse JSON dict into a Pydantic model instance.

function _inline_json_schema

Flatten a Pydantic$defsschema into a single object schema.

function _type_to_schema

Convert a Python type annotation to a JSON schema fragment.

function _extract_json

Extract JSON from LLM text, handling markdown code fences.

is_structured_type

function

def is_structured_type(cls: type[Any]) -> bool

Return True if cls is a supported structured-output target.

Parameters

Name	Type	Default	Description
`cls`	`type[Any]`	—

schema_for_type

function

def schema_for_type(cls: type[Any]) -> dict[str, Any]

Generate a JSON schema for a dataclass or Pydantic model.

Parameters

Name	Type	Default	Description
`cls`	`type[Any]`	—

dataclass_to_schema

function

def dataclass_to_schema(cls: type[T]) -> dict[str, Any]

Generate a JSON schema from a frozen dataclass.

Parameters

Name	Type	Default	Description
`cls`	`type[T]`	—

pydantic_to_schema

function

def pydantic_to_schema(cls: type[Any]) -> dict[str, Any]

Generate a JSON schema from a PydanticBaseModelsubclass.

Parameters

Name	Type	Default	Description
`cls`	`type[Any]`	—

parse_structured

function

def parse_structured(cls: type[T], text: str) -> T

Parse an LLM text response into a dataclass or Pydantic model.

Parameters

Name	Type	Default	Description
`cls`	`type[T]`	—
`text`	`str`	—

structured_repair_prompt

function

def structured_repair_prompt(*, error: StructuredOutputError, bad_text: str) -> str

Build a follow-up user message asking the model to repair bad JSON.

Parameters

Name	Type	Default	Description
`error`	`StructuredOutputError`	—
`bad_text`	`str`	—

_parse_pydantic_model

function

def _parse_pydantic_model(cls: type[Any], data: dict[str, Any]) -> Any

Parse JSON dict into a Pydantic model instance.

Parameters

Name	Type	Default	Description
`cls`	`type[Any]`	—
`data`	`dict[str, Any]`	—

_inline_json_schema

function

def _inline_json_schema(schema: dict[str, Any]) -> dict[str, Any]

Flatten a Pydantic$defsschema into a single object schema.

Parameters

Name	Type	Default	Description
`schema`	`dict[str, Any]`	—

_type_to_schema

function

def _type_to_schema(annotation: Any) -> dict[str, Any]

Convert a Python type annotation to a JSON schema fragment.

Parameters

Name	Type	Default	Description
`annotation`	`Any`	—

_extract_json

function

def _extract_json(text: str) -> str

Extract JSON from LLM text, handling markdown code fences.

Parameters

Name	Type	Default	Description
`text`	`str`	—

View source · /home/runner/work/chirp/chirp/site/../src/chirp/ai/_structured.py:1