Structured Responses

Generate structured, typed data from AI responses using Pydantic models. This ensures consistent output format and enables easy data validation.

Basic Usage

Simple Model

from datapizza.clients.openai import OpenAIClient
from pydantic import BaseModel

client = OpenAIClient(api_key="your-api-key", model="gpt-4o-mini")

class Person(BaseModel):
    name: str
    age: int
    occupation: str

response = client.structured_response(
    input="Create a profile for a software engineer",
    output_cls=Person
)

person = response.structured_data[0]
print(f"Name: {person.name}")
print(f"Age: {person.age}")
print(f"Occupation: {person.occupation}")

Complex Models

from typing import List
from pydantic import BaseModel, Field

class Product(BaseModel):
    name: str
    price: float = Field(gt=0, description="Price must be positive")
    tags: List[str]
    in_stock: bool

class Store(BaseModel):
    name: str
    location: str
    products: List[Product]

response = client.structured_response(
    input="Create a tech store with 3 products",
    output_cls=Store
)

store = response.structured_data[0]
print(f"Store: {store.name}")
for product in store.products:
    print(f"- {product.name}: ${product.price}")

Data Extraction

Extract Information from Text

class ContactInfo(BaseModel):
    name: str
    email: str
    phone: str
    company: str

text = """
Hi, I'm John Smith from TechCorp. 
You can reach me at john.smith@techcorp.com or call 555-0123.
"""

response = client.structured_response(
    input=f"Extract contact information from this text: {text}",
    output_cls=ContactInfo
)

contact = response.structured_data[0]
print(f"Contact: {contact.name} at {contact.company}")

Analyze and Categorize

from enum import Enum

class Sentiment(str, Enum):
    POSITIVE = "positive"
    NEGATIVE = "negative"
    NEUTRAL = "neutral"

class TextAnalysis(BaseModel):
    sentiment: Sentiment
    confidence: float = Field(ge=0, le=1)
    key_topics: List[str]
    summary: str

review = "This product is amazing! Great quality and fast shipping."

response = client.structured_response(
    input=f"Analyze this review: {review}",
    output_cls=TextAnalysis
)

analysis = response.structured_data[0]
print(f"Sentiment: {analysis.sentiment}")
print(f"Confidence: {analysis.confidence}")
print(f"Topics: {', '.join(analysis.key_topics)}")