Structured Outputs

Use Cases

Structured Outputs enables models to generate responses that conform to the JSON Schema you provide, making generated results more controllable and easier to parse. This feature facilitates parsing and processing in downstream logic, and also helps integrate results into business systems. It is suitable for a variety of automation and data processing scenarios.

Supported Models

The following models support structured outputs:

Usage

Add the following information to your request:

Set parameters: Specify your defined JSON Schema through the response_format parameter.
Prompt guidance: Instruct the model in your prompt to produce structured output.

Example

The following provides a complete Python code example demonstrating how to use Structured Outputs to generate a JSON response that conforms to the JSON Schema you provide.

1. Initialize the Client

You need to initialize the client with your JieKou AI API key.

from openai import OpenAI

client = OpenAI(
    base_url="https://api.highwayapi.ai/openai",
    api_key="<Your API Key>",
)

model = "qwen/qwen-2.5-72b-instruct"

2. Define the JSON Schema

You need to define a JSON Schema. The following example creates a JSON Schema that extracts expense information from user input.

# Define the system prompt for expense tracking.
system_prompt = """You are an expense tracking assistant.
Extract expense information from the user's input and format it according to the provided schema."""

# Define the JSON Schema for structured responses.
response_format = {
    "type": "json_schema",
    "json_schema": {
        "name": "expense_tracking_schema",
        "schema": {
            "type": "object",
            "properties": {
                "expenses": {
                    "type": "array",
                    "items": {
                        "type": "object",
                        "properties": {
                            "description": {
                                "type": "string",
                                "description": "Description of the expense"
                            },
                            "amount": {
                                "type": "number",
                                "description": "Amount spent in dollars"
                            },
                            "date": {
                                "type": "string",
                                "description": "When the expense occurred"
                            },
                            "category": {
                                "type": "string",
                                "description": "Category of expense (e.g., food, office, travel)"
                            }
                        },
                        "required": [
                            "description",
                            "amount"
                        ]
                    }
                },
                "total": {
                    "type": "number",
                    "description": "Total amount of all expenses"
                }
            },
            "required": [
                "expenses",
                "total"
            ],
        },
    },
}

3. Make an API Request

Create an API request. This request includes the response_format parameter, which specifies the JSON schema defined in the previous step.

chat_completion = client.chat.completions.create(
    model=model,
    messages=[
        {
            "role": "system",
            "content": system_prompt,
        },
        {
            "role": "user",
            "content": """I spent $120 on dinner at an Italian restaurant last Friday with my colleagues.
Also bought office supplies for $45 on Monday.""",
        },
    ],
    max_tokens=1024,
    temperature=0.8,
    stream=False,
    response_format=response_format,
)

response_content = chat_completion.choices[0].message.content

# Parse and prettify JSON
try:
    json_response = json.loads(response_content)
    prettified_json = json.dumps(json_response, indent=2)
    print(prettified_json)
except json.JSONDecodeError:
    print("Could not parse response as JSON. Raw response:")
    print(response_content)

Output:

{
  "expenses": [
    {
      "date": "2023-03-17",
      "description": "Dinner at Italian restaurant",
      "amount": 120,
      "category": "Food & Dining"
    },
    {
      "date": "2023-03-13",
      "description": "Office supplies",
      "amount": 45,
      "category": "Office Supplies"
    }
  ],
  "total": 165
}

Complete Code

from openai import OpenAI
import json

client = OpenAI(
    base_url="https://api.highwayapi.ai/openai",
    api_key="<Your API Key>",
)

model = "qwen/qwen-2.5-72b-instruct"

# Example of using JSON Schema for structured outputs
# This example creates a schema for extracting expense information

# Define the system prompt for expense tracking
system_prompt = """You are an expense tracking assistant.
Extract expense information from the user's input and format it according to the provided schema."""

# Define the JSON schema for structured responses
response_format = {
    "type": "json_schema",
    "json_schema": {
        "name": "expense_tracking_schema",
        "schema": {
            "type": "object",
            "properties": {
                "expenses": {
                    "type": "array",
                    "items": {
                        "type": "object",
                        "properties": {
                            "description": {
                                "type": "string",
                                "description": "Description of the expense"
                            },
                            "amount": {
                                "type": "number",
                                "description": "Amount spent in dollars"
                            },
                            "date": {
                                "type": "string",
                                "description": "When the expense occurred"
                            },
                            "category": {
                                "type": "string",
                                "description": "Category of expense (e.g., food, office, travel)"
                            }
                        },
                        "required": [
                            "description",
                            "amount"
                        ]
                    }
                },
                "total": {
                    "type": "number",
                    "description": "Total amount of all expenses"
                }
            },
            "required": [
                "expenses",
                "total"
            ],
        },
    },
}

chat_completion = client.chat.completions.create(
    model=model,
    messages=[
        {
            "role": "system",
            "content": system_prompt,
        },
        {
            "role": "user",
            "content": """I spent $120 on dinner at an Italian restaurant last Friday with my colleagues.
Also bought office supplies for $45 on Monday.""",
        },
    ],
    max_tokens=1024,
    temperature=0.8,
    stream=False,
    response_format=response_format,
)

response_content = chat_completion.choices[0].message.content

# Parse and prettify JSON
try:
    json_response = json.loads(response_content)
    prettified_json = json.dumps(json_response, indent=2)
    print(prettified_json)
except json.JSONDecodeError:
    print("Could not parse response as JSON. Raw response:")
    print(response_content)

Getting Started

LLM API

Model Providers

Model Features

Third-party Tool Setup

Use Cases

Supported Models

Usage

Example

1. Initialize the Client

2. Define the JSON Schema

3. Make an API Request

Complete Code

​Use Cases

​Supported Models

​Usage

​Example

​1. Initialize the Client

​2. Define the JSON Schema

​3. Make an API Request

​Complete Code

Use Cases

Supported Models

Usage

Example

1. Initialize the Client

2. Define the JSON Schema

3. Make an API Request

Complete Code