SchemaForge Python SDK

English | 中文

About

This is the official Python SDK for SchemaForge AI, a unified service for AI structured data processing using Pydantic models to define output formats, supporting multiple AI large language models.

Introduction

SchemaForge is a powerful Python SDK for quickly transforming unstructured data into structured formats and generating Pydantic models from text descriptions. It provides an intuitive API for:

• Structuring text data: Convert unstructured text into structured Pydantic models
• Generating models from sample data: Create Pydantic models based on sample data
• Async support: All operations support both synchronous and asynchronous calls

Installation

pip install schemaforge

Quick Start

Initialize Client

from schemaforge import SchemaForge

# Initialize client with API key
client = SchemaForge(api_key="your_secure_api_key_here")

# Or get API key from environment variable
# export SCHEMAFORGE_API_KEY=your_secure_api_key_here
client = SchemaForge()

Structure Text Data

from pydantic import BaseModel

# Define a Pydantic model
class Person(BaseModel):
    name: str
    age: int
    occupation: str

# Structure text using the model
person = client.structure(
    content="John is a 30-year-old software engineer",
    model_class=Person
)
print(person)

Generate Model from Sample Data

# Sample JSON data
json_sample = '''
{
    "product_id": "P12345",
    "name": "Smart Watch",
    "price": 199.99,
    "in_stock": true,
    "specifications": {
        "screen_size": "6.5 inches",
        "processor": "SnapDragon 8",
        "storage": "128GB",
        "camera": "48MP"
    },
    "colors": ["Black", "Silver", "Gold"],
    "release_date": "2024-01-15"
}
'''

# Generate model from sample data
result = client.generate_model(
    sample_data=json_sample,
    model_name="Product",
    description="Product information model with specifications"
)

if result.get("success"):
    # Load the generated model
    Product = client.load_model(result["model_code"])
    print(Product)

Async Support

import asyncio
from schemaforge import SchemaForge

async def main():
    client = SchemaForge(api_key="your_secure_api_key_here")
    
    # Async structure text
    person = await client.astructure(
        content="Jane is a 25-year-old data scientist",
        model_class=Person
    )
    print(person)
    
    # Async generate model
    result = await client.agenerate_model(
        sample_data='{"name": "Laptop", "price": 999.99}',
        model_name="Laptop",
        description="Laptop information model"
    )
    
    if result.get("success"):
        Laptop = client.load_model(result["model_code"])
        print(Laptop)

asyncio.run(main())

Advanced Usage

Custom Configuration

from schemaforge import SchemaForge

# Create client with custom configuration
client = SchemaForge(
    api_key="your_secure_api_key_here",
    api_base="https://custom-endpoint.schemaforge.ai",
    default_model="gpt-4",
    timeout=60.0,
    max_retries=3,
    retry_delay=1.0,
    retry_jitter=True,
    verbose=True
)

Error Handling

from schemaforge import SchemaForge
from schemaforge.exceptions.api_error import SchemaForgeError

client = SchemaForge(api_key="your_secure_api_key_here")

try:
    person = client.structure(
        content="Invalid text",
        model_class=Person
    )
except SchemaForgeError as e:
    print(f"Error: {e}")

Further Reading

For complete API documentation and more examples, visit the SchemaForge Documentation.