examples

SchemaForge AI Examples

This directory contains comprehensive examples demonstrating how to use SchemaForge AI for structuring data and generating Pydantic models.

中文文档

Quick Overview

The examples in this directory show main ways to use SchemaForge AI:

1. Using Predefined Models - Define Pydantic models manually and use them to structure data
2. Auto-Generating Models - Generate Pydantic models automatically from sample data

File Descriptions

predefined_models.py

Demonstrates using predefined Pydantic models to structure data from text inputs with several practical examples:

Example	Description	Model Used
Person Information	Extracts personal details from text	Person
Book Information	Structures book metadata from descriptive text	BookInfo
News Article	Parses news articles into structured data	NewsArticle
Model Comparison	Compares different AI models on the same task	Multiple models
Custom Recipe Model	Shows how to create and use a custom model	Recipe

model_generation_example.py

Shows how to automatically generate Pydantic models from different types of data samples:

Example	Description	Data Format
Product Model	Generates a model from JSON product data	JSON
Customer Model with Requirements	Creates a model with specific validation requirements	Text
Book Model	Generates a model from CSV-like data	CSV
Complex Nested Order Model	Creates models with nested structures from complex JSON	JSON

Setting Up

Before running these examples, make sure:

1. You have set up the SchemaForge AI service:

   git clone https://github.com/X-Zero-L/schemaforge-ai.git
   cd schemaforge-ai
   uv sync

2. You have an API key in the .env file:

   API_KEY=your_api_key_here
   

3. The API service is running:

   uvicorn app.main:app --reload

Running the Examples

To run any example:

python examples/predefined_models.py

or

python examples/model_generation_example.py

Creating Your Own Models

You can create custom Pydantic models following this pattern:

from pydantic import BaseModel, Field
from typing import List, Optional

class YourModel(BaseModel):
    field1: str = Field(..., description="Description of field1")
    field2: int = Field(..., description="Description of field2")
    field3: Optional[List[str]] = Field(default_factory=list, description="Description of field3")

The description in each Field is important - it helps the AI understand what information to extract.

Generating Models

If you don't want to define models manually, you can use the model generation API:

result = await generate_model(
    sample_data="Your sample data (JSON, text, CSV, etc.)",
    model_name="YourModelName",
    description="Description of what the model represents",
    requirements="Optional specific requirements",
    llm_model_name="openai:gpt-4o"  # Specify AI model to use
)

# The generated model code is available in:
generated_code = result["model_code"]

Best Practices

1. Field Descriptions: Always provide clear descriptions for each field
2. System Prompts: Use system prompts to improve structuring accuracy for complex tasks
3. Model Selection: Different models have different strengths - experiment to find the best for your use case
4. Validation Rules: When generating models, specify validation requirements for better data quality
5. Error Handling: Always check the success field in responses and handle errors appropriately
6. SDK Usage: For production integration, consider using the SDK approach to simplify your code

Supported AI Models

The examples support various AI models including:

• OpenAI: gpt-3.5-turbo, gpt-4, gpt-4o
• Anthropic: claude-3-opus-latest, claude-3-7-sonnet-latest
• Google: gemini-1.5-pro, gemini-1.5-flash
• Mistral: mistral-large-latest, mistral-small-latest

Specify the model using the format: provider:model_name (e.g., openai:gpt-4o)