from pydantic import BaseModel
from majordomo_llm import get_llm_instance

llm = get_llm_instance(provider="openai", model="gpt-4o")

# Simple request with automatic cost tracking
response = await llm.get_response(
    system_prompt="You are a helpful assistant.",
    user_prompt="Explain async/await in Python.",
)
print(response.text)
print(f"Cost: ${response.total_cost:.6f}")

# Structured output — one API for every provider
class MovieReview(BaseModel):
    title: str
    rating: float
    summary: str

review = await llm.get_structured_json_response(
    system_prompt="You are a movie critic.",
    user_prompt="Review the movie Inception.",
    output_type=MovieReview,
)
print(f"{review.title}: {review.rating}/10")

# Pydantic AI adapter
from pydantic_ai import Agent
from majordomo_frameworks.pydantic_ai import create_model, build_extra_headers

model = create_model(
    provider="openai",
    model_name="gpt-4o",
    gateway_url="http://localhost:7680",
)

agent = Agent(model)
result = await agent.run(
    "Summarize this document.",
    model_settings={"extra_headers": build_extra_headers(
        majordomo_key="your-key",
        app_name="my-agent",
    )},
)

# Agno adapter
from agno.agent import Agent as AgnoAgent
from agno.models.openai import OpenAIChat
from majordomo_frameworks.agno import patch_model

model = patch_model(
    OpenAIChat(id="gpt-4o"),
    gateway_url="http://localhost:7680",
    majordomo_key="your-key",
)

agent = AgnoAgent(model=model)