history = [
    {"role": "user", "content": "User message 1"},
    {"role": "assistant", "content": "Assistant response 1"},
    {"role": "user", "content": "User message 2"},
    # ... more turns
]

import asyncio
from openjudge.models import OpenAIChatModel
from openjudge.graders.multi_turn import ContextMemoryGrader

async def main():
    model = OpenAIChatModel(model="qwen-max")
    grader = ContextMemoryGrader(model=model)

    result = await grader.aevaluate(
        history=[
            {"role": "user", "content": "I'm allergic to nuts and prefer vegetarian food."},
            {"role": "assistant", "content": "I'll remember that for recommendations."},
            {"role": "user", "content": "What should I order for dinner?"},
        ],
        response="I recommend the walnut pesto pasta!",  # Forgot nut allergy!
    )

    print(f"Score: {result.score}")
    print(f"Reason: {result.reason}")

asyncio.run(main())

import asyncio
from openjudge.models import OpenAIChatModel
from openjudge.graders.multi_turn import AnaphoraResolutionGrader

async def main():
    model = OpenAIChatModel(model="qwen-max")
    grader = AnaphoraResolutionGrader(model=model)

    result = await grader.aevaluate(
        history=[
            {"role": "user", "content": "I bought a laptop and a phone yesterday."},
            {"role": "assistant", "content": "Nice purchases!"},
            {"role": "user", "content": "Is it good for programming?"},  # "it" = laptop
        ],
        response="Yes, the laptop is excellent for programming.",
    )

    print(f"Score: {result.score}")
    print(f"Reason: {result.reason}")

asyncio.run(main())

import asyncio
from openjudge.models import OpenAIChatModel
from openjudge.graders.multi_turn import TopicSwitchGrader

async def main():
    model = OpenAIChatModel(model="qwen-max")
    grader = TopicSwitchGrader(model=model)

    result = await grader.aevaluate(
        history=[
            {"role": "user", "content": "What's the best programming language?"},
            {"role": "assistant", "content": "Python is great for beginners."},
            {"role": "user", "content": "By the way, what's the weather in Tokyo?"},  # Topic switch!
        ],
        response="I don't have real-time weather data, but you can check weather.com.",
    )

    print(f"Score: {result.score}")
    print(f"Reason: {result.reason}")

asyncio.run(main())

import asyncio
from openjudge.models import OpenAIChatModel
from openjudge.graders.multi_turn import SelfCorrectionGrader

async def main():
    model = OpenAIChatModel(model="qwen-max")
    grader = SelfCorrectionGrader(model=model)

    result = await grader.aevaluate(
        history=[
            {"role": "user", "content": "What's the capital of Australia?"},
            {"role": "assistant", "content": "The capital is Sydney."},  # Wrong!
            {"role": "user", "content": "I think that's incorrect."},
        ],
        response="You're right, I apologize. The capital of Australia is Canberra, not Sydney.",
    )

    print(f"Score: {result.score}")
    print(f"Reason: {result.reason}")

asyncio.run(main())

import asyncio
from openjudge.models import OpenAIChatModel
from openjudge.graders.multi_turn import InstructionClarificationGrader

async def main():
    model = OpenAIChatModel(model="qwen-max")
    grader = InstructionClarificationGrader(model=model)

    result = await grader.aevaluate(
        history=[
            {"role": "user", "content": "Book me a flight."},  # Vague query
        ],
        response="I'd be happy to help! Could you tell me your departure city, destination, and travel dates?",
    )

    print(f"Score: {result.score}")
    print(f"Reason: {result.reason}")

asyncio.run(main())

import asyncio
from openjudge.models import OpenAIChatModel
from openjudge.graders.multi_turn import ProactiveInteractionGrader

async def main():
    model = OpenAIChatModel(model="qwen-max")
    grader = ProactiveInteractionGrader(model=model)

    result = await grader.aevaluate(
        history=[
            {"role": "user", "content": "I'm planning a trip to Japan."},
        ],
        response="Japan is wonderful! When are you planning to go? I can suggest activities based on the season.",
    )

    print(f"Score: {result.score}")
    print(f"Reason: {result.reason}")

asyncio.run(main())

import asyncio
from openjudge.models import OpenAIChatModel
from openjudge.graders.multi_turn import ResponseRepetitionGrader

async def main():
    model = OpenAIChatModel(model="qwen-max")
    grader = ResponseRepetitionGrader(model=model)

    result = await grader.aevaluate(
        history=[
            {"role": "user", "content": "What is Python?"},
            {"role": "assistant", "content": "Python is a high-level programming language."},
            {"role": "user", "content": "Tell me more."},
        ],
        response="Python is a high-level programming language.",  # Exact repeat!
    )

    print(f"Score: {result.score}")
    print(f"Reason: {result.reason}")

asyncio.run(main())

import asyncio
from openjudge.models import OpenAIChatModel
from openjudge.graders.multi_turn import ContextMemoryGrader

def split_session_to_steps(messages):
    """Split session into (history, response) tuples for each assistant turn."""
    steps = []
    history = []
    for msg in messages:
        if msg["role"] == "assistant":
            steps.append((list(history), msg["content"]))
        history.append(msg)
    return steps

async def main():
    model = OpenAIChatModel(model="qwen-max")
    grader = ContextMemoryGrader(model=model)

    # Complete conversation session
    session = [
        {"role": "user", "content": "I'm allergic to nuts."},
        {"role": "assistant", "content": "I'll remember that."},
        {"role": "user", "content": "Suggest a snack."},
        {"role": "assistant", "content": "Try some fruit or yogurt!"},
        {"role": "user", "content": "And a dessert?"},
        {"role": "assistant", "content": "How about almond cake?"},  # Forgot allergy!
    ]

    # Evaluate each step
    steps = split_session_to_steps(session)
    scores = []

    for i, (history, response) in enumerate(steps):
        result = await grader.aevaluate(history=history, response=response)
        scores.append(result.score)
        print(f"Step {i+1}: {result.score}/5")

    # Aggregate scores
    print(f"\nAverage: {sum(scores)/len(scores):.2f}")
    print(f"Min: {min(scores)}")

asyncio.run(main())