fix: propagate optimizer_model to algorithm

[Redtius]

Description

Hello! I've been experimenting with DeepEval lately—y'all have done a very good job on this project.
However, I noticed a blocker while experimenting with MIPROv2, especially when using a custom LLM implementing DeepEvalBaseLLM. Even though I passed the optimizer_model to the PromptOptimizer constructor, I ran into the following exception:
[MIPROv2] • error DeepEvalError: MIPROv2 requires an optimizer_model for instruction proposal. Set it via PromptOptimizer. • halted before first iteration

Root Cause

After digging into the source, I found that self.algorithm.optimizer_model was remaining None because it wasn't being assigned during the configuration step in PromptOptimizer. Since MIPROv2 relies on a proposer phase before the main optimization loop, it needs this model to be explicitly set on the algorithm instance.
I’ve submitted this PR as a quick fix to ensure the model is correctly propagated. If you have a better architectural idea for handling this assignment, I'd be glad to help adjust it!

Reproductible Case

I prepared a mock case that reproduces the issue without requiring any external API keys or complex setups:

from deepeval.models.base_model import DeepEvalBaseLLM
from deepeval.optimizer import PromptOptimizer
from deepeval.optimizer.algorithms import MIPROV2
from deepeval.dataset import Golden
from deepeval.prompt import Prompt
from deepeval.metrics import BaseMetric
from deepeval.test_case import LLMTestCase

class MockLLM(DeepEvalBaseLLM):
    def __init__(self):
        pass
    def load_model(self):
        return self
    def generate(self, prompt: str) -> str:
        return "response"
    async def a_generate(self, prompt: str) -> str:
        return "response"
    def get_model_name(self):
        return "LLM"

mock_model = MockLLM()
goldens = [Golden(input="Hi", expected_output="Hello"),Golden(input="Hru?", expected_output="Fine, U?")]
prompt = Prompt(text_template="You are a mock assistant.")

class MockMetric(BaseMetric):
    def __init__(self):
        self.threshold = 0.5
    def measure(self, test_case: LLMTestCase):
        self.score = 1.0
        return self.score
    async def a_measure(self, test_case: LLMTestCase):
        return self.measure(test_case)
    def is_successful(self):
        return True
    @property
    def __name__(self):
        return "Metric"

def model_callback(prompt: Prompt,thread_id: str) -> str:
    return "Mocked output"

optimizer = PromptOptimizer(
    algorithm=MIPROV2(),
    optimizer_model=mock_model,
    model_callback=model_callback,
    metrics=[MockMetric()]
)
try:
    result = optimizer.optimize(prompt=prompt, goldens=goldens)
    print(result.text_template)
except Exception as e:
    print(f"\nBUG: {e}")
    print("\nEvidence: Even though 'optimizer_model' was passed to PromptOptimizer,")
    print(f"the algorithm instance has optimizer_model = {optimizer.algorithm.optimizer_model}")

[vercel]

Someone is attempting to deploy a commit to the Confident AI Team on Vercel.

A member of the Team first needs to authorize it.

[greptile-apps]

PR author is not in the allowed authors list.

[Redtius]

@penguine-ip