Fix code example in MLFlow section of deployment docs

[jmho]

The current example code in the MLFlow section of the deployment doc:

import dspy
import mlflow

mlflow.set_tracking_uri("http://127.0.0.1:5000/")
mlflow.set_experiment("deploy_dspy_program")

lm = dspy.LM("openai/gpt-4o-mini")
dspy.settings.configure(lm=lm)
dspy_program = dspy.ChainOfThought("question -> answer")

with mlflow.start_run():
    mlflow.dspy.log_model(
        dspy_program,
        "dspy_program",
        input_example={"messages": [{"role": "user", "content": "What is LLM agent?"}]},
        task="llm/v1/chat",
    )

Does not work given dspy>=2.6.23 and mlflow>=2.22.0. This has been referenced in both #8134 and mlflow/mlflow#15546 where the resolution in the latter was that MLFlow expects custom modules for mlflow.dspy.log_model(...) and not the primitive modules which is currently in the example.

This PR changes the MLFlow example in the deploy docs to match the working MLFlow docstring example found here for the mlflow.dspy.log_model(...) function.

[chenmoneygithub]

Thanks for the PR! I was able to deploy with the existing code 3 days ago. May I have your exact DSPy and MLflow version?

[jmho]

Hey @chenmoneygithub thanks for the help!

I created a reproduction repo here that has all of my deps, the code I ran in script, and the same code in notebook form. For reference I am, running dspy==2.6.23 and mlflow==2.22.0 and python=3.11.

I run mlflow ui in the activated venv and then run either the jupyter notebook cell or the script using uv run main.py and get:

TypeError('ChainOfThought.forward() takes 1 positional argument but 2 were given')Traceback (most recent call last):
  File "/Users/jmho/Desktop/Personal-Documents/projects/dspy-reproduction-repo/.venv/lib/python3.11/site-packages/mlflow/utils/_capture_modules.py", line 166, in load_model_and_predict
    model.predict(input_example, params=params)
  File "/Users/jmho/Desktop/Personal-Documents/projects/dspy-reproduction-repo/.venv/lib/python3.11/site-packages/mlflow/dspy/wrapper.py", line 89, in predict
    outputs = self.model(converted_inputs)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/jmho/Desktop/Personal-Documents/projects/dspy-reproduction-repo/.venv/lib/python3.11/site-packages/dspy/utils/callback.py", line 326, in sync_wrapper
    return fn(instance, *args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/jmho/Desktop/Personal-Documents/projects/dspy-reproduction-repo/.venv/lib/python3.11/site-packages/dspy/primitives/program.py", line 32, in __call__
    return self.forward(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: ChainOfThought.forward() takes 1 positional argument but 2 were given

[chenmoneygithub]

LGTM, thanks for the fix!