[FEA]: Support Pydantic objects as input to `LLMEngine` stage

[ashsong-nv]

Is this a new feature, an improvement, or a change to existing functionality?

New Feature

How would you describe the priority of this feature request

Medium

Please provide a clear description of problem this feature solves

I want Morpheus to support Pydantic objects as inputs to the LLMEngine stage. The vulnerability-analysis blueprint pipeline mainly uses Pydantic objects throughout the pipeline and needs to convert to DF to use the LLMEngine stage, then convert back from DF to Pydantic.

Describe your ideal solution

N/A

Additional context

Relevant parts of the pipeline: https://github.com/NVIDIA-AI-Blueprints/vulnerability-analysis/blob/main/src/cve/pipeline/pipeline.py#L170-L201

Code of Conduct

• I agree to follow this project's Code of Conduct
• I have searched the open feature requests and have found no duplicates for this feature request

[dagardner-nv]

I was in this bit of code yesterday, nodes on the top of my head.

On the C++ side the LLMEngine and LLMContext code is currently directly tied to ControlMessage, however in both cases the actual code interactions is based on extracting the DataFrame from the ControlMessage and transforming it into a Python dictionary, and pulling the task from the metadata.

We would just need a pydantic overload.