assistant.py

from typing import Optional, Dict, Any, Union, Generator
from huggingface_hub import InferenceClient
from openai import OpenAI
from prompt_template import PromptTemplate

class AIAssistant:
    def __init__(self, client: Union[OpenAI, InferenceClient], model: str):
        self.client = client
        self.model = model

    def generate_response(
        self,
        prompt_template: PromptTemplate,
        messages: list[Dict[str, str]],
        generation_params: Optional[Dict] = None,
        stream: bool = True,
    ) -> Generator[str, None, None]:
        """
        Generate LLM response using the provided template and parameters.

        Args:
            prompt_template: PromptTemplate object containing template and parameters
            messages: List of message dictionaries with role and content
            generation_params: Optional generation parameters (overrides template parameters)
            stream: Whether to stream the response

        Yields:
            Streamed response text
        """
        params = generation_params or prompt_template.parameters

        # Ensure messages are in correct format
        formatted_messages = [
            {"role": msg["role"], "content": str(msg["content"])}
            for msg in messages
        ]

        try:
            completion = self.client.chat.completions.create(
                model=self.model,
                messages=formatted_messages,
                stream=stream,
                **params
            )

            if stream:
                response = ""
                for chunk in completion:
                    if chunk.choices[0].delta.content is not None:
                        response += chunk.choices[0].delta.content
                        yield response
            else:
                return completion.choices[0].message.content

        except Exception as e:
            yield f"Error generating response: {str(e)}"