Source code for sherpa_ai.actions.synthesize

from typing import Any

from langchain_core.language_models import BaseLanguageModel 
from loguru import logger 

from sherpa_ai.actions.base import BaseAction


SYNTHESIZE_DESCRIPTION = """{role_description}

Context: {context}

Action - Result History:
{history}

Given the context and the action-result history, please complete the task mentioned. Include any links you used from the context and history in the result.
Task: {task}
Result:
"""  # noqa: E501

SYNTHESIZE_DESCRIPTION_CITATION = """{role_description}

Context: {context}

Action - Result History:
{history}

Given the context and the action-result history, please complete the task mentioned. If you can cite some of the sentences in the Context, please use them in your response as much intact as possible. DO NOT Include any links you used from the context and history in the result.
Task: {task}
Result:
"""  # noqa: E501




class SynthesizeOutput(BaseAction):
    role_description: str
    llm: Any = None  # The BaseLanguageModel from LangChain is not compatible with Pydantic 2 yet
    description: str = SYNTHESIZE_DESCRIPTION
    add_citation: bool = False

    # Override the name and args from BaseAction
    name: str = "SynthesizeOutput"
    args: dict = {"task": "string", "context": "string", "history": "string"}
    usage: str = "Answer the question using conversation history with the user"

    def __init__(self, **kwargs):
        super().__init__(**kwargs)
        if self.add_citation:
            self.description = SYNTHESIZE_DESCRIPTION_CITATION



    def execute(self, task: str, context: str, history: str) -> str:
        prompt = self.description.format(
            task=task,
            context=context,
            history=history,
            role_description=self.role_description,
        )

        logger.debug("Prompt: {}", prompt)
        result = self.llm.predict(prompt)
        return result