from tools.base import Tool

from tools.utils import load_prompt

from typing import Any

class RecommendDeescalationTool(Tool):

    def openai_spec(self, legacy=False):
        return {
            "name": self.name,
            "description": self.description,
            "parameters": self.args_schema
        }
    """
    Tool to recommend narrowing (deescalation) of antibiotics based on culture and sensitivity results.

    This tool uses microbiology results and current antibiotic regimen to generate a prompt for an LLM or rules engine to recommend deescalation.
    """
    def __init__(self) -> None:
        """
        Initialize the RecommendDeescalationTool with its name, description, and argument schema.
        """
        super().__init__()
        self.name = "recommend_deescalation"
        self.description = (
            "Recommend narrowing antibiotics based on culture & sensitivity."
        )
        self.args_schema = {
            "type": "object",
            "properties": {
                "culture": {"type": "string", "description": "Culture & sensitivity results"},
                "meds":    {"type": "string", "description": "Current antibiotic regimen"},
                "site_of_infection": {"type": "string", "description": "Site of infection (e.g., lung, urine, blood, etc.)"},
                "risk_of_biofilm": {"type": "string", "description": "Risk or presence of biofilm (e.g., prosthetic material, indwelling device, etc.)"},
                "current_response": {"type": "string", "description": "Current response to antibiotics (e.g., improving, stable, worsening)"},
                "creatinine_clearance": {"type": "string", "description": "Current creatinine clearance or renal function"},
                "severity_of_infection": {"type": "string", "description": "Severity of infection (e.g., mild, moderate, severe, septic shock)"},
                "known_allergies": {"type": "string", "description": "Known drug allergies (comma-separated or free text)"}
            },
            "required": ["culture", "meds", "site_of_infection", "risk_of_biofilm", "current_response", "creatinine_clearance", "severity_of_infection", "known_allergies"]
        }

    async def run(
        self,
        culture: str,
        meds: str,
        site_of_infection: str,
        risk_of_biofilm: str,
        current_response: str,
        creatinine_clearance: str,
        severity_of_infection: str,
        known_allergies: str
    ) -> str:
        """
        Generate a deescalation recommendation based on all relevant clinical variables.

        Args:
            culture (str): Culture & sensitivity results.
            meds (str): Current antibiotic regimen.
            site_of_infection (str): Site of infection.
            risk_of_biofilm (str): Risk or presence of biofilm.
            current_response (str): Current response to antibiotics.
            creatinine_clearance (str): Renal function.
            severity_of_infection (str): Severity of infection.

        Returns:
            str: The deescalation recommendation (placeholder).
        """
        prompt: str = load_prompt(
            "deescalation.j2",
            culture=culture,
            meds=meds,
            site_of_infection=site_of_infection,
            risk_of_biofilm=risk_of_biofilm,
            current_response=current_response,
            creatinine_clearance=creatinine_clearance,
            severity_of_infection=severity_of_infection,
            known_allergies=known_allergies
        )
        # Placeholder for LLM call
        return (
            f"Deescalation recommendation for: {culture}, {meds}, "
            f"Site: {site_of_infection}, Biofilm: {risk_of_biofilm}, Response: {current_response}, "
            f"CrCl: {creatinine_clearance}, Severity: {severity_of_infection}, Allergies: {known_allergies}"
        )