import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
from typing import Dict, List, Any

class EndpointHandler():
    def __init__(self, path=""):
      self.quant = quantization_config = BitsAndBytesConfig(load_in_4bit=True, bnb_4bit_compute_dtype=torch.float16)
      self.model = AutoModelForCausalLM.from_pretrained(path, device_map="auto", quantization_config=self.quant, trust_remote_code=True)
      self.tokenizer = AutoTokenizer.from_pretrained(path)
      self.pipeline = pipeline("text-generation", model=self.model, tokenizer=self.tokenizer)

    def __call__(self, data: Dict[str, Dict[str, Any]]) -> Any:
        inputs = data["inputs"]["msg"]
        parameters = data["args"]
        prediction = self.pipeline(inputs, **parameters)
        output = prediction[0]['generated_text']
        return output